articles on correlational research

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Clin Kidney J
• v.14(11); 2021 Nov

Conducting correlation analysis: important limitations and pitfalls

Roemer j janse.

Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, The Netherlands

Tiny Hoekstra

Department of Nephrology, Amsterdam Cardiovascular Sciences, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands

Kitty J Jager

ERA-EDTA Registry, Department of Medical Informatics, Amsterdam Public Health Research Institute, Amsterdam UMC, University of Amsterdam, Amsterdam, The Netherlands

Carmine Zoccali

CNR-IFC, Center of Clinical Physiology, Clinical Epidemiology of Renal Diseases and Hypertension, Reggio Calabria, Italy

Giovanni Tripepi

Friedo w dekker, merel van diepen.

The correlation coefficient is a statistical measure often used in studies to show an association between variables or to look at the agreement between two methods. In this paper, we will discuss not only the basics of the correlation coefficient, such as its assumptions and how it is interpreted, but also important limitations when using the correlation coefficient, such as its assumption of a linear association and its sensitivity to the range of observations. We will also discuss why the coefficient is invalid when used to assess agreement of two methods aiming to measure a certain value, and discuss better alternatives, such as the intraclass coefficient and Bland–Altman’s limits of agreement. The concepts discussed in this paper are supported with examples from literature in the field of nephrology.

‘Correlation is not causation’: a saying not rarely uttered when a person infers causality from two variables occurring together, without them truly affecting each other. Yet, though causation may not always be understood correctly, correlation too is a concept in which mistakes are easily made. Nonetheless, the correlation coefficient has often been reported within the medical literature. It estimates the association between two variables (e.g. blood pressure and kidney function), or is used for the estimation of agreement between two methods of measurement that aim to measure the same variable (e.g. the Modification of Diet in Renal Disease (MDRD) formula and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula for estimating the glomerular filtration rate (eGFR)]. Despite the wide use of the correlation coefficient, limitations and pitfalls for both situations exist, of which one should be aware when drawing conclusions from correlation coefficients. In this paper, we aim to describe the correlation coefficient and its limitations, together with methods that can be applied to avoid these limitations.

The basics: the correlation coefficient

Fundamentals.

The correlation coefficient was described over a hundred years ago by Karl Pearson [ 1 ], taking inspiration from a similar idea of correlation from Sir Francis Galton, who developed linear regression and was the not-so-well-known half-cousin of Charles Darwin [ 2 ]. In short, the correlation coefficient, denoted with the Greek character rho ( ρ ) for the true (theoretical) population and r for a sample of the true population, aims to estimate the strength of the linear association between two variables. If we have variables X and Y that are plotted against each other in a scatter plot, the correlation coefficient indicates how well a straight line fits these data. The coefficient ranges from −1 to 1 and is dimensionless (i.e., it has no unit). Two correlations with r = −1 and r  = 1 are shown in Figure 1A and B , respectively. The values of −1 and 1 indicate that all observations can be described perfectly using a straight line, which in turn means that if X is known, Y can be determined deterministically and vice versa. Here, the minus sign indicates an inverse association: if X increases, Y decreases. Nonetheless, real-world data are often not perfectly summarized using a straight line. In a scatterplot as shown in Figure 1C , the correlation coefficient represents how well a linear association fits the data.

Different shapes of data and their correlation coefficients. ( A ) Linear association with r = −1. ( B ) A linear association with r  = 1. ( C ) A scatterplot through which a straight line could plausibly be drawn, with r  = 0.50. ( D ) A sinusoidal association with r  = 0. ( E ) A quadratic association with r  = 0. ( F ) An exponential association with r  = 0.50.

It is also possible to test the hypothesis of whether X and Y are correlated, which yields a P-value indicating the chance of finding the correlation coefficient’s observed value or any value indicating a higher degree of correlation, given that the two variables are not actually correlated. Though the correlation coefficient will not vary depending on sample size, the P-value yielded with the t -test will.

The value of the correlation coefficient is also not influenced by the units of measurement, but it is influenced by measurement error. If more error (also known as noise) is present in the variables X and Y , variability in X will be partially due to the error in X , and thus not solely explainable by Y . Moreover, the correlation coefficient is also sensitive to the range of observations, which we will discuss later in this paper.

An assumption of the Pearson correlation coefficient is that the joint distribution of the variables is normal. However, it has been shown that the correlation coefficient is quite robust with regard to this assumption, meaning that Pearson’s correlation coefficient may still be validly estimated in skewed distributions [ 3 ]. If desired, a non-parametric method is also available to estimate correlation; namely, the Spearman’s rank correlation coefficient. Instead of the actual values of observations, the Spearman’s correlation coefficient uses the rank of the observations when ordering observations from small to large, hence the ‘rank’ in its name [ 4 ]. This usage of the rank makes it robust against outliers [ 4 ].

Explained variance and interpretation

One may also translate the correlation coefficient into a measure of the explained variance (also known as R 2 ), by taking its square. The result can be interpreted as the proportion of statistical variability (i.e. variance) in one variable that can be explained by the other variable. In other words, to what degree can variable X be explained by Y and vice versa. For instance, as mentioned above, a correlation of −1 or +1 would both allow us to determine X from Y and vice versa without error, which is also shown in the coefficient of determination, which would be (−1) 2 or 1 2 = 1, indicating that 100% of variability in one variable can be explained by the other variable.

In some cases, the interpretation of the strength of correlation coefficient is based on rules of thumb, as is often the case with P-values (P-value <0.05 is statistically significant, P-value >0.05 is not statistically significant). However, such rules of thumb should not be used for correlations. Instead, the interpretation should always depend on context and purposes [ 5 ]. For instance, when studying the association of renin–angiotensin–system inhibitors (RASi) with blood pressure, patients with increased blood pressure may receive the perfect dosage of RASi until their blood pressure is exactly normal. Those with an already exactly normal blood pressure will not receive RASi. However, as the perfect dosage of RASi makes the blood pressure of the RASi users exactly normal, and thus equal to the blood pressure of the RASi non-users, no variation is left between users and non-users. Because of this, the correlation will be 0.

The linearity of correlation

An important limitation of the correlation coefficient is that it assumes a linear association. This also means that any linear transformation and any scale transformation of either variable X or Y , or both, will not affect the correlation coefficient. However, variables X and Y may also have a non-linear association, which could still yield a low correlation coefficient, as seen in Figure 1D and E , even though variables X and Y are clearly related. Nonetheless, the correlation coefficient will not always return 0 in case of a non-linear association, as portrayed in Figure 1F with an exponential correlation with r  = 0.5. In short, a correlation coefficient is not a measure of the best-fitted line through the observations, but only the degree to which the observations lie on one straight line.

In general, before calculating a correlation coefficient, it is advised to inspect a scatterplot of the observations in order to assess whether the data could possibly be described with a linear association and whether calculating a correlation coefficient makes sense. For instance, the scatterplot in Figure 1C could plausibly fit a straight line, and a correlation coefficient would therefore be suitable to describe the association in the data.

The range of observations for correlation

An important pitfall of the correlation coefficient is that it is influenced by the range of observations. In Figure 2A , we illustrate hypothetical data with 50 observations, with r  = 0.87. Included in the figure is an ellipse that shows the variance of the full observed data, and an ellipse that shows the variance of only the 25 lowest observations. If we subsequently analyse these 25 observations independently as shown in Figure 2B , we will see that the ellipse has shortened. If we determine the correlation coefficient for Figure 2B , we will also find a substantially lower correlation: r  = 0.57.

The effect of the range of observations on the correlation coefficient, as shown with ellipses. ( A ) Set of 50 observations from hypothetical dataset X with r  = 0.87, with an illustrative ellipse showing length and width of the whole dataset, and an ellipse showing only the first 25 observations. ( B ) Set of only the 25 lowest observations from hypothetical dataset X with r  = 0.57, with an illustrative ellipse showing length and width.

The importance of the range of observations can further be illustrated using an example from a paper by Pierrat et al. [ 6 ] in which the correlation between the eGFR calculated using inulin clearance and eGFR calculated using the Cockcroft–Gault formula was studied both in adults and children. Children had a higher correlation coefficient than adults ( r  = 0.81 versus r  = 0.67), after which the authors mentioned: ‘The coefficients of correlation were even better […] in children than in adults.’ However, the range of observations in children was larger than the range of observations in adults, which in itself could explain the higher correlation coefficient observed in children. One can thus not simply conclude that the Cockcroft–Gault formula for eGFR correlates better with inulin in children than in adults. Because the range of the correlation influences the correlation coefficient, it is important to realize that correlation coefficients cannot be readily compared between groups or studies. Another consequence of this is that researchers could inflate the correlation coefficient by including additional low and high eGFR values.

The non-causality of correlation

Another important pitfall of the correlation coefficient is that it cannot be interpreted as causal. It is of course possible that there is a causal effect of one variable on the other, but there may also be other possible explanations that the correlation coefficient does not take into account. Take for example the phenomenon of confounding. We can study the association of prescribing angiotensin-converting enzyme (ACE)-inhibitors with a decline in kidney function. These two variables would be highly correlated, which may be due to the underlying factor albuminuria. A patient with albuminuria is more likely to receive ACE-inhibitors, but is also more likely to have a decline in kidney function. So ACE-inhibitors and a decline in kidney function are correlated not because of ACE-inhibitors causing a decline in kidney function, but because they have a shared underlying cause (also known as common cause) [ 7 ]. More reasons why associations may be biased exist, which are explained elsewhere [ 8 , 9 ].

It is however possible to adjust for such confounding effects, for example by using multivariable regression. Whereas a univariable (or ‘crude’) linear regression analysis is no different than calculating the correlation coefficient, a multivariable regression analysis allows one to adjust for possible confounder variables. Other factors need to be taken into account to estimate causal effects, but these are beyond the scope of this paper.

Agreement between methods

We have discussed the correlation coefficient and its limitations when studying the association between two variables. However, the correlation coefficient is also often incorrectly used to study the agreement between two methods that aim to estimate the same variable. Again, also here, the correlation coefficient is an invalid measure.

The correlation coefficient aims to represent to what degree a straight line fits the data. This is not the same as agreement between methods (i.e. whether X  =  Y ). If methods completely agree, all observations would fall on the line of equality (i.e. the line on which the observations would be situated if X and Y had equal values). Yet the correlation coefficient looks at the best-fitted straight line through the data, which is not per se the line of equality. As a result, any method that would consistently measure a twice as large value as the other method would still correlate perfectly with the other method. This is shown in Figure 3 , where the dashed line shows the line of equality, and the other lines portray different linear associations, all with perfect correlation, but no agreement between X and Y . These linear associations may portray a systematic difference, better known as bias, in one of the methods.

A set of linear associations, with the dashed line (- - -) showing the line of equality where X  =  Y . The equations and correlations for the other lines are shown as well, which shows that only a linear association is needed for r  = 1, and not specifically agreement.

This limitation applies to all comparisons of methods, where it is studied whether methods can be used interchangeably, and it also applies to situations where two individuals measure a value and where the results are then compared (inter-observer variation or agreement; here the individuals can be seen as the ‘methods’), and to situations where it is studied whether one method measures consistently at two different time points (also known as repeatability). Fortunately, other methods exist to compare methods [ 10 , 11 ], of which one was proposed by Bland and Altman themselves [ 12 ].

Intraclass coefficient

One valid method to assess interchangeability is the intraclass coefficient (ICC), which is a generalization of Cohen’s κ , a measure for the assessment of intra- and interobserver agreement. The ICC shows the proportion of the variability in the new method that is due to the normal variability between individuals. The measure takes into account both the correlation and the systematic difference (i.e. bias), which makes it a measure of both the consistency and agreement of two methods. Nonetheless, like the correlation coefficient, it is influenced by the range of observations. However, an important advantage of the ICC is that it allows comparison between multiple variables or observers. Similar to the ICC is the concordance correlation coefficient (CCC), though it has been stated that the CCC yields values similar to the ICC [ 13 ]. Nonetheless, the CCC may also be found in the literature [ 14 ].

The 95% limits of agreement and the Bland–Altman plot

When they published their critique on the use of the correlation coefficient for the measurement of agreement, Bland and Altman also published an alternative method to measure agreement, which they called the limits of agreement (also referred to as a Bland–Altman plot) [ 12 ]. To illustrate the method of the limits of agreement, an artificial dataset was created using the MASS package (version 7.3-53) for R version 4.0.4 (R Corps, Vienna, Austria). Two sets of observations (two observations per person) were derived from a normal distribution with a mean ( µ ) of 120 and a randomly chosen standard deviation ( σ ) between 5 and 15. The mean of 120 was chosen with the aim to have the values resemble measurements of high eGFR, where the first set of observed eGFRs was hypothetically acquired using the MDRD formula, and the second set of observed eGFRs was hypothetically acquired using the CKD-EPI formula. The observations can be found in Table 1 .

Artificial data portraying hypothetically observed MDRD measurements and CKD-EPI measurements

The 95% limits of agreement can be easily calculated using the mean of the differences ( d ¯ ) and the standard deviation (SD) of the differences. The upper limit (UL) of the limits of agreement would then be UL = d ¯ + 1.96 * SD and the lower limit (LL) would be LL = d ¯ - 1.96 * SD . If we apply this to the data from Table 1 , we would find d ¯ = 0.32 and SD = 4.09. Subsequently, UL = 0.32 + 1.96 * 4.09 = 8.34 and LL = 0.32 − 1.96 * 4.09 = −7.70. Our limits of agreement are thus −7.70 to 8.34. We can now decide whether these limits of agreement are too broad. Imagine we decide that if we want to replace the MDRD formula with the CKD-EPI formula, we say that the difference may not be larger than 7 mL/min/1.73 m 2 . Thus, on the basis of these (hypothetical) data, the MDRD and CKD-EPI formulas cannot be used interchangeably in our case. It should also be noted that, as the limits of agreement are statistical parameters, they are also subject to uncertainty. The uncertainty can be determined by calculating 95% confidence intervals for the limits of agreement, on which Bland and Altman elaborate in their paper [ 12 ].

The limits of agreement are also subject to two assumptions: (i) the mean and SD of the differences should be constant over the range of observations and (ii) the differences are approximately normally distributed. To check these assumptions, two plots were proposed: the Bland–Altman plot, which is the differences plotted against the means of their measurements, and a histogram of the differences. If in the Bland–Altman plot the means and SDs of the differences appear to be equal along the x -axis, the first assumption is met. The histogram of the differences should follow the pattern of a normal distribution. We checked these assumptions by creating a Bland–Altman plot in Figure 4A and a histogram of the differences in Figure 4B . As often done, we also added the limits of agreement to the Bland–Altman plot, between which approximately 95% of datapoints are expected to be. In Figure 4A , we see that the mean of the differences appears to be equal along the x -axis; i.e., these datapoints could plausibly fit the horizontal line of the total mean across the whole x -axis. Nonetheless, the SD does not appear to be distributed equally: the means of the differences at the lower values of the x -axis are closer to the total mean (thus a lower SD) than the means of the differences at the middle values of the x -axis (thus a higher SD). Therefore, the first assumption is not met. Nonetheless, the second assumption is met, because our differences follow a normal distribution, as shown in Figure 4B . Our failure to meet the first assumption can be due to a number of reasons, for which Bland and Altman also proposed solutions [ 15 ]. For example, data may be skewed. However, in that case, log-transforming variables may be a solution [ 16 ].

Plots to check assumptions for the limits of agreement. ( A ) The Bland–Altman plot for the assumption that the mean and SD of the differences are constant over the range of observations. In our case, we see that the mean of the differences appears to be equal along the x -axis; i.e., these datapoints could plausibly fit the horizontal line of the total mean across the whole x -axis. Nonetheless, the SD does not appear to be distributed equally: the means of the differences at the lower values of the x -axis are closer to the total mean (thus a lower SD) than the means of the differences at the middle values of the x -axis (thus a higher SD). Therefore, the first assumption is not met. The limits of agreement and the mean are added as dashed (- - -) lines. ( B ) A histogram of the distribution of differences to ascertain the assumption of whether the differences are normally distributed. In our case, the observations follow a normal distribution and thus, the assumption is met.

It is often mistakenly thought that the Bland–Altman plot alone is the analysis to determine the agreement between methods, but the authors themselves spoke strongly against this [ 15 ]. We suggest that authors should both report the limits of agreement and show the Bland–Altman plot, to allow readers to assess for themselves whether they think the agreement is met.

The correlation coefficient is easy to calculate and provides a measure of the strength of linear association in the data. However, it also has important limitations and pitfalls, both when studying the association between two variables and when studying agreement between methods. These limitations and pitfalls should be taken into account when using and interpreting it. If necessary, researchers should look into alternatives to the correlation coefficient, such as regression analysis for causal research, and the ICC and the limits of agreement combined with a Bland–Altman plot when comparing methods.

CONFLICT OF INTEREST STATEMENT

None declared.

Contributor Information

Roemer J Janse, Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, The Netherlands.

Tiny Hoekstra, Department of Nephrology, Amsterdam Cardiovascular Sciences, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands.

Kitty J Jager, ERA-EDTA Registry, Department of Medical Informatics, Amsterdam Public Health Research Institute, Amsterdam UMC, University of Amsterdam, Amsterdam, The Netherlands.

Carmine Zoccali, CNR-IFC, Center of Clinical Physiology, Clinical Epidemiology of Renal Diseases and Hypertension, Reggio Calabria, Italy.

Giovanni Tripepi, CNR-IFC, Center of Clinical Physiology, Clinical Epidemiology of Renal Diseases and Hypertension, Reggio Calabria, Italy.

Friedo W Dekker, Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, The Netherlands.

Merel van Diepen, Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, The Netherlands.

6.2 Correlational Research

Learning objectives.

• Define correlational research and give several examples.
• Explain why a researcher might choose to conduct correlational research rather than experimental research or another type of non-experimental research.
• Interpret the strength and direction of different correlation coefficients.
• Explain why correlation does not imply causation.

What Is Correlational Research?

Correlational research is a type of non-experimental research in which the researcher measures two variables and assesses the statistical relationship (i.e., the correlation) between them with little or no effort to control extraneous variables. There are many reasons that researchers interested in statistical relationships between variables would choose to conduct a correlational study rather than an experiment. The first is that they do not believe that the statistical relationship is a causal one or are not interested in causal relationships. Recall two goals of science are to describe and to predict and the correlational research strategy allows researchers to achieve both of these goals. Specifically, this strategy can be used to describe the strength and direction of the relationship between two variables and if there is a relationship between the variables then the researchers can use scores on one variable to predict scores on the other (using a statistical technique called regression).

Another reason that researchers would choose to use a correlational study rather than an experiment is that the statistical relationship of interest is thought to be causal, but the researcher  cannot  manipulate the independent variable because it is impossible, impractical, or unethical. For example, while I might be interested in the relationship between the frequency people use cannabis and their memory abilities I cannot ethically manipulate the frequency that people use cannabis. As such, I must rely on the correlational research strategy; I must simply measure the frequency that people use cannabis and measure their memory abilities using a standardized test of memory and then determine whether the frequency people use cannabis use is statistically related to memory test performance. 

Correlation is also used to establish the reliability and validity of measurements. For example, a researcher might evaluate the validity of a brief extraversion test by administering it to a large group of participants along with a longer extraversion test that has already been shown to be valid. This researcher might then check to see whether participants’ scores on the brief test are strongly correlated with their scores on the longer one. Neither test score is thought to cause the other, so there is no independent variable to manipulate. In fact, the terms  independent variable  and dependent variabl e  do not apply to this kind of research.

Another strength of correlational research is that it is often higher in external validity than experimental research. Recall there is typically a trade-off between internal validity and external validity. As greater controls are added to experiments, internal validity is increased but often at the expense of external validity. In contrast, correlational studies typically have low internal validity because nothing is manipulated or control but they often have high external validity. Since nothing is manipulated or controlled by the experimenter the results are more likely to reflect relationships that exist in the real world.

Finally, extending upon this trade-off between internal and external validity, correlational research can help to provide converging evidence for a theory. If a theory is supported by a true experiment that is high in internal validity as well as by a correlational study that is high in external validity then the researchers can have more confidence in the validity of their theory. As a concrete example, correlational studies establishing that there is a relationship between watching violent television and aggressive behavior have been complemented by experimental studies confirming that the relationship is a causal one (Bushman & Huesmann, 2001) [1] .  These converging results provide strong evidence that there is a real relationship (indeed a causal relationship) between watching violent television and aggressive behavior.

Data Collection in Correlational Research

Again, the defining feature of correlational research is that neither variable is manipulated. It does not matter how or where the variables are measured. A researcher could have participants come to a laboratory to complete a computerized backward digit span task and a computerized risky decision-making task and then assess the relationship between participants’ scores on the two tasks. Or a researcher could go to a shopping mall to ask people about their attitudes toward the environment and their shopping habits and then assess the relationship between these two variables. Both of these studies would be correlational because no independent variable is manipulated. 

Correlations Between Quantitative Variables

Correlations between quantitative variables are often presented using scatterplots . Figure 6.3 shows some hypothetical data on the relationship between the amount of stress people are under and the number of physical symptoms they have. Each point in the scatterplot represents one person’s score on both variables. For example, the circled point in Figure 6.3 represents a person whose stress score was 10 and who had three physical symptoms. Taking all the points into account, one can see that people under more stress tend to have more physical symptoms. This is a good example of a positive relationship , in which higher scores on one variable tend to be associated with higher scores on the other. A  negative relationship  is one in which higher scores on one variable tend to be associated with lower scores on the other. There is a negative relationship between stress and immune system functioning, for example, because higher stress is associated with lower immune system functioning.

Figure 6.3 Scatterplot Showing a Hypothetical Positive Relationship Between Stress and Number of Physical Symptoms. The circled point represents a person whose stress score was 10 and who had three physical symptoms. Pearson’s r for these data is +.51.

The strength of a correlation between quantitative variables is typically measured using a statistic called  Pearson’s Correlation Coefficient (or Pearson’s  r ) . As Figure 6.4 shows, Pearson’s r ranges from −1.00 (the strongest possible negative relationship) to +1.00 (the strongest possible positive relationship). A value of 0 means there is no relationship between the two variables. When Pearson’s  r  is 0, the points on a scatterplot form a shapeless “cloud.” As its value moves toward −1.00 or +1.00, the points come closer and closer to falling on a single straight line. Correlation coefficients near ±.10 are considered small, values near ± .30 are considered medium, and values near ±.50 are considered large. Notice that the sign of Pearson’s  r  is unrelated to its strength. Pearson’s  r  values of +.30 and −.30, for example, are equally strong; it is just that one represents a moderate positive relationship and the other a moderate negative relationship. With the exception of reliability coefficients, most correlations that we find in Psychology are small or moderate in size. The website http://rpsychologist.com/d3/correlation/ , created by Kristoffer Magnusson, provides an excellent interactive visualization of correlations that permits you to adjust the strength and direction of a correlation while witnessing the corresponding changes to the scatterplot.

Figure 6.4 Range of Pearson’s r, From −1.00 (Strongest Possible Negative Relationship), Through 0 (No Relationship), to +1.00 (Strongest Possible Positive Relationship)

There are two common situations in which the value of Pearson’s  r  can be misleading. Pearson’s  r  is a good measure only for linear relationships, in which the points are best approximated by a straight line. It is not a good measure for nonlinear relationships, in which the points are better approximated by a curved line. Figure 6.5, for example, shows a hypothetical relationship between the amount of sleep people get per night and their level of depression. In this example, the line that best approximates the points is a curve—a kind of upside-down “U”—because people who get about eight hours of sleep tend to be the least depressed. Those who get too little sleep and those who get too much sleep tend to be more depressed. Even though Figure 6.5 shows a fairly strong relationship between depression and sleep, Pearson’s  r  would be close to zero because the points in the scatterplot are not well fit by a single straight line. This means that it is important to make a scatterplot and confirm that a relationship is approximately linear before using Pearson’s  r . Nonlinear relationships are fairly common in psychology, but measuring their strength is beyond the scope of this book.

Figure 6.5 Hypothetical Nonlinear Relationship Between Sleep and Depression

The other common situations in which the value of Pearson’s  r  can be misleading is when one or both of the variables have a limited range in the sample relative to the population. This problem is referred to as  restriction of range . Assume, for example, that there is a strong negative correlation between people’s age and their enjoyment of hip hop music as shown by the scatterplot in Figure 6.6. Pearson’s  r  here is −.77. However, if we were to collect data only from 18- to 24-year-olds—represented by the shaded area of Figure 6.6—then the relationship would seem to be quite weak. In fact, Pearson’s  r  for this restricted range of ages is 0. It is a good idea, therefore, to design studies to avoid restriction of range. For example, if age is one of your primary variables, then you can plan to collect data from people of a wide range of ages. Because restriction of range is not always anticipated or easily avoidable, however, it is good practice to examine your data for possible restriction of range and to interpret Pearson’s  r  in light of it. (There are also statistical methods to correct Pearson’s  r  for restriction of range, but they are beyond the scope of this book).

Figure 6.6 Hypothetical Data Showing How a Strong Overall Correlation Can Appear to Be Weak When One Variable Has a Restricted Range.The overall correlation here is −.77, but the correlation for the 18- to 24-year-olds (in the blue box) is 0.

Correlation Does Not Imply Causation

You have probably heard repeatedly that “Correlation does not imply causation.” An amusing example of this comes from a 2012 study that showed a positive correlation (Pearson’s r = 0.79) between the per capita chocolate consumption of a nation and the number of Nobel prizes awarded to citizens of that nation [2] . It seems clear, however, that this does not mean that eating chocolate causes people to win Nobel prizes, and it would not make sense to try to increase the number of Nobel prizes won by recommending that parents feed their children more chocolate.

There are two reasons that correlation does not imply causation. The first is called the  directionality problem . Two variables,  X  and  Y , can be statistically related because X  causes  Y  or because  Y  causes  X . Consider, for example, a study showing that whether or not people exercise is statistically related to how happy they are—such that people who exercise are happier on average than people who do not. This statistical relationship is consistent with the idea that exercising causes happiness, but it is also consistent with the idea that happiness causes exercise. Perhaps being happy gives people more energy or leads them to seek opportunities to socialize with others by going to the gym. The second reason that correlation does not imply causation is called the  third-variable problem . Two variables,  X  and  Y , can be statistically related not because  X  causes  Y , or because  Y  causes  X , but because some third variable,  Z , causes both  X  and  Y . For example, the fact that nations that have won more Nobel prizes tend to have higher chocolate consumption probably reflects geography in that European countries tend to have higher rates of per capita chocolate consumption and invest more in education and technology (once again, per capita) than many other countries in the world. Similarly, the statistical relationship between exercise and happiness could mean that some third variable, such as physical health, causes both of the others. Being physically healthy could cause people to exercise and cause them to be happier. Correlations that are a result of a third-variable are often referred to as  spurious correlations.

Some excellent and funny examples of spurious correlations can be found at http://www.tylervigen.com  (Figure 6.7  provides one such example).

“Lots of Candy Could Lead to Violence”

Although researchers in psychology know that correlation does not imply causation, many journalists do not. One website about correlation and causation, http://jonathan.mueller.faculty.noctrl.edu/100/correlation_or_causation.htm , links to dozens of media reports about real biomedical and psychological research. Many of the headlines suggest that a causal relationship has been demonstrated when a careful reading of the articles shows that it has not because of the directionality and third-variable problems.

One such article is about a study showing that children who ate candy every day were more likely than other children to be arrested for a violent offense later in life. But could candy really “lead to” violence, as the headline suggests? What alternative explanations can you think of for this statistical relationship? How could the headline be rewritten so that it is not misleading?

As you have learned by reading this book, there are various ways that researchers address the directionality and third-variable problems. The most effective is to conduct an experiment. For example, instead of simply measuring how much people exercise, a researcher could bring people into a laboratory and randomly assign half of them to run on a treadmill for 15 minutes and the rest to sit on a couch for 15 minutes. Although this seems like a minor change to the research design, it is extremely important. Now if the exercisers end up in more positive moods than those who did not exercise, it cannot be because their moods affected how much they exercised (because it was the researcher who determined how much they exercised). Likewise, it cannot be because some third variable (e.g., physical health) affected both how much they exercised and what mood they were in (because, again, it was the researcher who determined how much they exercised). Thus experiments eliminate the directionality and third-variable problems and allow researchers to draw firm conclusions about causal relationships.

Key Takeaways

• Correlational research involves measuring two variables and assessing the relationship between them, with no manipulation of an independent variable.
• Correlation does not imply causation. A statistical relationship between two variables,  X  and  Y , does not necessarily mean that  X  causes  Y . It is also possible that  Y  causes  X , or that a third variable,  Z , causes both  X  and  Y .
• While correlational research cannot be used to establish causal relationships between variables, correlational research does allow researchers to achieve many other important objectives (establishing reliability and validity, providing converging evidence, describing relationships and making predictions)
• Correlation coefficients can range from -1 to +1. The sign indicates the direction of the relationship between the variables and the numerical value indicates the strength of the relationship.
• A cognitive psychologist compares the ability of people to recall words that they were instructed to “read” with their ability to recall words that they were instructed to “imagine.”
• A manager studies the correlation between new employees’ college grade point averages and their first-year performance reports.
• An automotive engineer installs different stick shifts in a new car prototype, each time asking several people to rate how comfortable the stick shift feels.
• A food scientist studies the relationship between the temperature inside people’s refrigerators and the amount of bacteria on their food.
• A social psychologist tells some research participants that they need to hurry over to the next building to complete a study. She tells others that they can take their time. Then she observes whether they stop to help a research assistant who is pretending to be hurt.

2. Practice: For each of the following statistical relationships, decide whether the directionality problem is present and think of at least one plausible third variable.

• People who eat more lobster tend to live longer.
• People who exercise more tend to weigh less.
• College students who drink more alcohol tend to have poorer grades.
• Bushman, B. J., & Huesmann, L. R. (2001). Effects of televised violence on aggression. In D. Singer & J. Singer (Eds.), Handbook of children and the media (pp. 223–254). Thousand Oaks, CA: Sage. ↵
• Messerli, F. H. (2012). Chocolate consumption, cognitive function, and Nobel laureates. New England Journal of Medicine, 367 , 1562-1564. ↵

Share This Book

• Increase Font Size

• Bipolar Disorder
• Therapy Center
• When To See a Therapist
• Types of Therapy
• Best Online Therapy
• Best Couples Therapy
• Best Family Therapy
• Managing Stress
• Sleep and Dreaming
• Understanding Emotions
• Self-Improvement
• Healthy Relationships
• Student Resources
• Personality Types
• Verywell Mind Insights
• 2023 Verywell Mind 25
• Mental Health in the Classroom
• Editorial Process
• Meet Our Review Board
• Crisis Support

Correlation Studies in Psychology Research

Determining the relationship between two or more variables.

Kendra Cherry, MS, is a psychosocial rehabilitation specialist, psychology educator, and author of the "Everything Psychology Book."

Emily is a board-certified science editor who has worked with top digital publishing brands like Voices for Biodiversity, Study.com, GoodTherapy, Vox, and Verywell.

Verywell / Brianna Gilmartin

• Characteristics

Potential Pitfalls

Frequently asked questions.

A correlational study is a type of research design that looks at the relationships between two or more variables. Correlational studies are non-experimental, which means that the experimenter does not manipulate or control any of the variables.

A correlation refers to a relationship between two variables. Correlations can be strong or weak and positive or negative. Sometimes, there is no correlation.

There are three possible outcomes of a correlation study: a positive correlation, a negative correlation, or no correlation. Researchers can present the results using a numerical value called the correlation coefficient, a measure of the correlation strength. It can range from –1.00 (negative) to +1.00 (positive). A correlation coefficient of 0 indicates no correlation.

• Positive correlations : Both variables increase or decrease at the same time. A correlation coefficient close to +1.00 indicates a strong positive correlation.
• Negative correlations : As the amount of one variable increases, the other decreases (and vice versa). A correlation coefficient close to -1.00 indicates a strong negative correlation.
• No correlation : There is no relationship between the two variables. A correlation coefficient of 0 indicates no correlation.

Characteristics of a Correlational Study

Correlational studies are often used in psychology, as well as other fields like medicine. Correlational research is a preliminary way to gather information about a topic. The method is also useful if researchers are unable to perform an experiment.

Researchers use correlations to see if a relationship between two or more variables exists, but the variables themselves are not under the control of the researchers.

While correlational research can demonstrate a relationship between variables, it cannot prove that changing one variable will change another. In other words, correlational studies cannot prove cause-and-effect relationships.

When you encounter research that refers to a "link" or an "association" between two things, they are most likely talking about a correlational study.

Types of Correlational Research

There are three types of correlational research: naturalistic observation, the survey method, and archival research. Each type has its own purpose, as well as its pros and cons.

Naturalistic Observation

The naturalistic observation method involves observing and recording variables of interest in a natural setting without interference or manipulation.  

Can inspire ideas for further research

Option if lab experiment not available

Variables are viewed in natural setting

Can be time-consuming and expensive

Extraneous variables can't be controlled

No scientific control of variables

Subjects might behave differently if aware of being observed

This method is well-suited to studies where researchers want to see how variables behave in their natural setting or state.   Inspiration can then be drawn from the observations to inform future avenues of research.

In some cases, it might be the only method available to researchers; for example, if lab experimentation would be precluded by access, resources, or ethics. It might be preferable to not being able to conduct research at all, but the method can be costly and usually takes a lot of time.  

Naturalistic observation presents several challenges for researchers. For one, it does not allow them to control or influence the variables in any way nor can they change any possible external variables.

However, this does not mean that researchers will get reliable data from watching the variables, or that the information they gather will be free from bias.

For example, study subjects might act differently if they know that they are being watched. The researchers might not be aware that the behavior that they are observing is not necessarily the subject's natural state (i.e., how they would act if they did not know they were being watched).

Researchers also need to be aware of their biases, which can affect the observation and interpretation of a subject's behavior.  

Surveys and questionnaires are some of the most common methods used for psychological research. The survey method involves having a  random sample  of participants complete a survey, test, or questionnaire related to the variables of interest.   Random sampling is vital to the generalizability of a survey's results.

Cheap, easy, and fast

Can collect large amounts of data in a short amount of time

Results can be affected by poor survey questions

Results can be affected by unrepresentative sample

Outcomes can be affected by participants

If researchers need to gather a large amount of data in a short period of time, a survey is likely to be the fastest, easiest, and cheapest option.  

It's also a flexible method because it lets researchers create data-gathering tools that will help ensure they get the information they need (survey responses) from all the sources they want to use (a random sample of participants taking the survey).

Survey data might be cost-efficient and easy to get, but it has its downsides. For one, the data is not always reliable—particularly if the survey questions are poorly written or the overall design or delivery is weak.   Data is also affected by specific faults, such as unrepresented or underrepresented samples .

The use of surveys relies on participants to provide useful data. Researchers need to be aware of the specific factors related to the people taking the survey that will affect its outcome.

For example, some people might struggle to understand the questions. A person might answer a particular way to try to please the researchers or to try to control how the researchers perceive them (such as trying to make themselves "look better").

Sometimes, respondents might not even realize that their answers are incorrect or misleading because of mistaken memories .

Archival Research

Many areas of psychological research benefit from analyzing studies that were conducted long ago by other researchers, as well as reviewing historical records and case studies.

For example, in an experiment known as  "The Irritable Heart ," researchers used digitalized records containing information on American Civil War veterans to learn more about post-traumatic stress disorder (PTSD).

Large amount of data

Can be less expensive

Researchers cannot change participant behavior

Can be unreliable

Information might be missing

No control over data collection methods

Using records, databases, and libraries that are publicly accessible or accessible through their institution can help researchers who might not have a lot of money to support their research efforts.

Free and low-cost resources are available to researchers at all levels through academic institutions, museums, and data repositories around the world.

Another potential benefit is that these sources often provide an enormous amount of data that was collected over a very long period of time, which can give researchers a way to view trends, relationships, and outcomes related to their research.

While the inability to change variables can be a disadvantage of some methods, it can be a benefit of archival research. That said, using historical records or information that was collected a long time ago also presents challenges. For one, important information might be missing or incomplete and some aspects of older studies might not be useful to researchers in a modern context.

A primary issue with archival research is reliability. When reviewing old research, little information might be available about who conducted the research, how a study was designed, who participated in the research, as well as how data was collected and interpreted.

Researchers can also be presented with ethical quandaries—for example, should modern researchers use data from studies that were conducted unethically or with questionable ethics?

You've probably heard the phrase, "correlation does not equal causation." This means that while correlational research can suggest that there is a relationship between two variables, it cannot prove that one variable will change another.

For example, researchers might perform a correlational study that suggests there is a relationship between academic success and a person's self-esteem. However, the study cannot show that academic success changes a person's self-esteem.

To determine why the relationship exists, researchers would need to consider and experiment with other variables, such as the subject's social relationships, cognitive abilities, personality, and socioeconomic status.

The difference between a correlational study and an experimental study involves the manipulation of variables. Researchers do not manipulate variables in a correlational study, but they do control and systematically vary the independent variables in an experimental study. Correlational studies allow researchers to detect the presence and strength of a relationship between variables, while experimental studies allow researchers to look for cause and effect relationships.

If the study involves the systematic manipulation of the levels of a variable, it is an experimental study. If researchers are measuring what is already present without actually changing the variables, then is a correlational study.

The variables in a correlational study are what the researcher measures. Once measured, researchers can then use statistical analysis to determine the existence, strength, and direction of the relationship. However, while correlational studies can say that variable X and variable Y have a relationship, it does not mean that X causes Y.

The goal of correlational research is often to look for relationships, describe these relationships, and then make predictions. Such research can also often serve as a jumping off point for future experimental research. 

Heath W. Psychology Research Methods . Cambridge University Press; 2018:134-156.

Schneider FW. Applied Social Psychology . 2nd ed. SAGE; 2012:50-53.

Curtis EA, Comiskey C, Dempsey O. Importance and use of correlational research .  Nurse Researcher . 2016;23(6):20-25. doi:10.7748/nr.2016.e1382

Carpenter S. Visualizing Psychology . 3rd ed. John Wiley & Sons; 2012:14-30.

Pizarro J, Silver RC, Prause J. Physical and mental health costs of traumatic war experiences among civil war veterans .  Arch Gen Psychiatry . 2006;63(2):193. doi:10.1001/archpsyc.63.2.193

Post SG. The echo of Nuremberg: Nazi data and ethics .  J Med Ethics . 1991;17(1):42-44. doi:10.1136/jme.17.1.42

Lau F. Chapter 12 Methods for Correlational Studies . In: Lau F, Kuziemsky C, eds. Handbook of eHealth Evaluation: An Evidence-based Approach . University of Victoria.

Akoglu H. User's guide to correlation coefficients .  Turk J Emerg Med . 2018;18(3):91-93. doi:10.1016/j.tjem.2018.08.001

Price PC. Research Methods in Psychology . California State University.

By Kendra Cherry, MSEd Kendra Cherry, MS, is a psychosocial rehabilitation specialist, psychology educator, and author of the "Everything Psychology Book."

Want to create or adapt books like this? Learn more about how Pressbooks supports open publishing practices.

7.2 Correlational Research

Learning objectives.

• Define correlational research and give several examples.
• Explain why a researcher might choose to conduct correlational research rather than experimental research or another type of nonexperimental research.

What Is Correlational Research?

Correlational research is a type of nonexperimental research in which the researcher measures two variables and assesses the statistical relationship (i.e., the correlation) between them with little or no effort to control extraneous variables. There are essentially two reasons that researchers interested in statistical relationships between variables would choose to conduct a correlational study rather than an experiment. The first is that they do not believe that the statistical relationship is a causal one. For example, a researcher might evaluate the validity of a brief extraversion test by administering it to a large group of participants along with a longer extraversion test that has already been shown to be valid. This researcher might then check to see whether participants’ scores on the brief test are strongly correlated with their scores on the longer one. Neither test score is thought to cause the other, so there is no independent variable to manipulate. In fact, the terms independent variable and dependent variable do not apply to this kind of research.

The other reason that researchers would choose to use a correlational study rather than an experiment is that the statistical relationship of interest is thought to be causal, but the researcher cannot manipulate the independent variable because it is impossible, impractical, or unethical. For example, Allen Kanner and his colleagues thought that the number of “daily hassles” (e.g., rude salespeople, heavy traffic) that people experience affects the number of physical and psychological symptoms they have (Kanner, Coyne, Schaefer, & Lazarus, 1981). But because they could not manipulate the number of daily hassles their participants experienced, they had to settle for measuring the number of daily hassles—along with the number of symptoms—using self-report questionnaires. Although the strong positive relationship they found between these two variables is consistent with their idea that hassles cause symptoms, it is also consistent with the idea that symptoms cause hassles or that some third variable (e.g., neuroticism) causes both.

A common misconception among beginning researchers is that correlational research must involve two quantitative variables, such as scores on two extraversion tests or the number of hassles and number of symptoms people have experienced. However, the defining feature of correlational research is that the two variables are measured—neither one is manipulated—and this is true regardless of whether the variables are quantitative or categorical. Imagine, for example, that a researcher administers the Rosenberg Self-Esteem Scale to 50 American college students and 50 Japanese college students. Although this “feels” like a between-subjects experiment, it is a correlational study because the researcher did not manipulate the students’ nationalities. The same is true of the study by Cacioppo and Petty comparing college faculty and factory workers in terms of their need for cognition. It is a correlational study because the researchers did not manipulate the participants’ occupations.

Figure 7.2 “Results of a Hypothetical Study on Whether People Who Make Daily To-Do Lists Experience Less Stress Than People Who Do Not Make Such Lists” shows data from a hypothetical study on the relationship between whether people make a daily list of things to do (a “to-do list”) and stress. Notice that it is unclear whether this is an experiment or a correlational study because it is unclear whether the independent variable was manipulated. If the researcher randomly assigned some participants to make daily to-do lists and others not to, then it is an experiment. If the researcher simply asked participants whether they made daily to-do lists, then it is a correlational study. The distinction is important because if the study was an experiment, then it could be concluded that making the daily to-do lists reduced participants’ stress. But if it was a correlational study, it could only be concluded that these variables are statistically related. Perhaps being stressed has a negative effect on people’s ability to plan ahead (the directionality problem). Or perhaps people who are more conscientious are more likely to make to-do lists and less likely to be stressed (the third-variable problem). The crucial point is that what defines a study as experimental or correlational is not the variables being studied, nor whether the variables are quantitative or categorical, nor the type of graph or statistics used to analyze the data. It is how the study is conducted.

Figure 7.2 Results of a Hypothetical Study on Whether People Who Make Daily To-Do Lists Experience Less Stress Than People Who Do Not Make Such Lists

Data Collection in Correlational Research

Again, the defining feature of correlational research is that neither variable is manipulated. It does not matter how or where the variables are measured. A researcher could have participants come to a laboratory to complete a computerized backward digit span task and a computerized risky decision-making task and then assess the relationship between participants’ scores on the two tasks. Or a researcher could go to a shopping mall to ask people about their attitudes toward the environment and their shopping habits and then assess the relationship between these two variables. Both of these studies would be correlational because no independent variable is manipulated. However, because some approaches to data collection are strongly associated with correlational research, it makes sense to discuss them here. The two we will focus on are naturalistic observation and archival data. A third, survey research, is discussed in its own chapter.

Naturalistic Observation

Naturalistic observation is an approach to data collection that involves observing people’s behavior in the environment in which it typically occurs. Thus naturalistic observation is a type of field research (as opposed to a type of laboratory research). It could involve observing shoppers in a grocery store, children on a school playground, or psychiatric inpatients in their wards. Researchers engaged in naturalistic observation usually make their observations as unobtrusively as possible so that participants are often not aware that they are being studied. Ethically, this is considered to be acceptable if the participants remain anonymous and the behavior occurs in a public setting where people would not normally have an expectation of privacy. Grocery shoppers putting items into their shopping carts, for example, are engaged in public behavior that is easily observable by store employees and other shoppers. For this reason, most researchers would consider it ethically acceptable to observe them for a study. On the other hand, one of the arguments against the ethicality of the naturalistic observation of “bathroom behavior” discussed earlier in the book is that people have a reasonable expectation of privacy even in a public restroom and that this expectation was violated.

Researchers Robert Levine and Ara Norenzayan used naturalistic observation to study differences in the “pace of life” across countries (Levine & Norenzayan, 1999). One of their measures involved observing pedestrians in a large city to see how long it took them to walk 60 feet. They found that people in some countries walked reliably faster than people in other countries. For example, people in the United States and Japan covered 60 feet in about 12 seconds on average, while people in Brazil and Romania took close to 17 seconds.

Because naturalistic observation takes place in the complex and even chaotic “real world,” there are two closely related issues that researchers must deal with before collecting data. The first is sampling. When, where, and under what conditions will the observations be made, and who exactly will be observed? Levine and Norenzayan described their sampling process as follows:

Male and female walking speed over a distance of 60 feet was measured in at least two locations in main downtown areas in each city. Measurements were taken during main business hours on clear summer days. All locations were flat, unobstructed, had broad sidewalks, and were sufficiently uncrowded to allow pedestrians to move at potentially maximum speeds. To control for the effects of socializing, only pedestrians walking alone were used. Children, individuals with obvious physical handicaps, and window-shoppers were not timed. Thirty-five men and 35 women were timed in most cities. (p. 186)

Precise specification of the sampling process in this way makes data collection manageable for the observers, and it also provides some control over important extraneous variables. For example, by making their observations on clear summer days in all countries, Levine and Norenzayan controlled for effects of the weather on people’s walking speeds.

The second issue is measurement. What specific behaviors will be observed? In Levine and Norenzayan’s study, measurement was relatively straightforward. They simply measured out a 60-foot distance along a city sidewalk and then used a stopwatch to time participants as they walked over that distance. Often, however, the behaviors of interest are not so obvious or objective. For example, researchers Robert Kraut and Robert Johnston wanted to study bowlers’ reactions to their shots, both when they were facing the pins and then when they turned toward their companions (Kraut & Johnston, 1979). But what “reactions” should they observe? Based on previous research and their own pilot testing, Kraut and Johnston created a list of reactions that included “closed smile,” “open smile,” “laugh,” “neutral face,” “look down,” “look away,” and “face cover” (covering one’s face with one’s hands). The observers committed this list to memory and then practiced by coding the reactions of bowlers who had been videotaped. During the actual study, the observers spoke into an audio recorder, describing the reactions they observed. Among the most interesting results of this study was that bowlers rarely smiled while they still faced the pins. They were much more likely to smile after they turned toward their companions, suggesting that smiling is not purely an expression of happiness but also a form of social communication.

Naturalistic observation has revealed that bowlers tend to smile when they turn away from the pins and toward their companions, suggesting that smiling is not purely an expression of happiness but also a form of social communication.

sieneke toering – bowling big lebowski style – CC BY-NC-ND 2.0.

When the observations require a judgment on the part of the observers—as in Kraut and Johnston’s study—this process is often described as coding . Coding generally requires clearly defining a set of target behaviors. The observers then categorize participants individually in terms of which behavior they have engaged in and the number of times they engaged in each behavior. The observers might even record the duration of each behavior. The target behaviors must be defined in such a way that different observers code them in the same way. This is the issue of interrater reliability. Researchers are expected to demonstrate the interrater reliability of their coding procedure by having multiple raters code the same behaviors independently and then showing that the different observers are in close agreement. Kraut and Johnston, for example, video recorded a subset of their participants’ reactions and had two observers independently code them. The two observers showed that they agreed on the reactions that were exhibited 97% of the time, indicating good interrater reliability.

Archival Data

Another approach to correlational research is the use of archival data , which are data that have already been collected for some other purpose. An example is a study by Brett Pelham and his colleagues on “implicit egotism”—the tendency for people to prefer people, places, and things that are similar to themselves (Pelham, Carvallo, & Jones, 2005). In one study, they examined Social Security records to show that women with the names Virginia, Georgia, Louise, and Florence were especially likely to have moved to the states of Virginia, Georgia, Louisiana, and Florida, respectively.

As with naturalistic observation, measurement can be more or less straightforward when working with archival data. For example, counting the number of people named Virginia who live in various states based on Social Security records is relatively straightforward. But consider a study by Christopher Peterson and his colleagues on the relationship between optimism and health using data that had been collected many years before for a study on adult development (Peterson, Seligman, & Vaillant, 1988). In the 1940s, healthy male college students had completed an open-ended questionnaire about difficult wartime experiences. In the late 1980s, Peterson and his colleagues reviewed the men’s questionnaire responses to obtain a measure of explanatory style—their habitual ways of explaining bad events that happen to them. More pessimistic people tend to blame themselves and expect long-term negative consequences that affect many aspects of their lives, while more optimistic people tend to blame outside forces and expect limited negative consequences. To obtain a measure of explanatory style for each participant, the researchers used a procedure in which all negative events mentioned in the questionnaire responses, and any causal explanations for them, were identified and written on index cards. These were given to a separate group of raters who rated each explanation in terms of three separate dimensions of optimism-pessimism. These ratings were then averaged to produce an explanatory style score for each participant. The researchers then assessed the statistical relationship between the men’s explanatory style as college students and archival measures of their health at approximately 60 years of age. The primary result was that the more optimistic the men were as college students, the healthier they were as older men. Pearson’s r was +.25.

This is an example of content analysis —a family of systematic approaches to measurement using complex archival data. Just as naturalistic observation requires specifying the behaviors of interest and then noting them as they occur, content analysis requires specifying keywords, phrases, or ideas and then finding all occurrences of them in the data. These occurrences can then be counted, timed (e.g., the amount of time devoted to entertainment topics on the nightly news show), or analyzed in a variety of other ways.

Key Takeaways

• Correlational research involves measuring two variables and assessing the relationship between them, with no manipulation of an independent variable.
• Correlational research is not defined by where or how the data are collected. However, some approaches to data collection are strongly associated with correlational research. These include naturalistic observation (in which researchers observe people’s behavior in the context in which it normally occurs) and the use of archival data that were already collected for some other purpose.

Discussion: For each of the following, decide whether it is most likely that the study described is experimental or correlational and explain why.

• An educational researcher compares the academic performance of students from the “rich” side of town with that of students from the “poor” side of town.
• A cognitive psychologist compares the ability of people to recall words that they were instructed to “read” with their ability to recall words that they were instructed to “imagine.”
• A manager studies the correlation between new employees’ college grade point averages and their first-year performance reports.
• An automotive engineer installs different stick shifts in a new car prototype, each time asking several people to rate how comfortable the stick shift feels.
• A food scientist studies the relationship between the temperature inside people’s refrigerators and the amount of bacteria on their food.
• A social psychologist tells some research participants that they need to hurry over to the next building to complete a study. She tells others that they can take their time. Then she observes whether they stop to help a research assistant who is pretending to be hurt.

Kanner, A. D., Coyne, J. C., Schaefer, C., & Lazarus, R. S. (1981). Comparison of two modes of stress measurement: Daily hassles and uplifts versus major life events. Journal of Behavioral Medicine, 4 , 1–39.

Kraut, R. E., & Johnston, R. E. (1979). Social and emotional messages of smiling: An ethological approach. Journal of Personality and Social Psychology, 37 , 1539–1553.

Levine, R. V., & Norenzayan, A. (1999). The pace of life in 31 countries. Journal of Cross-Cultural Psychology, 30 , 178–205.

Pelham, B. W., Carvallo, M., & Jones, J. T. (2005). Implicit egotism. Current Directions in Psychological Science, 14 , 106–110.

Peterson, C., Seligman, M. E. P., & Vaillant, G. E. (1988). Pessimistic explanatory style is a risk factor for physical illness: A thirty-five year longitudinal study. Journal of Personality and Social Psychology, 55 , 23–27.

Research Methods in Psychology Copyright © 2016 by University of Minnesota is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License , except where otherwise noted.

• Cancer Nursing Practice
• Emergency Nurse
• Evidence-Based Nursing
• Learning Disability Practice
• Mental Health Practice
• Nurse Researcher
• Nursing Children and Young People
• Nursing Management
• Nursing Older People
• Nursing Standard
• Primary Health Care
• RCN Nursing Awards
• Nursing Live
• Nursing Careers and Job Fairs
• CPD webinars on-demand
• --> Advanced -->

• Clinical articles
• Expert advice
• Career advice
• Revalidation

Evidence & Practice Previous     Next

Importance and use of correlational research, elizabeth a curtis assistant professor, school of nursing and midwifery, trinity college dublin, dublin, republic of ireland, catherine comiskey head of school and professor of healthcare statistics, school of nursing and midwifery, trinity college dublin, dublin, republic of ireland, orla dempsey adjunct quantitative healthcare lead, school of nursing and midwifery, trinity college dublin, dublin, republic of ireland.

Background The importance of correlational research has been reported in the literature yet few research texts discuss design in any detail.

Aim To discuss important issues and considerations in correlational research, and suggest ways to avert potential problems during the preparation and application of the design.

Discussion: This article targets the gap identified in the literature regarding correlational research design. Specifically, it discusses the importance and purpose of correlational research, its application, analysis and interpretation with contextualisations to nursing and health research.

Conclusion: Findings from correlational research can be used to determine prevalence and relationships among variables, and to forecast events from current data and knowledge. In spite of its many uses, prudence is required when using the methodology and analysing data. To assist researchers in reducing mistakes, important issues are singled out for discussion and several options put forward for analysing data.

Implications for practice Correlational research is widely used and this paper should be particularly useful for novice nurse researchers. Furthermore, findings generated from correlational research can be used, for example, to inform decision-making, and to improve or initiate health-related activities or change.

Nurse Researcher . 23, 6, 20-25. doi: 10.7748/nr.2016.e1382

[email protected]

This article has been subject to double-blind review and has been checked using antiplagiarism software

None declared

Received: 16 March 2015

Accepted: 11 November 2015

nurses - nursing research - correlational research - correlation - variables - quantitative - measurement tools - data analysis

User not found

Want to read more?

Already have access log in, 3-month trial offer for £5.25/month.

• Unlimited access to all 10 RCNi Journals
• RCNi Learning featuring over 175 modules to easily earn CPD time
• NMC-compliant RCNi Revalidation Portfolio to stay on track with your progress
• Personalised newsletters tailored to your interests
• A customisable dashboard with over 200 topics

Alternatively, you can purchase access to this article for the next seven days. Buy now

Are you a student? Our student subscription has content especially for you. Find out more

18 July 2016 / Vol 23 issue 6

TABLE OF CONTENTS

DIGITAL EDITION

• LATEST ISSUE
• SIGN UP FOR E-ALERT
• WRITE FOR US
• PERMISSIONS

Share article: Importance and use of correlational research

We use cookies on this site to enhance your user experience.

By clicking any link on this page you are giving your consent for us to set cookies.

• Open access
• Published: 01 February 2023

Editorial: Best practices for correlational research in CAPMH

• Gerrit I. van Schalkwyk 1  

Child and Adolescent Psychiatry and Mental Health volume  17 , Article number:  18 ( 2023 ) Cite this article

1975 Accesses

1 Citations

8 Altmetric

Metrics details

The majority of manuscripts submitted to Child and Adolescent Psychiatry and Mental Health describe relationships between observed variables, either at a moment in time or longitudinally. This type of work has the potential to enhance our understanding of how specific exposures may increase risks of a certain outcome, as well as to understand variables that may mediate or moderate such relationships. Given the large number of submissions that employ this approach, the editorial team must increasingly apply a high standard during initial review. In order to provide authors with a more transparent review experience, this editorial will outline some considerations for best practices when submitting such work to our journal. These considerations will be illustrated using a case example.

Research needs to be hypothesis driven: it is common for manuscripts to draw on data from large datasets, and then describe statistically significant correlations between variables without a specific reason to have interrogated these variables. This represents a major limitation that will significantly reduce our interest in the submission. When authors describe a specific, well-substantiated hypothesis that guided their enquiry, the work has greater scientific validity, and will likely be of greater interest to our readers—even if the findings are statistically non-significant. For example:

“There is considerable interest in whether early screen exposure impacts executive function. To test this hypothesis, we administered a validated measure of screen time combined with a measure of executive function to a sample of 250 youth. Our findings revealed no significant relationship between screen time and executive function”

Would represent a more valid approach than:

“In a sample of 250 youth, data was collected on anxiety levels, parent education, trauma exposure, and screen time. We found a statistically significant correlation between paternal high school completion and screen time exposure”

The work needs to occur in the context of a theoretical framework. Beyond stating a hypothesis, high quality correlational research should be situated within existing theory, and should attempt to advance our existing understanding by asking questions that exist at the edge of our current knowledge. This will increase the impact of the work relative to research that situates its hypothesis in an arbitrary way that is entirely disconnect from the state of the field. As an example of a best practice:

“Two main theories exist regarding how screen time impacts cognitive function. The first argues for the impact screen time has on ‘displacing’ other activities that are valuable to normal development. The second argues that screen time has a direct negative impact that is independent of displacement. The present study therefore sought to measure both screen time and time spent on other activities, such that our results could be interpreted in the context of these competing theories.”

By contrast, a disconnected approach could be:

“Much is unknown about how screen time impacts cognitive function. One area that is unexplored is the impact of nutrition. We therefore decided to study nutritional status and screen time and how it correlates to cognitive function”

Less is more. There is a temptation when conducting correlational research to measure a large number of variables in order to increase the odds of finding a statistically significant signal. However, this approach makes the research inherently less hypothesis driven, and suggests that the researchers are less focused on answering a question to advance our understanding than they are on finding something that will achieve some measure of significance. When a large number of variables are tested, there are likely to be correlations identified by chance which do not represent true relationships, and may obscure rather than advance our understanding of a phenomenon. In order to reinforce more disciplined approaches to correlational research, our editorial board commits to considering negative studies that are rigorous in their design as superior to positive studies that came to significance by testing too many relationships.

Tell the entire story. No research is perfect, and good findings do not always come about in the context of perfectly designed studies. It is preferable that authors present an accurate and honest account of how their work was conducted, rather than to attempt to provide a tidier narrative that is not reflective of the reality of what was done. Most often, this problem will occur when a hypothesis was developed after the fact to ‘fit’ a significant finding, or when the initial hypothesis was changed for the same purpose:

“Given the interest of the authors in the impacts of paternal education level, we sought to test the hypothesis that paternal education was correlated with excess screen time in a sample of youth”

A more honest presentation could be:

“We tested the hypothesis that socio-economic status would be correlated with increased screen time in a sample of youth, but this finding proved to not be significant in our sample. Of interest for further work, a secondary analysis did identify a strong relationship between paternal education and screen time exposure, which may relate to an emerging theory on how education level impacts aspects of child supervision. A future study could aim to develop and test a specific hypothesis around this possible correlation.”

Make the impetus clear. Particularly for a clinically-focused journal like CAPMH , it is of importance that correlational work be driven by a specific clinical or theoretical problem that it is attempting to address. A significant number of manuscripts justify the topic under investigation by a vague statement of needing to ‘better understand the factors’ that drive or influence an outcome, and then conclude that ‘further research is required’, or that more resources be dedicated to addressing which ever risk factor or mediator was identified. By contrast, a more robust approach is to articulate why a specific knowledge gap is salient and important, and how the present study has the potential to provide a real improvement in our understanding that can guide specific decisions about treatment, diagnosis, resource allocation or policy. A less desirable but common approach could be described as follows:

“Given the importance of cognitive function, it is important to understand factors by which it may be impacted. Given our finding that cognitive function is lower in youth who had one parent that did not complete high school, more resources should be given to screen these youth for cognitive problems’.

A contrasting and superior approach would be:

“The lack of understanding on whether screen time impacts cognitive function a) directly, or, b) by limiting other important activities, is a critical limitation in our ability to issue guidance to patients and families. Our study found that increased screen time was only correlated with reduced cognitive function in youth who also had decreased engagement in other daily activities. This supports an emerging theory that the impacts of screen time on cognitive function can be overcome by ensuring youth are also able to spend enough time on other stimulating tasks, which has both clinical and policy implications.”

We recognize that every study is unique, and will continue to employ a review process that is individualized, and balances the strengths and weaknesses of each submission. This editorial should not be viewed as a checklist that will either guarantee or prevent submission, but a set of considerations that we hope will aid authors in the development of their work, and provide clarity on how the editorial board might evaluate a submission.

Availability of data and materials

Acknowledgements.

The author would like to acknowledge Andres Martin, Anna Ordonez, Andreas Witt, Susan van Schalkwyk and Jame Agapoff for their helpful input on this manuscript

Author information

Authors and affiliations.

Intermountain Healthcare, Salt Lake City, UT, USA

Gerrit I. van Schalkwyk

You can also search for this author in PubMed   Google Scholar

Contributions

GIS conceived and wrote the manuscript entirely. The author read and approved the final manuscript.

Corresponding author

Correspondence to Gerrit I. van Schalkwyk .

Ethics declarations

Ethics approval and consent to participate, consent for publication.

The author provides full consent for the publication of this manuscript.

Competing interests

The author declares no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

van Schalkwyk, G.I. Editorial: Best practices for correlational research in CAPMH. Child Adolesc Psychiatry Ment Health 17 , 18 (2023). https://doi.org/10.1186/s13034-023-00562-6

Download citation

Accepted : 19 January 2023

Published : 01 February 2023

DOI : https://doi.org/10.1186/s13034-023-00562-6

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Child and Adolescent Psychiatry and Mental Health

ISSN: 1753-2000

• Submission enquiries: [email protected]

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 07 November 2022

A systematic review of worldwide causal and correlational evidence on digital media and democracy

• Philipp Lorenz-Spreen   ORCID: orcid.org/0000-0001-6319-4154 1   na1 ,
• Lisa Oswald   ORCID: orcid.org/0000-0002-8418-282X 2   na1 ,
• Stephan Lewandowsky   ORCID: orcid.org/0000-0003-1655-2013 3 , 4 &
• Ralph Hertwig   ORCID: orcid.org/0000-0002-9908-9556 1  

Nature Human Behaviour volume  7 ,  pages 74–101 ( 2023 ) Cite this article

46k Accesses

53 Citations

612 Altmetric

Metrics details

• Communication
• Science, technology and society

One of today’s most controversial and consequential issues is whether the global uptake of digital media is causally related to a decline in democracy. We conducted a systematic review of causal and correlational evidence ( N  = 496 articles) on the link between digital media use and different political variables. Some associations, such as increasing political participation and information consumption, are likely to be beneficial for democracy and were often observed in autocracies and emerging democracies. Other associations, such as declining political trust, increasing populism and growing polarization, are likely to be detrimental to democracy and were more pronounced in established democracies. While the impact of digital media on political systems depends on the specific variable and system in question, several variables show clear directions of associations. The evidence calls for research efforts and vigilance by governments and civil societies to better understand, design and regulate the interplay of digital media and democracy.

Similar content being viewed by others

Persistent interaction patterns across social media platforms and over time

Michele Avalle, Niccolò Di Marco, … Walter Quattrociocchi

Global warming and heat extremes to enhance inflationary pressures

Maximilian Kotz, Friderike Kuik, … Christiane Nickel

Negativity drives online news consumption

Claire E. Robertson, Nicolas Pröllochs, … Stefan Feuerriegel

The ongoing heated debate on the opportunities and dangers that digital media pose to democracy has been hampered by disjointed and conflicting results (for recent overviews, see refs. 1 , 2 , 3 , 4 ). Disagreement about the role of new media is not a novel phenomenon; throughout history, evolving communication technologies have provoked concerns and debates. One likely source of concern is the dual-use dilemma, that is, the inescapable fact that technologies can be used for both noble and malicious aims. For instance, during the Second World War, radio was used as a propaganda tool by Nazi Germany 5 , whereas allied radio, such as the BBC, supported resistance against the Nazi regime, for example, by providing tactical information on allied military activities 6 , 7 . In the context of the Rwandan genocide, radio was used to incite Rwandan Hutus to massacre the country’s Tutsi minority 8 . In the aftermath of the genocide, using the same means to cause different ends, the radio soap opera ‘Musekeweya’ successfully reduced intergroup prejudice in a year-long field experiment 9 , 10 .

Digital media appears to be another double-edged sword. On the one hand, it can empower citizens, as demonstrated in movements such as the Arab Spring 11 , Fridays for Future and #MeToo 12 . On the other hand, digital media can also be instrumental in inciting destructive behaviours and tendencies such as polarization and populism 13 , as well as fatal events such as the attack on the United States Capitol in January 2021. Relatedly, the way political leaders use or avoid digital media can vary greatly depending on the political context. Former US President Trump used it to spread numerous lies ranging from claims about systematic voter fraud in the 2020 presidential election to claims about the harmlessness of Covid-19. In spring 2022, Russian President Putin had banned most social media platforms that would bypass the state-controlled classical media, probably to prevent access to information about his army’s attack on Ukraine 14 . At the same time, Ukrainian President Zelensky has skilfully used social media to boost Ukrainian morale and engage in the information war with Russia. Examples of the dual-use dilemma of digital media abound.

Clearly, digital media can foster liberation, democratization and participation, but can also play an important role in eroding democracy. The role of digital media is further complicated because unlike other communication technologies, it enables individuals to easily produce and disseminate content themselves, and offers largely frictionless interaction between users. These properties have not only moved the self-organized political behaviour of citizens into the spotlight 15 , but have also shifted power to large digital media platforms. Unlike broadcasters, digital media platforms typically do not create content; instead, their power lies in providing and governing a digital infrastructure. Although that infrastructure could serve as an online public sphere 16 , it is the platforms that exert much control over the dynamics of information flow.

Our goal is to advance the scientific and public debate on the relationship between digital media and democracy by providing an evidence-based picture of this complex constellation. To this end, we comprehensively reviewed and synthesized the available scientific knowledge 17 on the link between digital media and various politically important variables such as participation, trust and polarization.

We aimed to answer the pre-registered question “If, to what degree and in which contexts, do digital media have detrimental effects on democracy?” (pre-registered protocol, including research question and search strategy, at https://osf.io/7ry4a/ ). This two-stage question encompasses, first, the assessment of the direction of effects and, second, how these effects play out as a function of political contexts.

A major difficulty facing researchers and policy makers is that most studies relating digital media use to political attitudes and behaviours are correlational. Because it is nearly impossible to simulate democracy in the laboratory, researchers are forced to rely on observational data that typically only provide correlational evidence. We therefore pursued two approaches. First, we collected and synthesized a broad set of articles that examine associations between digital media use and different political variables. We then conducted an in-depth analysis of the small subset of articles reporting causal evidence. This two-step approach permitted us to focus on causal effects while still taking the full spectrum of correlational evidence into account.

For the present purpose, we adopted a broad understanding of digital media, ranging from general internet access to the use of specific social media platforms, including exposure to certain types of content on these platforms. To be considered as a valid digital media variable in our review, information or discussion forums must be hosted via the internet or need to describe specific features of online communication. For example, we considered the online outlets of traditional newspapers or TV channels as digital source of political information but not the original traditional media themselves. We provide an overview of digital media variables present in our review sample in Fig. 1d and discriminate in our analyses between the two overarching types of digital media: internet, broadly defined, on the one hand and social media in particular on the other hand.

a , Combinations of variables in the sample: digital media (A), political variables (B) and content features such as selective exposure or misinformation (C). Numbers in brackets count articles in our sample that measure an association between variables. b , Geographic distribution of articles that reported site of data collection. c , d , Distribution of measurements (counted separately whenever one article reported several variables) over combinations of outcome variables and methods ( c ) and over combinations of outcome variables and digital media variables ( d ).

We further aimed to synthesize evidence on a broad spectrum of political attitudes and behaviours that are relevant to basic democratic principles 18 . We therefore grounded our assessment of political variables in the literature that examines elements of modern democracies that are considered essential to their functioning, such as citizens’ basic trust in media and institutions 19 , a well-informed public 20 , an active civil society 21 , 22 and exposure to a variety of opinions 23 , 24 . We also included phenomena that are considered detrimental to the functioning of democracies, including open discrimination against people 25 , political polarization to the advantage of political extremists and populists 26 and social segregation in homogeneous networks 23 , 27 .

The political variables in focus are themselves multidimensional and may be heterogeneous and conflicting. For example, polarization encompasses partisan sorting 28 , affective polarization 29 , issue alignment 30 , 31 and a number of other phenomena (see ref. 32 for an excellent literature review on media effects on variations of ideological and affective polarization). For our purpose, however, we take a broader perspective, examining and comparing across different political variables the directions—beneficial or detrimental to democracy—in which digital media effects play out.

Notwithstanding the nuances within each dimension of political behaviour, wherever possible we explicitly interpreted each change in a political variable as tending to be either beneficial or detrimental to democracy. Even though we tried to refrain from normative judgements, the nature of our research question required us to interpret the reported evidence regarding its relation to democracy. For example, an increase in political knowledge is generally considered to be beneficial under the democratic ideal of an informed citizenry 20 . Similarly, a certain level of trust in democratic institutions is crucial for a functioning democracy 33 . By contrast, various forms of polarization (particularly affective polarization) tend to split societies into opposing camps and threaten democratic decision-making 34 , 35 . Likewise, populist politics that are often coupled with right-wing nationalist ideologies, artificially divide society into a corrupt ‘elite’ that is opposed by ‘the people’, which runs counter to the ideals of a pluralistic democracy and undermines citizens’ trust in politics and the media 36 , 37 . We therefore considered polarization and populism, for example, to be detrimental to democracy.

There are already some systematic reviews of subsets of associations between political behaviour and media use that fall within the scope of our analysis, including reviews of the association between media and radicalization 38 , 39 , polarization 32 , hate speech 40 , participation 41 , 42 , 43 , 44 , 45 , echo chambers 46 and campaigning on Twitter 47 . These extant reviews, however, did not contrast and integrate the wide range of politically relevant variables into one comprehensive analysis—an objective that we pursue here. For the most relevant review articles, we matched the references provided in them with our reference list (see Materials and Methods for details). Importantly, and unlike some extant reviews, our focus is not on institutions, the political behaviour of political elites (for example, their strategic use of social media; see refs. 47 , 48 ), or higher-level outcomes (for example, policy innovation in governments 49 ). We also did not consider the effects of traditional media (for example, television or radio) or consumption behaviours that are not specific to digital media (for example, selective exposure 50 ). Furthermore, we did not focus on the microscopic psychological mechanisms that could shape polarization on social media (for a review, see ref. 51 ). For reasons of external validity, we omitted small-scale laboratory-only experiments (for example, see ref. 52 ), but included field experiments in our review. We included studies using a variety of methods—from surveys to large-scale analyses of social media data—and across different disciplines that are relevant to our research question. Details on the inclusion and exclusion criteria are provided in Materials and Methods. Our goal for this knowledge synthesis is to provide a nuanced foundation of shared facts for a constructive stage in the academic but also societal debate about the future of digital media and their role in democracy. In our view, this debate and the future design of digital media for democracy require a comprehensive assessment of its impact. We therefore not only focus on individual dimensions of political behaviour but also compare these dimensions and the methods by which they have been researched so far, thus going beyond the extant reviews. This approach aims to stimulate research that fills evidence gaps and establishes missing links that only become apparent when comparing the dimensions.

After conducting a pre-registered search (most recent update 15 September 2021) and selection process, we arrived at a final sample of N  = 496 articles. For further analysis, we classified them by the set of variables between which they report associations: type of digital media (for example, social media, online news), political variables (for example, trust, participation) and characteristics of the information ecology (for example, misinformation, selective exposure), as depicted in Fig. 1a . Each article was coded according to the combination of these variables as well as the method, specific outcome variable and, if applicable, the direction of association and potential moderator variables (see Materials and Methods for details). The resulting table of the fully coded set of studies can be found at https://osf.io/7ry4a/ , alongside the code for the analyses and visualizations offered here.

Figure 1 reports the composition of the set of included articles. Figure 1a confirms that the search query mainly returned articles concerned with the most relevant associations between digital media and political outcomes. Most of the articles were published in the last 5 years, highlighting the fast growth of interest in the link between digital media and democracy. Articles span a range of disciplines, including political science, psychology, computational science and communication science. Although a preponderance of articles focused on the United States, there was still a large geographical variation overall (see Fig. 1b ).

Figure 1c shows the distribution of measurements (counted separately when one article reported several outcomes) across methods and political variables. Our search query was designed to capture a broad range of politically relevant variables, which meant that we had to group them into broader categories. The ten most frequently reported categories of variables were trust in institutions, different variants of political participation (for example, voter turnout or protest participation), exposure to diverse viewpoints in the news, political knowledge, political expression, measures of populism (for example, support for far-right parties or anti-minority rhetoric), prevalence and spread of misinformation, measures of polarization (for example, negative attitudes towards political opponents or fragmented and adversarial discourse), homophily in social networks (that is, social connections between like-minded individuals) and online hate (that is, hate speech or hate crime). Similarly, the distribution of outcomes and associated digital media variables in Fig. 1d shows that many studies focused on political information online, and specifically political information on social media, in combination with political polarization and participation, while other digital media variables, such as messenger platforms are less explored. The full table, including the reported political variables within each category, can be found at https://osf.io/7ry4a/ . Figure 1 also reveals gaps in the literature, such as rarely explored geographical regions (for example, Africa) and under-studied methods–variable combinations (for example, involving the combination of data sources such as social media data with survey or secondary data).

Direction of associations

In the first part of our research question, we ask whether the available evidence suggests that the effects of digital media are predominantly beneficial or detrimental to democracy. To find an answer, we first selected subsets of articles that addressed the ten most frequently studied categories of political variables (hereafter simply referred to as political variables). We did not test specific hypotheses in our review. A total of N  = 354 associations were reported for these variables (when an article examined two relevant outcome variables, two associations were counted). The independent variable across these articles was always a measure of the usage of some type of digital media, such as online news consumption or social media uptake. Statistically speaking, the independent variables can be positively or negatively associated with the political outcome variable. For instance, more digital media use could be associated with more expression of hate (positive association), less expression of hate (negative association), or not associated at all. We decided to present relationships not at a statistical level but at a conceptual level. We therefore classified each observed statistical association as beneficial or detrimental depending on whether its direction was aligned or misaligned with democracy. For example, a positive statistical association between digital media use and hate speech was coded as a detrimental association; by contrast, a positive statistical association between digital media use and participation was coded as beneficial. Throughout, we represent beneficial associations in turquoise and detrimental associations in orange, irrespective of the underlying statistical polarity.

Figure 2 provides an overview of the ten most frequently studied political variables and the reported directions—colour-coded in terms of whether they are beneficial or detrimental to democracy—of each of their associations with digital media use. This overview encompasses both correlational and causal evidence. Some findings in Fig. 2 suggest that digital media can foster democratic objectives. First, the associations reported for participation point mostly in beneficial directions for democracy (aligned with previous results 45 ), including a wide range of political and civic behaviour (Fig. 1d ), from low-effort participation such as liking/sharing political messages on social media to high-cost activities such as protesting in oppressive regimes. Second, measures of political knowledge and diversity of news exposure appear to be associated with digital media in beneficial ways, but the overall picture was slightly less clear. Third, the literature is also split on how political expression is associated with digital media. Articles reporting beneficial associations between digital media and citizens’ political expression were opposed by a number of articles describing detrimental associations. These detrimental associations relate to the ‘spiral of silence’ idea, that is, the notion that people’s willingness to express their political opinions online depends on the perceived popularity of their opinions (see relevant overview articles 53 , 54 ).

Directions of associations are reported for various political variables (see Fig. 1d for a breakdown). Insets show examples of the distribution of associations with trust, news exposure, polarization and network homophily over the different digital media variables with which they were associated.

Fourth, we observed consistent detrimental associations for a number of variables. Specifically, the associations with trust in institutions were overwhelmingly pointing in directions detrimental to a functioning democracy. Measures of hate, polarization and populism were also widely reported to have detrimental associations with digital media use in the clear majority of articles. Likewise, increased digital media use was often associated with a greater exposure to misinformation. Finally, we also found that digital media were associated with homophily in social networks in detrimental ways (mostly measured on social media, and here especially on Twitter), but the pattern of evidence was a little less consistent. Differences in the consistency of results were also reflected when broken down along associated digital media variables (see insets in Fig. 2 ). For instance, both trust and polarization measures were consistently associated with media use across types of digital media ranging from social media to political information online; in contrast, results for homophily were concentrated on social media and especially on Twitter, while measurements of news exposure were mostly concentrated on political information online.This points not only to different operationalizations of related outcome measures, such as diverse information exposure and homophilic network structures, but also to differences between the distinct domains of digital media in which these very related phenomena are measured. Similar observations can be made when separating associations between general types of digital media: social media vs internet more broadly (Supplementary Fig. 1 ).

Next, we distinguished between articles reporting correlational versus causal evidence and focused on the small subset of articles reporting the latter ( N  = 24). We excluded causal evidence on the effects of voting advice applications from our summary as a very specific form of digital media, explicitly constructed to inform vote choices, and already extensively discussed in a meta-analysis 55 .

Causal inference

Usually, the absence of randomized treatment assignment, an inescapable feature of observational data (for example, survey data), precludes the identification of causal effects because individuals differ systematically on variables other than the treatment (or independent) variable. However, under certain conditions, it is possible to rule out non-causal explanations for associations, even in studies without random assignment that are based on observational data (see refs. 56 , 57 , 58 ). For a more detailed explanation of the fundamental principles of causal inference, see Supplementary Material page 5 and, for example, the work of the 2021 laureates of the Nobel Memorial Prize in Economics 56 , 57 , 58 .

Common causal inference techniques that were used in our sample include instrumental variable designs that introduce exogenous variation in the treatment variable 59 , 60 , 61 , 62 , 63 , matching approaches to explicitly balance treatment and control groups 64 , 65 , 66 , and panel designs that account for unobserved confounders with unit and/or time-fixed effects 67 , 68 . We also found multiple large-scale field experiments conducted on social media platforms 69 , 70 , 71 , 72 as well as various natural experiments 59 , 61 , 62 , 73 .

Figure 3 summarizes the findings and primary causal inference techniques of these articles. Again, causal effects were coded as beneficial for or detrimental to democracy. This figure is structured according to whether evidence stemmed from established democracies or from emerging democracies and authoritarian regimes, adopting classifications from the Liberal Democracy Index provided by the Varieties of Democracy project 18 . In some autocratic regimes (for example, China), it is particularly difficult to interpret certain effects. For example, a loss of trust in government suggests a precarious development for an established democracy; in authoritarian regimes, however, it may indicate a necessary step toward overcoming an oppressive regime and, eventually, progressing towards a more liberal and democratic system. Instead of simply adopting the authors’ interpretation of the effects or imposing our own interpretation of effects in authoritarian contexts, we leave this interpretation to the reader (denoted in purple in the figure). The overall picture converges closely with the one drawn in Fig. 2 . We found general trends of digital media use increasing participation and knowledge but also increasing political polarization and decreasing trust that mostly aligned with correlational evidence.

Each box represents one article. Treatments (T) are in white boxes on the left, political outcome (O) variables in coloured boxes on the right; M denotes mediators; H represents sources of effect heterogeneity or moderators. Positive (+) and negative (−) signs at paths indicate reported direction of effects. Location of sample indicated in top right corner of boxes, primary causal inference strategy in bottom left. Strategies include statistical estimation strategies such as instrumental variables (IV), matching and panel designs (PD) that use, for example, fixed effects (FE) or difference in difference (DiD) for causal estimation, as well as lab or field experiments (for example, field experiments rolled out on various platforms that are often supplemented with IV estimation to account for imperfect compliance). Detrimental effects on liberal democracy are shown in orange, beneficial effects in turquoise, effects open to interpretation in purple and null effects in grey. Solid arrows represent pathways for which authors provide causal identification strategies, dashed arrows represent descriptive (mediation) pathways.

Effects on key political variables

In the following sections, we provide a short synopsis of the results, point to conflicting trends and highlight some examples of the full set of correlational and causal evidence, reported in Figs. 2 and 3 , for six variables that we found to be particularly crucial for democracy: participation, trust, political knowledge, polarization, populism, network structures and news exposure. The chosen examples are stand-ins and illustrations of the general trends.

Participation

Consistent with past meta-analyses 42 , 43 , 45 , the body of correlational evidence supported a beneficial association between digital media use and political participation and mobilization.

Causal analyses of the effects of digital media on political participation in established democracies mostly studied voting and voter turnout 64 , 67 , 71 , 74 , 75 , 76 ; articles concerned with other regions of the world rather focused on political protest behaviour 59 , 61 , 66 . Other articles considered online political participation 65 , 71 . One study, applying causal mediation analysis to assess a causal mechanism 77 , found that information-oriented social media use affects political participation, mediated or enabled through the user’s online political efficacy 65 . Overall, our evidence synthesis found largely beneficial mobilizing effects for political participation across this set of articles. Our search did not identify any studies that examined causal effects of digital media on political participation in authoritarian regimes in Africa or the Middle East.

Many articles in our sample found detrimental associations between digital media and various dimensions of trust (Fig. 2 ). For example, detrimental associations were found for trust in governments and politics 59 , 60 , 66 , 78 , 79 , 80 , 81 , 82 , trust in media 83 , and social and institutional trust 84 . During the COVID-19 pandemic, digital media use was reported to be negatively associated with trust in vaccines 85 , 86 . Yet the results about associations with trust are not entirely homogeneous. One multinational survey found beneficial associations with trust in science 87 ; others found increasing trust in democracy with digital media use in Eastern and Central European samples 88 , 89 . Nevertheless, the large majority of reported associations between digital media use and trust appear to be detrimental for democracy. While the evidence stems mostly from surveys, results gathered with other methods underpin these findings (Fig. 2 inset).

The majority of articles identifying causal effects also find predominantly detrimental effects of digital media on trust. A field experiment in the United States that set browser defaults to partisan media outlets 37 found a long-term loss of trust in mainstream media. Studies examining social trust as a central component of social capital find consistent detrimental effects of social media use 84 ; in contrast, no effects of broadband internet in general on social trust was found 90 . In authoritarian regimes in Asia, increasing unrestricted internet access decreased levels of trust in the political system 59 , 73 , 91 . This finding confirms the predominant association observed in most other countries. Yet it also illustrates how digital media is a double-edged sword, depending on the political context: by reducing trust in institutions, digital media can threaten existing democracies as well as foster emerging democratic developments in authoritarian regimes.

Political knowledge

The picture was less clear for associations between the consumption of digital media and political knowledge. Still, the majority of associations point in beneficial directions and were found in both cross-sectional surveys 92 , 93 , 94 , 95 , 96 , 97 , 98 , 99 and panel surveys 100 , 101 , 102 . Studies linking web-tracking and survey data showed increased learning about politics 103 , but also a turning away from important topics 104 , whereas other experiments demonstrated an overall beneficial effect of digital media on issue salience 105 . These findings, however, stand in contrast to other studies that find a detrimental association between political knowledge and digital media use 106 , 107 , 108 , 109 , 110 .

The body of causal evidence on political knowledge also tends to paint a relatively promising picture. Multiple articles found that engagement with digital media increased political knowledge 67 , 70 , 72 , 74 and that engagement with political content on social media increased political interest among adolescents 111 . In line with these findings, it has been reported that political messages on social media, as well as faster download speed, can increase information-seeking in the political domain 67 , 71 . By contrast, there is evidence for a decrease in political knowledge 112 , which is mediated through the news-finds-me effect: social media users believe that actively seeking out news is no longer required to stay informed, as they expect to be presented with important information.

It is important to note that most of these effects are accompanied by considerable heterogeneity in the population that benefits and the type of digital media. For example, politically interested individuals showed higher knowledge acquisition when engaging with Twitter, whereas the opposite effects emerged for engagement with Facebook 113 . Furthermore, there is evidence that the news-finds-me effect on social media can be mitigated when users consult alternative news sources 112 .

Polarization

Most articles found detrimental associations between digital media and different forms of political polarization 114 , 115 , 116 , 117 , 118 . Our review obtained evidence for increasing outgroup polarization on social media in a range of political contexts and on various platforms 119 , 120 , 121 , 122 . Increasing polarization was also linked to exposure to viewpoints opposed to one’s own on social media feeds 69 , 123 . Articles comparing several political systems found associations that were country-dependent 124 , again highlighting the importance of political context 125 . Nevertheless, high digital media use was for the most part linked to higher levels of polarization, although there was some evidence for balanced online discourse without pronounced patterns of polarization 126 , 127 , 128 , as well as evidence for potentially depolarizing tendencies 129 .

The body of causal articles largely supported the detrimental associations of digital media that emerged, by and large, in the correlational articles. Among established democracies, both social media use and overall internet use increased political polarization 63 , 70 . This was also the case for an experimental treatment that exposed users to opposing views on Twitter 69 . However, some findings run counter to the latter result 130 : in a 2 month field experiment, exposure to counter-attitudinal news on Facebook reduced affective polarization (the authors used opposing news outlets as treatment instead of opinions on social media). Furthermore, one other field experiment did not find evidence that exposure to partisan online news substantively shifted political opinions but found a long-term loss of trust in mainstream media 37 . Still, taking all evidence into account, the overall picture remains largely consistent on the detrimental association between digital media and political polarization, including some but not all causal evidence.

Articles on populism in our review examined either vote share and other popularity indicators for populist parties or the prevalence of populist messages and communication styles on digital media. Overall, articles using panel surveys, tracking data and methods linking surveys to social media data consistently found that digital media use was associated with higher levels of populism. For example, digital platforms were observed to benefit populist parties more than they benefit established politicians 131 , 132 , 133 , 134 . In a panel survey in Germany, a decline in trust that accompanied increasing digital media consumption was also linked to a turn towards the hard-right populist AfD party 80 . This relationship might be connected to AfD’s greater online presence, relative to other German political parties 132 , even though these activities might be partly driven by automated accounts. There is also evidence for an association between increased social media use and online right-wing radicalization in Austria, Sweden and Australia 135 , 136 , 137 . Only a minority of articles found no relationship or the reverse relationship between digital media and populism 138 , 139 , 140 . For instance, in Japan, internet exposure was associated with increased tolerance towards foreigners 141 .

Similarly, most causal inference studies linked increased populism to digital media use. For instance, digital media use in Europe led to increased far-right populist support 63 , 142 , and there was causal evidence that digital media can propagate ethnic hate crimes in both democratic and authoritarian countries 62 , 68 . Leaving the US and European political context, in Malaysia, internet exposure was found to cause decreasing support for the authoritarian, populist government 60 .

Echo chambers and news exposure

The evidence on echo chambers points in different directions depending on the outcome measure. On the one hand, when looking at news consumption, several articles showed that social media and search engines diversify people’s news diets 67 , 143 , 144 , 145 , 146 . On the other hand, when considering social networks and the impact of digital media on homophilic structures, the literature contains consistent reports of ideologically homogeneous social clusters 147 , 148 , 149 , 150 , 151 . This underscores an important point: some seemingly paradoxical results can potentially be resolved by looking more closely at context and specific outcome measurement (see also Supplementary Fig. 2 ). The former observation of diverse news exposure might fit with the beneficial relationship between digital media and knowledge reported in refs. 67 , 74 , 94 , 95 , 102 , and the homophilic social structures could be connected to the prevalence of hate speech and anti-outgroup sentiments 120 , 152 , 153 , 154 , 155 .

Heterogeneity

We now turn to the second part of our research question and analyse the effects of digital media use in light of different political contexts. Figure 4 shows the geographical distribution of effect directions around the globe. Notably, most beneficial effects on democracy were found in emerging democracies in South America, Africa and South Asia. Mixed effects, by contrast, were distributed across Europe, the United States, Russia and China. Similarly, detrimental outcomes were mainly found in Europe, the United States and partly Russia, although this may reflect a lack of studies undertaken in authoritarian contexts. These patterns are also shown in Fig. 4c,d , where countries are listed according to the Liberal Democracy Index. Moderators—variables such as partisanship and news consumption that are sources of effect heterogeneity—displayed in Supplementary Fig. 3 also show slight differences between outcomes. Beneficial outcomes seemed to be more often moderated by political interest and news consumption, whereas detrimental outcomes tended to be moderated by political position and partisanship.

a , Geographical distribution of reported associations for the variables trust, knowledge, participation, exposure and expression. Pie charts show the composition of directions for each country studied. b , Geographic representation of reported associations for the variables hate, polarization, populism, homophily and misinformation. c , Data and variables in a , in absolute numbers of reported associations and sorted along the Liberal Democracy Index 18 . d , Data and variables in b , in absolute numbers of reported associations and sorted along the Liberal Democracy Index.

Furthermore, many causal articles acknowledge that effects differ between subgroups of their sample when including interaction terms in their statistical models. For example, the polarizing effects of digital media differ between Northern and Southern European media systems 142 : while consumption of right-leaning digital media increased far-right votes, especially in Southern Europe, the consumption of news media and public broadcasting in Northern European media systems with high journalistic standards appears to mitigate these effects. Another example of differential effects between subgroups was found in Russia, where the effects of social media on xenophobic violence were only present in areas with pre-existing nationalist sentiment. This effect was especially pronounced for hate crimes with a larger number of perpetrators, indicating that digital media was serving a coordinating function. In summary, a range of articles found heterogeneity in effects for varying levels of political interest 67 , 113 , political orientation 63 , 69 , 70 and different characteristics of online content 111 .

Most authors, particularly those of the causal inference articles in our body of evidence, explicitly emphasized the national, cultural, temporal and political boundary conditions for interpreting and generalizing their results (see, for example, ref. 111 ). By contrast, especially in articles conducted on US samples, the national context and the results’ potential conditionality was often not highlighted. We strongly caution against a generalization of findings that are necessarily bound to a specific political setting (for example, the United States) to other contexts.

Sampling methods and risk of bias

To assess study quality and risk of bias, we additionally coded important methodological aspects of the studies, specifically, the sampling method, sample size and transparency indicators, such as competing interest, open data practices and pre-registrations. In Fig. 5 , we show an excerpt from that analysis. Different sampling methods naturally result in different sample sizes as shown in Fig. 5a,b . Furthermore, behavioural data are much more prevalent for studies that look at detrimental outcomes, such as polarization and echo chambers. Classic surveys with probability samples or quota samples, in contrast, are often used to examine beneficial outcome measures such as trust and participation (Fig. 5c,d ). Overall, however, no coherent pattern emerges in terms of the reported directions of associations. If anything, large probabilistic samples report relatively less beneficial associations for both types of outcomes (Fig. 5 ). Generally, different types of data have different advantages, such as probability and quota samples approximating more closely the ideal of representativeness, whereas the observation of actual behaviour on social media escaping the potential downsides of self-reporting. A potential blind spot in studies working with behavioural data from social media, inaccessible to both us and the original authors of the studies, is the selection of data provided by platforms. Therefore, it is tremendously important for researchers to get unrestricted access or, at least, transparent provision of random samples of data by platforms. The selection of users into the platforms, however, remains an open issue for behavioural data as it is often unclear who the active users are and why they are active online. We find that political outcome measures studied with behavioural data appear to show quite distinct results compared with those studied with large-scale survey data. Combining both data types would probably maximize the chances for reliable conclusions about the impact of digital media on democracy.

a , Sample size vs sampling methods for variables of trust, knowledge, participation, exposure and expression. Each dot represents one measurement, colour coded according to the direction of the reported association. b , Sample size vs sampling method for variables of hate, polarization, populims, network homophily and misinformation. c , More detailed breakdown for the same varibales as in a of sampling methods and their respective counts of reported associations and their direction. d , Breakdown of sampling methods and counts of associations for the same variables as in b .

We found relatively few null effects for some variables. This could be accurate, but it could also be driven by the file-drawer problem—the failure to publish null results. To examine the extent of a potential file-drawer problem, we contacted authors via large mailing lists but did not receive any unpublished work that fitted our study selection criteria. Regarding possible risk of bias, we found that only in 143 out of 354 measurements did authors clearly communicate that no conflict of interest was present (beyond the usual funding statement). However, we did not find a striking imbalance in the distribution of reported associations between those articles that did not explicitly state competing interest and those that did. Of the few associations for which conflicts of interest were stated, 4 pointed in beneficial, 3 in detrimental and 2 reported lack of directionality. In only 79 of 354 measurements did the researchers use open data practices. Considering articles that reported detrimental associations, we did not find a clear difference in the directions between those with and without open data. However, considering articles that reported beneficial outcomes, the numbers of positive findings in the studies without open data are relatively much larger than for the open science studies. Namely, 103 beneficial and 33 detrimental associations were reported in those without open data, while 19 beneficial versus 14 detrimental were reported in studies with open data practices. This observation might be due to the large number of survey-based studies about participation, which often do not follow open data practices. Even fewer of the studies in our sample were pre-registerd, namely, 13 of the 354, where 9 reported detrimental associations, only 3 reported beneficial associations and 1 found no direction of association. To shed light on other potential biases, we additionally examined temporal variations in the directions of reported associations and found, besides the general explosive growth of studies in this domain, a slight trend towards an increasing number of both detrimental directions and null effects over time (Supplementary Fig. 4 ). At the author level, there was no clear pattern in the associations reported by those authors who published the greatest number of articles in our sample; several authors variously reported detrimental and beneficial effects as well as null effects, with a few exceptions (Supplementary Fig. 5 ). Their co-authorship network in Supplementary Fig. 6 , split for the two types of outcomes measures, shows some communities of co-authors; however, no clear pattern of preferred direction of reported association can be spotted. Overall, we did not find evidence of a systematic bias in either direction driven by temporal trends or particular authors.

Regardless of whether they are authoritarian, illiberal, or democratic, governments around the world are concerned with how digital media affect governance and their citizenry’s political beliefs and behaviours. A flurry of recent interdisciplinary research, stimulated in part by new methodological possibilities and data sources, has shed light on this potential interplay.

Although classical survey methods are still predominant, novel ways of linking data types, for example linking URL tracking data or social media data with surveys, permit more complex empirical designs and analyses. Furthermore, digital trace data allow an expansion in sample size. The articles we reviewed included surveys with a few hundred, up to a few thousand participants, but also large-scale social media analyses that included behavioural traces of millions. Yet with computational social science still in its early days, the amount of evidence supporting and justifying causal conclusions is still limited. Causal effects of digital media on political variables are also hard to pin down empirically due to a plethora of complexities and context factors, as well as the highly dynamic technological developments that make predicting the future difficult. While emergent political phenomena are hard to simulate in the lab, the value of estimation and data collection strategies to draw causal inferences from real-life data is enormous. However, the long-established trade-off between internal and external validity still applies, which also highlights the value of high-quality descriptive work.

Taking into account both correlational and causal evidence, our review suggests that digital media use is clearly associated with variables such as trust, participation and polarization. They are critical for the functioning of any political system, in particular democracies. Extant research reports relatively few null effects. However, the trends on each factor mostly converge, both across research methods and across correlative and causal evidence.

Our results also highlight that digital media are a double-edged sword, with both beneficial and detrimental effects on democracy. What is considered beneficial or detrimental will, at least partly, hinge on the political system in question: intensifying populism and network homophily may benefit a populist regime or a populist politician but undermine a pluralistic democracy. For democratic countries, evidence clearly indicates that digital media increase political participation. Less clear but still suggestive are the findings that digital media have positive effects on political knowledge and exposure to diverse viewpoints in news. On the negative side, however, digital media use is associated with eroding the ‘glue that keeps democracies together’ 33 : trust in political institutions. The results indicating this danger converge across methods. Furthermore, our results also suggest that digital media use is associated with increases in hate, populism and polarization. Again, the findings converge across causal and correlational articles.

Alongside the need for more causal evidence, we found several research gaps, including the relationship between trust and digital media and the seeming contradiction between network homophily and diverse news exposure. Methods that link tracking data for measuring news exposure with behavioural data from social media (for example, sharing activities or the sentiment of commenting) are crucial to a better understanding of this apparent contradiction.

Limitations

The articles in our sample incorporate a plethora of methods and measures. As a result, it was necessary to classify variables and effects into broad categories. This is a trade-off we had to make in exchange for the breadth of our overview of the landscape of evidence across disciplines. For the same reason, we could not provide a quantitative comparison across the diverse sample of articles. We believe that digital media research would benefit from more unified measures (for example, for polarization), methods across disciplines to allow for better comparability in the future, a systematic comparison of different types of digital media (that is, Facebook and Twitter are neither of one kind nor, in all likelihood, are their effects) and extensions of outcome measurements beyond certain types of digital media. This follows other recent calls for commensurate measures of political and affective polarization 156 . The breadth of our review and the large number of political outcome measures in particular, made it necessary to be quite restrictive on other ends (see Fig. 6 for our exclusion process and Supplementary Table 1 for the detailed criteria). We explicitly decided to prioritize the selection of causal evidence (see Fig. 7 for an overview of the causal inference techniques that we considered) and other large-sample, quantitative, published evidence. However, following this pre-registered search strategy led to the selection of unequal numbers of studies for different outcome variables. For example, our search query selected considerably more studies examining political participation than political expression or trust, while at the same time, it did not include all studies that are included in other systematic reviews 45 due to stricter exclusion criteria.

a , Keywords included in our search query, run on Web of Science and Scopus, with logical connectors. Focus was on causal inference methods (method column), but also inclusion of descriptive quantitative evidence, relationships between digital media (cause column) and political outcomes (direct effect box) or content features (indirect effect box). b , Flowchart representing the stepwise exclusion process, starting with title-based exclusion, followed by abstract-based exclusion. c , Example illustration of outcome variable extraction from the abstracts. d , Breakdown of the most frequently reported political variables into top 10 categories. Numbers in brackets are counts of measurements in the set.

Fundamental principles of causal inference techniques and statistical strategies used in our sample of causal evidence (excluding field experiments).

The interpretation of our results was in several cases hampered by a lack of appropriate baseline measures. There is no clear measure of what constitutes a reasonable benchmark of desirable political behaviour in a healthy democracy. In addition, there were no means of quantification of some of these behaviours in the past, outside of digital media. This problem is particularly pronounced for factors such as exposure to diverse news, social network homophily, misinformation and hate speech. Measuring these phenomena at scale is possible through digital media (for example, by analysing social network structure); much less is known about their prevalence and dynamics in offline settings. Many articles therefore lacked a baseline. For instance, it is neither clear what level of homophily in social networks is desirable or undesirable in a democratic society, nor is it clear how to interpret the results of certain studies on polarization 69 , 130 , whose findings depend on whether one assumes that social media have increased or decreased exposure to opposing views relative to some offline benchmark. For example, if exposure to opposing views is increased on social media, the conclusion of one study 130 would be that it reduces polarization, but if exposure is decreased, one would come to the opposite conclusion. Notably, in this study, counter-attitudinal exposure was found to be down-ranked by Facebook’s news feed—hence supporting a process that fosters polarization instead of counteracting it. Furthermore, results about populism might be skewed: descriptive evidence on the relative activity and popularity of right-wing populist parties in Europe suggests their over-representation, as in the case of Germany’s AfD, on social media, relative to established democratic parties (see, for example, ref. 132 ). Therefore, it is difficult to interpret even causal effects of digital media use on populist support in isolation from the relative preponderance of right-wing content online.

Our results provide grounds for concern. Alongside the positive effects of digital media for democracy, there is clear evidence of serious threats to democracy. Considering the importance of these corrosive and potentially difficult-to-reverse effects for democracy, a better understanding of the diverging effects of digital media in different political contexts (for example, authoritarian vs democratic) is urgently needed. To this end, methodological innovation is required. This includes, for instance, more research using causal inference methodologies, as well as research that examines digital media use across multiple and interdependent measures of political behaviour. More research and better study designs will, however, also depend on access to data collected by the platforms. This access has been restricted or foreclosed. Yet without independent research that has unhampered access to all relevant data, the effects of digital media can hardly be understood in time. This is even more concerning because digital media can implement architectural changes that, even if seemingly small, can scale up to widespread behavioural effects. Regulation may be required to facilitate this access 157 . Most importantly, we suggest that the bulk of empirical findings summarized here can be attributed to the current status quo of an information ecosystem produced and curated by large, commercial platforms. They have succeeded in attracting a vast global audience of users. The sheer size of their audience as well as their power over what content and how content gets the most attention has led, in the words of the philosopher Jürgen Habermas, to a new structural transformation of the public sphere 16 . In this new public sphere, everybody can be a potential author spontaneously producing content, both right-wing radical networks as well as the courageous Belarusian women standing up for human rights and against a repressive regime. One need not share Habermas’ conception of ‘deliberate democracy’ to see that current platforms fail to produce an information ecosystem that empowers citizens to make political choices that are as rationally motivated as possible. Our results show how this ecosystem plays out to have important consequences for political behaviours and attitudes. They further underscore that finding out which aspects of this relationship are detrimental to democracy and how they can be contained while actively preserving and fostering the emancipatory potential of digital media is, perhaps, one of the most important global tasks of the present. Our analysis hopes to contribute to the empirical basis of this endeavour.

This systematic review follows the MOOSE Guidelines for Meta-Analyses and Systematic Reviews of Observational Studies 158 . The detailed protocol of the review process was pre-registered on the Open Science Framework (OSF) at https://osf.io/7ry4a/ . The repository also contains the completed MOOSE checklist showing where each guideline is addressed in the text.

Figure 6 summarizes the search query that we used on two established academic databases, Scopus and Web of Science (both highly recommended search tools), the resulting number of articles from the query and the subsequent exclusion steps, leading to the final sample size of N  = 496 articles under consideration 159 , 160 , 161 , 162 , 163 , 164 , 165 , 166 , 167 , 168 , 169 , 170 , 171 , 172 , 173 , 174 , 175 , 176 , 177 , 178 , 179 , 180 , 181 , 182 , 183 , 184 , 185 , 186 , 187 , 188 , 189 , 190 , 191 , 192 , 193 , 194 , 195 , 196 , 197 , 198 , 199 , 200 , 201 , 202 , 203 , 204 , 205 , 206 , 207 , 208 , 209 , 210 , 211 , 212 , 213 , 214 , 215 , 216 , 217 , 218 , 219 , 220 , 221 , 222 , 223 , 224 , 225 , 226 , 227 , 228 , 229 , 230 , 231 , 232 , 233 , 234 , 235 , 236 , 237 , 238 , 239 , 240 , 241 , 242 , 243 , 244 , 245 , 246 , 247 , 248 , 249 , 250 , 251 , 252 , 253 , 254 , 255 , 256 , 257 , 258 , 259 , 260 , 261 , 262 , 263 , 264 , 265 , 266 , 267 , 268 , 269 , 270 , 271 , 272 , 273 , 274 , 275 , 276 , 277 , 278 , 279 , 280 , 281 , 282 , 283 , 284 , 285 , 286 , 287 , 288 , 289 , 290 , 291 , 292 , 293 , 294 , 295 , 296 , 297 , 298 , 299 , 300 , 301 , 302 , 303 , 304 , 305 , 306 , 307 , 308 , 309 , 310 , 311 , 312 , 313 , 314 , 315 , 316 , 317 , 318 , 319 , 320 , 321 , 322 , 323 , 324 , 325 , 326 , 327 , 328 , 329 , 330 , 331 , 332 , 333 , 334 , 335 , 336 , 337 , 338 , 339 , 340 , 341 , 342 , 343 , 344 , 345 , 346 , 347 , 348 , 349 , 350 , 351 , 352 , 353 , 354 , 355 , 356 , 357 , 358 , 359 , 360 , 361 , 362 , 363 , 364 , 365 , 366 , 367 , 368 , 369 , 370 , 371 , 372 , 373 , 374 , 375 , 376 , 377 , 378 , 379 , 380 , 381 , 382 , 383 , 384 , 385 , 386 , 387 , 388 , 389 , 390 , 391 , 392 , 393 , 394 , 395 , 396 , 397 , 398 , 399 , 400 , 401 , 402 , 403 , 404 , 405 , 406 , 407 , 408 , 409 , 410 , 411 , 412 , 413 , 414 , 415 , 416 , 417 , 418 , 419 , 420 , 421 , 422 , 423 , 424 , 425 , 426 , 427 , 428 , 429 , 430 , 431 , 432 , 433 , 434 , 435 , 436 , 437 , 438 , 439 , 440 , 441 , 442 , 443 , 444 , 445 , 446 , 447 , 448 , 449 , 450 , 451 , 452 , 453 , 454 , 455 , 456 , 457 , 458 , 459 , 460 , 461 , 462 , 463 , 464 , 465 , 466 , 467 , 468 , 469 , 470 , 471 , 472 , 473 , 474 , 475 , 476 , 477 , 478 , 479 , 480 , 481 , 482 , 483 , 484 , 485 , 486 , 487 , 488 , 489 , 490 , 491 , 492 , 493 , 494 , 495 , 496 , 497 , 498 , 499 , 500 , 501 , 502 , 503 , 504 , 505 , 506 , 507 , 508 , 509 , 510 , 511 , 512 , 513 , 514 , 515 , 516 , 517 , 518 , 519 , 520 , 521 , 522 , 523 , 524 , 525 , 526 , 527 , 528 , 529 , 530 , 531 , 532 , 533 , 534 , 535 , 536 , 537 , 538 , 539 , 540 , 541 , 542 , 543 , 544 , 545 , 546 , 547 , 548 , 549 , 550 , 551 , 552 , 553 , 554 , 555 , 556 , 557 , 558 , 559 , 560 , 561 , 562 , 563 , 564 , 565 , 566 , 567 , 568 , 569 , 570 , 571 , 572 , 573 , 574 , 575 .

Study selection criteria

We included only original, empirical work. Conceptual or theoretical work, simulation studies and evidence synthesizing studies were excluded. Articles had to be published in academic journals in English. Unpublished studies for which only the abstract or a preprinted version was available were excluded from the review. We excluded small- N laboratory experiments and small- N student surveys ( N  < 100) from our body of original work due to validity concerns. Although correlational evidence cannot establish a causal direction, we focused on articles that examined effects of digital media on democracy but not the opposite. We therefore excluded, for example, articles that examined ways to digitize democratic procedures. To be included, articles had to include at least two distinct variables, a digital media variable and a political outcome. Articles measuring a single variable were only included if this variable was a feature of digital media (for example, hate speech prevalence, homophily in online social networks, prevalence of misinformation in digital media).

Search strategy, study selection, coding and data extraction

Articles eligible for our study had to be published before 15 September 2021. We sourced our review database from Scopus and Web of Science, as suggested by ref. 159 . The search query (Fig. 6 ) was constructed in consultation with professional librarians and was designed to be as broad as possible to pick up any articles containing original empirical evidence of direct or indirect effects of digital media on democracy (including correlational evidence). We further consulted recent, existing review articles in the field 32 , 39 , 40 to check for important articles that did not appear in the review body. Articles that were included manually are referenced separately in the flowchart (Fig. 6 ). In addition, we contacted authors via large mailing lists of researchers working on computational social science and misinformation but did not receive any unpublished work that fitted our study selection criteria. The query retrieved N  = 3,509 articles. Of these, 1,349 were retained after screening the titles for irrelevant topics. This first coding round, whether an article, based on the title, fits the review frame or not, was split between two coders. Coders could flag articles that are subject to discussion to let the other coder double check the decision. In this round, only clearly not fitting articles were excluded from the sample. A list of exclusion criteria can be found in SuppIementary Information .

The next coding round, whether an article, based on the abstract, fits the review frame, was conducted in parallel by two coders. The inter-coder reliability, after this round of article selection, was Krippendorff’s alpha of 0.66 (87% agreement). After calculating this value, disagreement between coders was solved through discussion. At this stage, we excluded all studies that were not original empirical work, such as other reviews or conceptual articles, simulation studies and purely methodological articles (for example, hate speech or misinformation detection approaches). This coding round was followed by a more in-depth coding round. Here we refined our exclusion decisions; for example, we excluded studies that examined the digitization of government, preprints, small-scale lab experiments, small-scale convenience or student samples and studies that only included one variable (for example, description of online forums) (see Supplementary Table 1 for a detailed list of criteria). A full-text screen was performed in cases where the relevant information could not be retrieved from the abstract and for all articles implying causal evidence.

After both rounds of abstract screening, 474 articles remained in our sample. After cross-checking the results of our literature search against the references from existing reviews, we found and included further N  = 22 articles that met our thematic criteria but were not identified by our search string. Ultimately, a total of 496 articles were selected into the final review sample. Figure 6b summarizes the selection procedure.

The following information was extracted from each article using a standardized data extraction form: variable groups under research (digital media, features of media and/or political outcome variables), the concrete digital media under research, the explicit political outcome variable, the methods used, the country of origin, causal claims, possible effect heterogeneity (moderation) as well as various potential sources of bias. To assess various quality criteria of the studies, the coders had to visit the full text of the articles (for example, to find the declaration of competing interests, pre-registration or data availability statements, or to consider the methods section). Therefore, and facing the large number of articles under consideration, blinding could not be established during this procedure.

When conducting a systematic review with a broad scope, categories of the variables cannot be exhaustively defined before coding. Therefore, variable categories, especially for the digital media variables and the political outcome variables, were chosen inductively. In the first extraction step, coders stuck closely to the phrasing of the authors of the respective study. To reduce redundancy and refine the clustering of the variables, we iteratively generated frequency tables and manually sorted single variables to the best-fitting categories until a small number of clearly distinct categories was selected. After the categories were defined, both coders re-coded 10% of the sample to calculate inter-coder reliabilities for all key variables. We provide a table of inter-coder reliabilities (percentage agreements and Krippendorff’s alphas) (Supplementary Table 2 ).

Data synthesis and analysis

Due to considerable heterogeneity in methods in the articles—including self-report surveys through network analysis of social media data, URL tracking data and field experiments—no calculation of meta-analytic effect sizes was possible. The final table of selected articles with coded variables will be published alongside this article as a major result of this review project. The effect directions of 10 important political outcome variables (4 consistent with liberal democracy, 4 opposing democratic values) are summarized in Fig. 2 . For articles dealing with these political variables, we also assessed the country in which the study was conducted (Fig. 4 ), as well as explicit sources of effect heterogeneity such as demographic characteristics of study participants or characteristics of the digital media platform.

For the overview analysis, which includes both correlational and causal evidence, we mainly restricted ourselves to the evaluation effects reported in the abstracts. Articles making explicit causal claims and/or using causal inference methods (Fig. 7 ) were examined in-depth and summarized as simplified path diagrams with information on mediators, moderators, country of origin and method used (Fig. 3 ).

Deviations from the protocol

The volume of papers our query returned prevented an in-depth analysis of confounding variables. Instead, our assessment of quality relied on the sampling strategy and sample size, the method used, sources of heterogeneity and transparency criteria, such as open data practices and pre-registration. Furthermore, we were able to construct the co-author network by matching the author’s names, but were unable to produce a meaningful co-citation network due to the incompleteness and ambiguity of references in the export format that we used.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The dataset including all originally collected studies with decision stages ( N  = 3,531, ‘full_data.xlsx’), the table including all papers within our review sample ( N  = 496, ‘data_review.xlsx’) and the table including all effects reported within papers dealing with the top ten outcome measures ( N  = 354, ‘data_effects.xlsx’) are available at https://osf.io/7ry4a/ .

Code availability

R scripts for all analyses and figures are available at https://osf.io/7ry4a/ .

Persily, N. & Tucker, J. A. Social Media and Democracy: The State of the Field, Prospects for Reform (Cambridge Univ. Press, 2020).

Tucker, J. A. et al. Social Media, Political Polarization, and Political Disinformation: A Review of the Scientific Literature https://hewlett.org/library/social-media-political-polarization-political-disinformation-review-scientific-literature/ (2018).

Rau, J. P. & Stier, S. Die echokammer-hypothese: fragmentierung der öffentlichkeit und politische polarisierung durch digitale medien? Z. fur Vgl. Polit. 13 , 399–417 (2019).

Article   Google Scholar  

Lewandowsky, S. et al. Technology and democracy: Understanding the influence of online technologies on political behaviour and decision-making. JRC Publications Repository https://doi.org/10.2760/709177 (2020).

Adena, M., Enikolopov, R., Petrova, M., Santarosa, V. & Zhuravskaya, E. Radio and the rise of the nazis in prewar Germany. Q. J. Econ. 130 , 1885–1939 (2015).

Adena, M., Enikolopov, R., Petrova, M. & Voth, H.-J. Bombs, Broadcasts and Resistance: Allied Intervention and Domestic Opposition to hhe Nazi Regime During World War II https://doi.org/10.2139/ssrn.3661643 (2021).

Gagliarducci, S., Onorato, M. G., Sobbrio, F. & Tabellini, G. War of the waves: radio and resistance during World War II. Am. Econ. J. Appl. Econ. 12 , 1–38 (2020).

Li, D. Echoes of violence: considerations on radio and genocide in Rwanda. J. Genocide Res. 6 , 9–27 (2004).

Paluck, E. L. Reducing intergroup prejudice and conflict using the media: a field experiment in Rwanda. J. Pers. Soc. Psychol. 96 , 574–587 (2009).

Staub, E. & Pearlman, L. A. Reducing intergroup prejudice and conflict: a commentary. J. Pers. Soc. Psychol. 96 , 588–593 (2009).

Howard, P. N. & Hussain, M. M. Democracy’s Fourth Wave? Digital Media and the Arab Spring (Oxford Univ. Press, 2013).

Jackson, S. J., Bailey, M. & Welles, B. F. #HashtagActivism: Networks of Race and Gender Justice (MIT Press, 2020).

Engesser, S., Ernst, N., Esser, F. & Büchel, F. Populism and social media: how politicians spread a fragmented ideology. Inform. Commun. Soc. 20 , 1109–1126 (2017).

Warzel, C. The information war isn’t over yet. The Atlantic (8 March 2022). https://www.theatlantic.com/technology/archive/2022/03/russia-ukraine-war-propaganda/626975/

Bak-Coleman, J. B. et al. Stewardship of global collective behavior. Proc. Natl Acad. Sci. USA 118 , e2025764118 (2021).

Article   CAS   Google Scholar  

Habermas, J. Überlegungen und hypothesen zu einem erneuten strukturwandel der politischen öffentlichkeit. Leviathan 470 , 470–500 (2021).

Google Scholar  

The value of evidence synthesis. Nat. Hum. Behav. 5 , 539 (2021).

Coppedge, M. et al. Varieties of Democracy: Measuring Two Centuries of Political Change (Cambridge Univ. Press, 2020).

Warren, M. E. in Handbook on Political Trust (eds Zmerli, S. & Van der Meer, T. W.) 33–52 (Edward Elgar Publishing, 2017).

Milner, H. Civic Literacy: How Informed Citizens Make Democracy Work (UPNE, 2002).

Kohler-Koch, B. & Quittkat, C. De-mystification of Participatory Democracy: EU-Governance and Civil Society (OUP Oxford, 2013).

O’Connell, B. & Gardner, J. W. Civil Society: The Underpinnings of American Democracy (UPNE, 1999).

Sunstein, C. R. The law of group polarization. J. Polit. Philos. 10 , 175–195 (2002).

Habermas, J., Lennox, S. & Lennox, F. The public sphere: an encyclopedia article (1964). New Ger. Crit. 3 , 49–55 (1974).

Howard, J. W. Free speech and hate speech. Annu. Rev. Polit. Sci. 22 , 93–109 (2019).

Müller, J.-W. What is Populism? (Univ. Pennsylvania Press, 2016).

Pariser, E. The Filter Bubble: What the Internet is Hiding from You (Penguin, 2011).

Davis, N. T. & Dunaway, J. L. Party polarization, media choice, and mass partisan-ideological sorting. Public Opin. Q. 80 , 272–297 (2016).

Iyengar, S., Lelkes, Y., Levendusky, M., Malhotra, N. & Westwood, S. J. The origins and consequences of affective polarization in the united states. Annu. Rev. Polit. Sci. 22 , 129–146 (2019).

Kozlowski, A. C. & Murphy, J. P. Issue alignment and partisanship in the american public: revisiting the partisans without constraint thesis. Soc. Sci. Res. 94 , 102498 (2021).

Baldassarri, D. & Gelman, A. Partisans without constraint: political polarization and trends in American public opinion. Am. J. Soc. 114 , 408–446 (2008).

Kubin, E. & von Sikorski, C. The role of (social) media in political polarization: a systematic review. Ann. Int. Commun. Assoc. 45 , 188–206 (2021).

Dodsworth, S. & Cheeseman, N. Political Trust: The Glue That Keeps Democracies Together (Westminster Foundation for Democracy, 2020).

McCoy, J. & Somer, M. Toward a theory of pernicious polarization and how it harms democracies: comparative evidence and possible remedies. Ann. Am. Acad. Polit. Soc. Sci. 681 , 234–271 (2019).

Lührmann, A. et al. Democracy Facing Global Challenges V-Dem Annual Democracy Report 2019 (V-Dem Institute, University of Gothenburg, 2019).

Bächtiger, A., Dryzek, J. S., Mansbridge, J., & Warren, M. E. (eds) The Oxford Handbook of Deliberative Democracy (Oxford University Press, 2018).

Guess, A. M., Barberá, P., Munzert, S. & Yang, J. The consequences of online partisan media. Proc. Natl Acad. Sci. USA 118 , e2013464118 (2021).

Odağ, Ö., Leiser, A. & Boehnke, K. Reviewing the role of the internet in radicalization processes. J. Deradicalization 21 , 261–300 (2019).

Hassan, G. et al. Exposure to extremist online content could lead to violent radicalization: a systematic review of empirical evidence. Int. J. Dev. Sci. 12 , 71–88 (2018).

Castano-Pulgarín, S. A., Suárez-Betancur, N., Vega, L. M. T. & López, H. M. H. Internet, social media and online hate speech. Systematic review. Aggress. Violent Behav. 58 , 101608 (2021).

Angyal, E. & Fellner, Z. How are online and offline political activities connected? A comparison of studies. Intersections EEJSP 6 , 81–98 (2020).

Chae, Y., Lee, S. & Kim, Y. Meta-analysis of the relationship between Internet use and political participation: examining main and moderating effects. Asian J. Commun. 29 , 35–54 (2019).

Oser, J. & Boulianne, S. Reinforcement effects between digital media use and political participation : a meta-analysis of repeated-wave panel data. Public Opin. Q. 84 , 355–365 (2020).

Boulianne, S. Social media use and participation: a meta-analysis of current research. Inf. Commun. Soc. 18 , 524–538 (2015).

Boulianne, S. Twenty years of digital media effects on civic and political participation. Commun. Res. 47 , 947–966 (2020).

Terren, L. & Borge-Bravo, R. Echo chambers on social media: a systematic review of the literature. Rev. Commun. Res. 9 , 99–118 (2021).

Jungherr, A. Twitter use in election campaigns: a systematic literature review. J. Inf. Technol. Polit. 13 , 72–91 (2016).

Lewandowsky, S., Jetter, M. & Ecker, U. K. Using the presidents tweets to understand political diversion in the age of social media. Nat. Commun. 11 , 5764 (2020).

LaCombe, S. J., Tolbert, C. & Mossberger, K. Information and policy innovation in U.S. states. Polit. Res. Q. 75 , 353–365 (2022).

Stroud, N. J. Polarization and partisan selective exposure. J. Commun. 60 , 556–576 (2010).

Van Bavel, J. J., Rathje, S., Harris, E., Robertson, C. & Sternisko, A. How social media shapes polarization. Trends Cogn. Sci. 25 , 913–916 (2021).

Banks, A., Calvo, E., Karol, D. & Telhami, S. #PolarizedFeeds: three experiments on polarization, framing, and social media. Int. J. Press Polit. 26 , 609–634 (2021).

Scheufle, D. A. & Moy, P. Twenty-five years of the spiral of silence: a conceptual review and empirical outlook. Int. J. Public Opin. Res. 12 , 3–28 (2000).

Matthes, J., Knoll, J. & von Sikorski, C. The spiral of silence revisited: a meta-analysis on the relationship between perceptions of opinion support and political opinion expression. Commun. Res. 45 , 3–33 (2018).

Munzert, S. & Ramirez-Ruiz, S. Meta-Analysis of the effects of voting advice applications. Polit. Commun. 38 , 691–706 (2021).

The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2021 (NobelPrize.org, 2022).

Angrist, J. D. & Pischke, J.-S. Mostly Harmless Econometrics (Princeton Univ. Press, 2008).

Angrist, J. D. & Pischke, J.-S. The credibility revolution in empirical economics: how better research design is taking the con out of econometrics. J. Econ. Perspect. 24 , 3–30 (2010).

Zhou, D., Deng, W. & Wu, X. Impacts of internet use on political trust: new evidence from China. Emerg. Mark. Finance Trade 56 , 3235–3251 (2020).

Miner, L. The unintended consequences of internet diffusion: evidence from Malaysia. J. Public Econ. 132 , 66–78 (2015).

Enikolopov, R., Makarin, A. & Petrova, M. Social media and protest participation: evidence from Russia. Econometrica 88 , 1479–1514 (2020).

Bursztyn, L., Egorov, G., Enikolopov, R. & Petrova, M. Social Media and Xenophobia: Evidence from Russia https://www.nber.org/papers/w26567 (NBER, 2019).

Schaub, M. & Morisi, D. Voter mobilisation in the echo chamber: broadband internet and the rise of populism in Europe. Eur. J. Polit. Res. 59 , 752–773 (2020).

Mousavi, R. & Gu, B. The impact of twitter adoption on lawmakers’ voting orientations. Inf. Syst. Res. 30 , 133–153 (2019).

Chen, C., Bai, Y. & Wang, R. Online political efficacy and political participation: a mediation analysis based on the evidence from Taiwan. New Media Soc. 21 , 1667–1696 (2019).

Bekmagambetov, A. et al. Critical social media information flows: political trust and protest behaviour among Kazakhstani college students. Central Asian Surv. 37 , 526–545 (2018).

Lelkes, Y. A bigger pie: the effects of high-speed internet on political behavior. J. Comput. Mediat. Commun. 25 , 199–216 (2020).

Müller, K. & Schwarz, C. Fanning the flames of hate: social media and hate crime. J. Eur. Econ. Assoc. 19 , 2131–2167 (2018).

Bail, C. A. et al. Exposure to opposing views on social media can increase political polarization. Proc. Natl Acad. Sci. USA 115 , 9216–9221 (2018).

Allcott, H., Braghieri, L., Eichmeyer, S. & Gentzkow, M. The welfare effects of social media. Am. Econ. Rev. 110 , 629–676 (2020).

Bond, R. M. et al. A 61-million-person experiment in social influence and political mobilization. Nature 489 , 295–298 (2012).

Asimovic, N., Nagler, J., Bonneau, R. & Tucker, J. A. Testing the effects of Facebook usage in an ethnically polarized setting. Proc. Natl Acad. Sci. USA 118 , e2022819118 (2021).

Wong, S. H.-W. & Wong, M. Y. H. ‘Distant participation’ and youth political attitudes: evidence from a natural experiment. Soc. Sci. Q. 101 , 1489–1512 (2020).

Kleinberg, M. & Lau, R. Googling politics: how offloading affects voting and political knowledge. Polit. Psychol. 42 , 93–110 (2021).

Poy, S. & Schüller, S. Internet and voting in the social media era: evidence from a local broadband policy. Res. Policy 49 , 103861 (2020).

Lee, S. & Xenos, M. Incidental news exposure via social media and political participation: evidence of reciprocal effects. New Media Soc. 24 , 178–201 (2020).

Imai, K., Keele, L. & Tingley, D. A general approach to causal mediation analysis. Psychol. Methods 15 , 309–334 (2010).

Porumbescu, G. Not all bad news after all? exploring the relationship between citizens’ use of online mass media for government information and trust in government. Int. Public Manage. J. 20 , 409–441 (2017).

Zhu, Z., Liu, Y., Kapucu, N. & Peng, Z. Online media and trust in government during crisis: the moderating role of sense of security. Int. J. Disaster Risk Reduct. 50 , 101717 (2020).

Zimmermann, F. & Kohring, M. Mistrust, disinforming news, and vote choice: a panel survey on the origins and consequences of believing disinformation in the 2017 german parliamentary election. Polit. Commun. 37 , 215–237 (2020).

Bucy, E. P. & Groshek, J. Empirical support for the media participation hypothesis: trends across presidential elections, 1992–2012. New Media Soc. 20 , 1889–1909 (2018).

Arlt, D. Who trusts the news media? Exploring the factors shaping trust in the news media in German-speaking Switzerland. Stud. Commun. Sci. 18 , 231–245 (2019).

Park, S., Fisher, C., Flew, T. & Dulleck, U. Global mistrust in news: the impact of social media on trust. Int. J. Media Manage. 22 , 83–96 (2020).

Sabatini, F. & Sarracino, F. Online social networks and trust. Soc. Indic. Res. 142 , 229–260 (2019).

Carrieri, V., Madio, L. & Principe, F. Vaccine hesitancy and (fake) news: quasi-experimental evidence from Italy. Health Econ. 28 , 1377–1382 (2019).

Casara, B., Suitner, C. & Bettinsoli, M. Viral suspicions: vaccine hesitancy in the web 2.0. J. Exp. Psychol. Appl. 25 , 354–371 (2019).

Huber, B., Barnidge, M., Gil de Zuniga, H. & Liu, J. Fostering public trust in science: the role of social media. Public Underst. Sci. 28 , 759–777 (2019).

Placek, M. A. #Democracy: social media use and democratic legitimacy in Central and Eastern Europe. Democratization 24 , 632–650 (2017).

Placek, M. Can the internet aid democratic consolidation? Online news and legitimacy in Central and Eastern Europe. Int. J. Commun. 12 , 2810–2831 (2018).

Geraci, A., Nardotto, M., Reggiani, T. & Sabatini, F. Broadband internet and social capital. J. Public Econ. 206 , 104578 (2022).

Min, G., Yu, Z. & Li, F. Analysis of moral deviation in netnews post-bumping. In Proc. 12th International Conference on Innovation and Management (eds Wang, Y. & Xu, H.) 1222–1226 (Wuhan University of Technology Press, 2015).

Di, C. & Fang, W. New channels, new ways of becoming informed? Examining the acquisition of public affairs knowledge by young people in China. Inf. Dev. 35 , 688–702 (2019).

Wei, K., Lin, Y.-R. & Yan, M. Examining protest as an intervention to reduce online prejudice: a case study of prejudice against immigrants. In Proc. of The Web Conference 2020 2443–2454 (Association for Computing Machinery, 2020).

Alam, A., Adnan, H. M. & Kotamjani, S. S. Examining the impact of using social networks on political knowledge and political attitude by iranian university students. J. Komun. Malays. J. Commun. 35 , 125–140 (2019).

Beaudoin, C. E. The internet’s impact on international knowledge. New Media Soc. 10 , 455–474 (2008).

Ida, R., Saud, M. & Mashud, M. An empirical analysis of social media usage, political learning and participation among youth: a comparative study of Indonesia and Pakistan. Qual. Quant. 54 , 1285–1297 (2020).

Salaudeen, M. & Onyechi, N. Digital media vs mainstream media: exploring the influences of media exposure and information preference as correlates of media credibility. Cogent Arts Humanit. 7 , 1837461 (2020).

Imran, M. S., Fatima, M. & Kosar, G. Connectivism: E-learning of democratic values on social media public spheres. In 2017 International Conference on Information and Communication Technologies 82–89 (IEEE, 2018).

Park, C. & Kaye, B. News engagement on social media and democratic citizenship: direct and moderating roles of curatorial news use in political involvement. Journal. Mass Commun. Q. 95 , 1103–1127 (2018).

Gottfried, J. A., Hardy, B. W., Holbert, R. L., Winneg, K. M. & Jamieson, K. H. The changing nature of political debate consumption: social media, multitasking, and knowledge acquisition. Polit. Commun. 34 , 172–199 (2017).

Kim, D. H. & Kwak, N. Media diversity policies for the public: empirical evidence examining exposure diversity and democratic citizenship. J. Broadcast. Electron. Media 61 , 682–702 (2017).

Beam, M. A., Hutchens, M. J. & Hmielowski, J. D. Clicking vs. sharing: the relationship between online news behaviors and political knowledge. Comput. Hum. Behav. 59 , 215–220 (2016).

Edgerly, S., Thorson, K. & Wells, C. Young citizens, social media, and the dynamics of political learning in the U.S. presidential primary election. Am. Behav. Sci. 62 , 1042–1060 (2018).

Cardenal, A., Galais, C. & Majó-Vázquez, S. Is Facebook eroding the public agenda? Evidence from survey and web-tracking data. Int. J. Public Opin. Res. 31 , 589–608 (2018).

Feezell, J. T. Agenda setting through social media: the importance of incidental news exposure and social filtering in the digital era. Polit. Res. Q. 71 , 482–494 (2018).

Kelly Garrett, R. Social media’s contribution to political misperceptions in U.S. Presidential elections. PLoS ONE 14 , e0213500 (2019).

Lee, S. & Xenos, M. Social distraction? Social media use and political knowledge in two U.S. Presidential elections. Comput. Hum. Behav. 90 , 18–25 (2019).

van Erkel, P. & Van Aelst, P. Why dont we learn from social media? Studying effects of and mechanisms behind social media news use on general surveillance political knowledge. Polit. Commun. 38 , 407–425 (2021).

Lee, S. Connecting social media use with gaps in knowledge and participation in a protest context: the case of candle light vigil in South Korea. Asian J. Commun. 29 , 111–127 (2019).

Cacciatore, M. A. et al. Is Facebook making us dumber? Exploring social media use as a predictor of political knowledge. Journal. Mass Commun. Q. 95 , 404–424 (2018).

Moeller, J., Shehata, A. & Kruikemeier, S. Internet use and political interest: growth curves, reinforcing spirals, and causal effects during adolescence. J. Commun. 68 , 1052–1078 (2018).

Lee, S. Probing the mechanisms through which social media erodes political knowledge: the role of the news-finds-me perception. Mass Commun. Soc. 23 , 810–832 (2020).

Boukes, M. Social network sites and acquiring current affairs knowledge: the impact of Twitter and Facebook usage on learning about the news. J. Inf. Technol. Polit. 16 , 36–51 (2019).

Adam, S., Haussler, T., Schmid-Petri, H. & Reber, U. Coalitions and counter-coalitions in online contestation: an analysis of the German and British climate change debate. New Media Soc. 21 , 2671–2690 (2019).

North, S., Piwek, L. & Joinson, A. Battle for Britain: analyzing events as drivers of political tribalism in Twitter discussions of Brexit. Policy Internet 13 , 185–208 (2020).

Bryson, B. Polarizing the middle: internet exposure and public opinion. Int. J. Sociol. Soc. Policy 40 , 99–113 (2019).

Lee, C., Shin, J. & Hong, A. Does social media use really make people politically polarized? Direct and indirect effects of social media use on political polarization in South Korea. Telemat. Inform. 35 , 245–254 (2018).

Cho, J., Ahmed, S., Keum, H., Choi, Y. & Lee, J. Influencing myself: self-reinforcement through online political expression. Commun. Res. 45 , 83–111 (2018).

Yarchi, M., Baden, C. & Kligler-Vilenchik, N. Political polarization on the digital sphere: a cross-platform, over-time analysis of interactional, positional, and affective polarization on social media. Polit. Commun. 38 , 98–139 (2020).

Workneh, T. Social media, protest, and outrage communication in Ethiopia: toward fractured publics or pluralistic polity?. Inf. Commun. Soc. 24 , 309–328 (2021).

Hawdon, J., Ranganathan, S., Leman, S., Bookhultz, S. & Mitra, T. Social media use, political polarization, and social capital: is social media tearing the U.S. apart? In Social Computing and Social Media. Design, Ethics, User Behavior, and Social Network Analysis (ed. Meiselwitz, G.) 243–260 (Springer, 2020).

Urman, A. News consumption of Russian Vkontakte users: polarization and news avoidance. Int. J. Commun. 13 , 5158–5182 (2019).

Kibet, A. & Ward, S. Socially networked heterogeneity: the influence of WhatsApp as a social networking site on polarisation in Kenya. Afr. Journal. Stud. 39 , 42–66 (2018).

Fletcher, R., Cornia, A. & Nielsen, R. How polarized are online and offline news audiences? A comparative analysis of twelve countries. Int. J. Press Polit. 25 , 169–195 (2020).

Lu, J. & Luo, C. Development consensus in the Internet context: penetration, freedom, and participation in 38 countries. Inf. Dev. 36 , 288–300 (2020).

Lai, M., Tambuscio, M., Patti, V., Ruffo, G. & Rosso, P. Stance polarity in political debates: a diachronic perspective of network homophily and conversations on Twitter. Data Knowl. Eng. 124 , 101738 (2019).

Kobayashi, T., Ogawa, Y., Suzuki, T. & Yamamoto, H. News audience fragmentation in the Japanese Twittersphere. Asian J. Commun. 29 , 274–290 (2019).

Nguyen, A. & Vu, H. Testing popular news discourse on the echo chamber effect: does political polarisation occur among those relying on social media as their primary politics news source? First Monday https://doi.org/10.5210/fm.v24i6.9632 (2019).

Beam, M., Hutchens, M. & Hmielowski, J. Facebook news and (de)polarization: reinforcing spirals in the 2016 US election. Inf. Commun. Soc. 21 , 940–958 (2018).

Levy, R. Social media, news consumption, and polarization: evidence from a field experiment. Am. Econ. Rev. 111 , 831–70 (2021).

Carrella, F. #Populism on Twitter: statistical analysis of the correlation between tweet popularity and populist discursive features. Brno Stud. Engl. 46 , 5–23 (2020).

Serrano, J., Shahrezaye, M., Papakyriakopoulos, O. & Hegelich, S. The rise of Germany’s AfD: a social media analysis. In Proc. 10th International Conference on Social Media and Society 214–223 (ACM, 2019).

Schumann, S., Boer, D., Hanke, K. & Liu, J. Social media use and support for populist radical right parties: assessing exposure and selection effects in a two-wave panel study. Inf. Commun. Soc. 24 , 921–940 (2019).

Schumann, S., Thomas, F., Ehrke, F., Bertlich, T. & Dupont, J. C. Maintenance or change? Examining the reinforcing spiral between social media news use and populist attitudes. Inf. Commun. Soc. https://doi.org/10.1080/1369118X.2021.1907435 (2021).

Heiss, R. & Matthes, J. Stuck in a nativist spiral: content, selection, and effects of right-wing populists communication on Facebook. Polit. Commun. 37 , 303–328 (2020).

Bliuc, A.-M. et al. The effects of local socio-political events on group cohesion in online far-right communities. PLoS ONE 15 , e0230302 (2020).

Schulze, H. Who uses right-wing alternative online media? An exploration of audience characteristics. Polit. Gov. 8 , 6–18 (2020).

Boulianne, S., Koc-Michalska, K. & Bimber, B. Right-wing populism, social media and echo chambers in Western democracies. New Media Soc. 22 , 683–699 (2020).

Jeroense, T., Luimers, J., Jacobs, K. & Spierings, N. Political social media use and its linkage to populist and postmaterialist attitudes and vote intention in the Netherlands. Eur. Political Sci. 21 , 193–215 (2022).

Bosilkov, I. Media populism in Macedonia: right-wing populist style in the coverage of the migrant crisis. Cent. Eur. J. Commun. 12 , 206–223 (2019).

Seebruck, R. Technology and tolerance in Japan: internet use and positive attitudes and behaviors toward foreigners. Soc. Sci. Jpn J. 16 , 279–300 (2013).

Doroshenko, L. Far-right parties in the European Union and media populism: a comparative analysis of 10 countries during European Parliament elections. Int. J. Commun. 12 , 3186–3206 (2018).

Fletcher, R. & Nielsen, R. Are people incidentally exposed to news on social media? A comparative analysis. New Media Soc. 20 , 2450–2468 (2018).

Guess, A. M. (Almost) everything in moderation: new evidence on americans’ online media diets. Am. J. Pol. Sci. https://doi.org/10.1111/ajps.12589 (2021).

Strauss, N., Huber, B. & Gil de Zuniga, H. ‘Yes, i saw it - but didn’t read it…’ a cross-country study, exploring relationships between incidental news exposure and news use across platforms. Digit. Journal. 8 , 1181–1205 (2020).

Yang, T., Majó-Vázquez, S., Nielsen, R. & González-Bailón, S. Exposure to news grows less fragmented with an increase in mobile access. Proc. Natl Acad. Sci. USA 117 , 28678–28683 (2020).

Rivero, G. Preaching to the choir: ideology and following behaviour in social media. Contemp. Soc. Sci. 14 , 54–70 (2019).

Cinelli, M., de Francisci Morales, G., Galeazzi, A., Quattrociocchi, W. & Starnini, M. The echo chamber effect on social media. Proc. Natl Acad. Sci. USA 118 , e2023301118 (2021).

Cota, W., Ferreira, S., Pastor-Satorras, R. & Starnini, M. Quantifying echo chamber effects in information spreading over political communication networks. EPJ Data Sci. 8 , 35 (2019).

Guerrero-Solé, F. & Lopez-Gonzalez, H. Government formation and political discussions in Twitter: an extended model for quantifying political distances in multiparty democracies. Soc. Sci. Comput. Rev. 37 , 3–21 (2019).

Koiranen, I., Koivula, A., Keipi, T. & Saarinen, A. Shared contexts, shared background, shared values - homophily in Finnish parliament members social networks on Twitter. Telemat. Inform. 36 , 117–131 (2019).

Barnidge, M., Macafee, T., Alvarez, G. & Rojas, H. Citizenship and political participation in Colombia: how orientations toward citizenship associate with political and civic behaviors. Int. J. Commun. 8 , 1831–1850 (2014).

Soral, W., Liu, J. & Bilewicz, M. Media of contempt: social media consumption predicts normative acceptance of anti-muslim hate speech and islamoprejudice. Int. J. Conf. Violence https://doi.org/10.4119/ijcv-3774 (2020).

Tornberg, A. & Wahlstrom, M. Unveiling the radical right online: exploring framing and identity in an online anti-immigrant discussion group. Sociol. Forsk. 55 , 267–292 (2018).

Rathje, S., Van Bavel, J. J. & van der Linden, S. Out-group animosity drives engagement on social media. Proc. Natl Acad. Sci. USA 118 , e2024292118 (2021).

Pasquetto, I. V. et al. Tackling misinformation: what researchers could do with social media data. The Harvard Kennedy School Misinformation Review (December 9, 2020).

Moher, D. et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 6 , e1000097 (2009).

Gusenbauer, M. & Haddaway, N. R. Which academic search systems are suitable for systematic reviews or meta-analyses? Evaluating retrieval qualities of Google Scholar, Pubmed, and 26 other resources. Res. Synth. Methods 11 , 181–217 (2020).

Park, I. & Lee, D. Understanding news-sharing across different platforms: the effects of newsworthiness and gratifications from news-sharing. Inf. Res. 25 , 882 (2020).

Machackova, H. & Tkaczyk, M. The effect of media and political beliefs and attitudes on trust in political institutions: a multilevel analysis on data from 21 European countries. Commun. Today 11 , 64–82 (2020).

Magalhaes, P. C., Aldrich, J. H. & Gibson, R. K. New forms of mobilization, new people mobilized? Evidence from the comparative study of electoral systems. Party Polit. 26 , 605–618 (2020).

Kwak, N., Lane, D. S., Zhu, Q., Lee, S. S. & Weeks, B. E. Political rumor communication on instant messaging platforms: relationships with political participation and knowledge. Int. J. Commun. 14 , 5663–5685 (2020).

Powers, E., Koliska, M. & Guha, P. ‘Shouting matches and echo chambers’: perceived identity threats and political self-censorship on social media. Int. J. Commun. 13 , 3630–3649 (2019).

Adegbola, O. & Gearhart, S. Examining the relationship between media use and political engagement: a comparative study among the United States, Kenya, and Nigeria. Int. J. Commun. 13 , 1231–1251 (2019).

Liu, Y.-I. Online and offline communication and political knowledge and participation in presidential campaigns: effects of geographical context. Int. J. Commun. 13 , 1438–1461 (2019).

Taneja, H. & Yaeger, K. Do people consume the news they trust? Incidental news usage and the high-choice media environment. In Proc. 2019 CHI Conference on Human Factors in Computing Systems 1–10 (ACM, 2019).

Goebel, S. & Munzert, S. Political advertising on the Wikipedia marketplace of information. Soc. Sci. Comput. Rev. 36 , 157–175 (2018).

Wen, N. & Wei, R. Examining effects of informational use of social media platforms and social capital on civic engagement regarding genetically modified foods in China. Int. J. Commun. 12 , 3729–3750 (2018).

Quenette, A. M. & Velasquez, A. Shifting demographics: understanding how ethnically diverse networks influence Latinos’ political uses of social media and offline political engagement. Int. J. Commun. 12 , 4839–4859 (2018).

Zhang, N. & Skoric, M. M. Media use and environmental engagement: examining differential gains from news media and social media. Int. J. Commun. 12 , 380–403 (2018).

Mustapha, L. K., Gbonegun, V. O. & Mustapha, M. L. Social media use, social capital, and political participation among Nigerian university students. Tripodos 39 , 127–143 (2016).

Barredo Ibanez, D., Arcila Calderon, C., Arroyave, J. & Silva, R. Influence of social networks in the decision to vote: an exploratory survey on the Ecuadorian electorate. Int. J. E Polit. 6 , 15–34 (2015).

Kim, Y. & Chen, H.-T. Discussion network heterogeneity matters: examining a moderated mediation model of social media use and civic engagement. Int. J. Commun. 9 , 2344–2365 (2015).

Krolo, K. & Puzek, I. Usage of internet social networks and participatory dimensions of social capital of youth - the example of Facebook. Drustvena Istraz. 23 , 383–405 (2014).

Gil de Zuniga, H. & Valenzuela, S. The mediating path to a stronger citizenship: online and offline networks, weak ties, and civic engagement. Commun. Res. 38 , 397–421 (2011).

Popova, O. & Negrov, E. Political communication of youth in the internet space: effects on influence on political consciousness and behavior. In Proc. International Conference "Internet and Modern Society" Vol. 2813 (eds Bolgov. R. V. et al.) 181–195 (RWTH Aachen Univ., 2021).

Panizo-LLedot, A., Torregrosa, J., Bello-Orgaz, G., Thorburn, J. & Camacho, D. Describing alt-right communities and their discourse on Twitter during the 2018 US mid-term elections. Stud. Comput. Intell. 882 , 427–439 (2020).

Serhan, F. & Elareshi, M. New media and hate speech: a study of university students in Jordan. Opcion 36 , 166–184 (2020).

Riikonen, R., Huhtinen, A.-M. & Norri-Sederholm, T. Not a problem for me: young men’s conceptions of their social media use and false information. In Proc. 7th European Conference on Social Media (eds Karpasitis, C. & Varda, C.) 240–245 (Academic Conferences International, 2020). https://www.scopus.com/inward/record.uri?eid=2-s2.0-85097756848&doi=10.34190%2fESM.20.031&partnerID=40&md5=debc7038da814e1795cf153561edf5a3

Ida, R., Saud, M. & Mashud, M. Persistence of social media on political activism and engagement among Indonesian and Pakistani youths. Int. J. Web Based Communities 16 , 378–395 (2020).

Kim, H., Kim, Y. & Lee, D. Understanding the role of social media in political participation: integrating political knowledge and bridging social capital from the social cognitive approach. Int. J. Commun. 14 , 4803–4824 (2020).

Steffan, D. Visual self-presentation strategies of political candidates on social media platforms: a comparative study. Int. J. Commun. 14 , 3096–3118 (2020).

Supovitz, J., Kolouch, C. & Daly, A. The social psychology of homophily: the collective sentiments of education advocacy groups. Teach. Coll. Rec. 122 , 49–66 (2020).

Zannettou, S., Finkelstein, J., Bradlyn, B. & Blackburn, J. A quantitative approach to understanding online antisemitism. In Proc. 14th International AAAI Conference on Web and Social Media 786–797 (AAAI Press, 2020).

Halpern, D., Valenzuela, S., Katz, J. & Miranda, J. From belief in conspiracy theories to trust in others: which factors influence exposure, believing and sharing fake news. Lecture Notes in Computer Science 11578 , 217–232 (2019).

Jones, M. Propaganda, fake news, and fake trends: the weaponization of Twitter bots in the Gulf crisis. Int. J. Commun. 13 , 1389–1415 (2019).

Jiang, S., Robertson, R. & Wilson, C. Bias misperceived: the role of partisanship and misinformation in YouTube comment moderation. In Proc. 13th International Conference on Web and Social Media 278–289 (2019). https://www.scopus.com/inward/record.uri?eid=2-s2.0-85070399362&partnerID=40&md5=2d70fd6bbdf3f70d6f91743032db1880

Quintanilha, T., Da Silva, M. & Lapa, T. Fake news and its impact on trust in the news. Using the Portuguese case to establish lines of differentiation. Commun. Soc. 32 , 17–33 (2019).

Emamjomehzadeh, S., Masoudnia, H. & Rahbarqazi, M. The role of online social media in peoples political orientations and tendency to participate in presidential elections. Teor. Praksa 55 , 666–683 (2018).

Bodrunova, S., Blekanov, I. & Kukarkin, M. Multi-dimensional echo chambers: language and sentiment structure of Twitter discussions on the Charlie Hebdo case. Commun. Comput. Inf. Sci. 850 , 393–400 (2018).

Mohamad, B., Dauda, S. & Halim, H. Youth offline political participation: trends and role of social media. J. Komun. Malays. J. Commun. 34 , 172–192 (2018).

Lee, S. The role of social media in protest participation: the case of candlelight vigils in South Korea. Int. J. Commun. 12 , 1523–1540 (2018).

Herrero-Jiménez, B., Carratalá, A. & Berganza, R. Violent conflicts and the new mediatization: the impact of social media on the European parliamentary agenda regarding the Syrian war. Commun. Soc. 31 , 141–157 (2018).

Xenos, M. et al. News media use and the informed public in the digital age. Int. J. Commun. 12 , 706–724 (2018).

Garimella, K., Smith, T., Weiss, R. & West, R. Political polarization in online news consumption. In Proc. International AAAI Conference on Web and Social Media Vol. 15, 152–162 (AAAI Press, 2021).

Pagoto, L. & Longhi, R. Platformization, techno populism and disintermediation of sources in attacks on journalism on Instagram. Chasqui Rev. Latinoam. Comun. 147 , 181–200 (2021).

Gil de Zuniga, H., Borah, P. & Goyanes, M. How do people learn about politics when inadvertently exposed to news? Incidental news paradoxical Direct and indirect effects on political knowledge. Comput. Hum. Behav. https://doi.org/10.1016/j.chb.2021.106803 (2021).

Olise, F. Level of Acceptance of News Stories on Social Media Platforms Among Youth in Nigeria. J. Komun. Malays. J. Commun. 37 , 210–225 (2021).

Santini, R., Salles, D. & Tucci, G. When machine behavior targets future voters: the use of social bots to test narratives for political campaigns in Brazil. Int. J. Commun. 15 , 1220–1243 (2021).

Choi, J. Cross-cutting scanning, integrating, and interacting: dimensions of cross-cutting exposure on social media and political participation. Int. J. Commun. 15 , 1595–1616 (2021).

Gerosa, T., Gui, M., Hargittai, E. & Nguyen, M. (Mis)informed during COVID-19: how education level and information sources contribute to knowledge gaps. Int. J. Commun. 15 , 2196–2217 (2021).

Villanueva, C. & Toscano, G. Legitimation of hate and political violence through memetic images: the Bolsonaro campaign. Commun. Soc. 34 , 449–468 (2021).

Ahmed, M., Riaz, M., Qamar, M. & Asghar, R. Fake news shared on WhatsApp during Covid-19: an analysis of groups and statuses in Pakistan. Media Educ. Mediaobrazovanie 1 , 4–17 (2021).

Gorski, L. C. & Thomas, F. Staying tuned or tuning out? A longitudinal analysis of news-avoiders on the micro and macro-level. Commun. Res. https://doi.org/10.1177/00936502211025907 (2021).

Enders, A. M. et al. The relationship between social media use and beliefs in conspiracy theories and misinformation. Polit. Behav. https://doi.org/10.1007/s11109-021-09734-6 (2021).

Müller, P. & Bach, R. L. Populist alternative news use and its role for elections: web-tracking and survey evidence from two campaign periods. New Media Soc. https://doi.org/10.1177/14614448211032969 (2021).

Ozeren, S., Cubukcu, S. & Cash, G. Exposure to extremist content and public sympathy for ISIS. Stud. Confl. Terror. https://doi.org/10.1080/1057610X.2021.1965728 (2021).

Hashemi, M. Discovering social media topics and patterns in the coronavirus and election era. J. Inf. Commun. 20 , 1–17 (2022).

CAS   Google Scholar  

Wolfowicz, M., Weisburd, D. & Hasisi, B. Examining the interactive effects of the filter bubble and the echo chamber on radicalization. J. Exp. Criminol. https://link.springer.com/10.1007/s11292-021-09471-0 (2021).

Yamamoto, M. & Yang, F. Does news help us become knowledgeable or think we are knowledgeable? Examining a linkage of traditional and social media use with political knowledge. J. Inf. Technol. Polit. 19 , 269–283 (2022).

Goyanes, M., Borah, P. & Gil de Zúñiga, H. Social media filtering and democracy: effects of social media news use and uncivil political discussions on social media unfriending. Comput. Hum. Behav. 120 , 106759 (2021).

Choli, M. & Kuss, D. J. Perceptions of blame on social media during the coronavirus pandemic. Comput. Hum. Behav. 124 , 106895 (2021).

Cano-Orón, L., Calvo, D., Llorca-Abad, G. & Mestre-Pérez, R. Media crisis and disinformation: the participation of digital newspapers in the dissemination of a denialist hoax. Prof. Inf. https://doi.org/10.3145/epi.2021.jul.12 (2021). https://revista.profesionaldelainformacion.com/index.php/EPI/article/view/86394/version/4386

Kim, B., Cooks, E. & Kim, S.-K. Exploring incivility and moral foundations toward Asians in English-speaking tweets in hate crime-reporting cities during the COVID-19 pandemic. Internet Res. 32 , 362–378 (2022).

Melki, J. et al. Mitigating infodemics: the relationship between news exposure and trust and belief in COVID-19 fake news and social media spreading. PLoS ONE 16 , e0252830 (2021).

Martin, J. D. & Hassan, F. Testing classical predictors of public willingness to censor on the desire to block fake news online. Convergence 28 , 867–887 (2022).

Bermes, A. Information overload and fake news sharing: a transactional stress perspective exploring the mitigating role of consumers’ resilience during COVID-19. J. Retail. Consum. Serv. 61 , 102555 (2021).

Shin, S. H., Ji, H. & Lim, H. Heterogeneity in preventive behaviors during COVID-19: health risk, economic insecurity, and slanted information. Soc. Sci. Med. 278 , 113944 (2021).

Nah, S., Lee, S. & Liu, W. Community storytelling network, expressive digital media use, and civic engagement. Commun. Res. 49 , 327–352 (2022).

Dias, T., von Bülow, M. & Gobbi, D. Populist framing mechanisms and the rise of right-wing activism in Brazil. Latin Am. Polit. Soc. 63 , 69–92 (2021).

Sihombing, S. O. & Pramono, R. The integration of social media to the theory of planned behavior: a case study in Indonesia. J. Asian Finance Econ. Bus. 8 , 445–454 (2021).

Rao, A. et al. Political partisanship and antiscience attitudes in online discussions about COVID-19: Twitter content analysis. J. Med. Internet Res. 23 , e26692 (2021).

Visentin, M., Tuan, A. & Di Domenico, G. Words matter: how privacy concerns and conspiracy theories spread on twitter. Psychol. Market. 38 , 1828–1846 (2021).

Osmundsen, M., Bor, A., Vahlstrup, P. B., Bechmann, A. & Petersen, M. B. Partisan polarization is the primary psychological motivation behind political fake news sharing on Twitter. Am. Polit. Sci. Rev. 115 , 999–1015 (2021).

Herrera-Peco, I. et al. Antivaccine movement and COVID-19 negationism: a content analysis of Spanish-written messages on Twitter. Vaccines 9 , 656 (2021).

Nazar, S. & Pieters, T. Plandemic revisited: a product of planned disinformation amplifying the COVID-19 infodemic. Front. Public Health 9 , 649930 (2021).

Hollewell, G. F. & Longpré, N. Radicalization in the social media era: understanding the relationship between self-radicalization and the internet. Int. J. Offender Ther. Comp. Criminol. 66 , 896–913 (2022).

Tal-Or, N., Cohen, J., Tsfati, Y. & Gunther, A. C. Testing causal direction in the influence of presumed media influence. Commun. Res. 37 , 801–824 (2010).

Weeks, B. E., Menchen-Trevino, E., Calabrese, C., Casas, A. & Wojcieszak, M. Partisan media, untrustworthy news sites, and political misperceptions. New Media Soc. https://doi.org/10.1177/14614448211033300 (2021).

Gil de Zúñiga, H., Barnidge, M. & Diehl, T. Political persuasion on social media: a moderated moderation model of political discussion disagreement and civil reasoning. Inf. Soc. 34 , 302–315 (2018).

Chayinska, M., Miranda, D. & González, R. A longitudinal study of the bidirectional causal relationships between online political participation and offline collective action. Comput. Hum. Behav. 121 , 106810 (2021).

Criss, S. et al. Advocacy, hesitancy, and equity: exploring U.S. race-related discussions of the COVID-19 vaccine on Twitter. Int. J. Environ. Res. Public Health 18 , 5693 (2021).

Onat, I., Guler, A., Kula, S. & Bastug, M. F. Fear of terrorism and fear of violent crimes in the United States: a comparative analysis. Crime Delinq. https://doi.org/10.1177/00111287211036130 (2021).

Chen, H.-T. Second screening and the engaged public: the role of second screening for news and political expression in an O-S-R-O-R model. J. Mass Commun. Q. 98 , 526–546 (2021).

Lin, W.-Y., Cheong, P., Kim, Y.-C. & Jung, J.-Y. Becoming citizens: youths civic uses of new media in five digital cities in East Asia. J. Adolesc. Res. 25 , 839–857 (2010).

Young, L. E. Mobilization under threat: emotional appeals and pro-opposition political participation online. Polit. Behav. https://link.springer.com/10.1007/s11109-021-09711-z (2021).

Wang, D. & Qian, Y. Echo chamber effect in rumor rebuttal discussions about COVID-19 in China: social media content and network analysis study. J. Med. Internet Res. 23 , e27009 (2021).

Diehl, T., Huber, B., Gil de Zúñiga, H. & Liu, J. Social media and beliefs about climate change: a cross-national analysis of news use, political ideology, and trust in science. Int. J. Public Opin. Res. 33 , 197–213 (2021).

Biancovilli, P., Makszin, L. & Jurberg, C. Misinformation on social networks during the novel coronavirus pandemic: a quali-quantitative case study of Brazil. BMC Public Health 21 , 1200 (2021).

Apuke, O. D. & Omar, B. Social media affordances and information abundance: enabling fake news sharing during the COVID-19 health crisis. Health Informatics J. 27 , 146045822110214 (2021).

Pérez-Curiel, C., Rivas-de Roca, R. & García-Gordillo, M. Impact of Trumps digital rhetoric on the US elections: a view from worldwide far-right populism. Soc. Sci. 10 , 152 (2021).

Elliott, T. & Earl, J. Online protest participation and the digital divide: modeling the effect of the digital divide on online petition-signing. New Media Soc. 20 , 698–719 (2018).

Hawkins, I. & Saleem, M. Rise UP! A content analytic study of how collective action is discussed within White nationalist videos on YouTube. New Media Soc. https://doi.org/10.1177/14614448211040520 (2021).

Gorodnichenko, Y., Pham, T. & Talavera, O. Social media, sentiment and public opinions: evidence from #Brexit and #USElection. Eur. Econ. Rev. 136 , 103772 (2021).

Chan, M., Chen, H.-T. & Lee, F. L. F. Cross-cutting discussion on social media and online political participation: a cross-national examination of information seeking and social accountability explanations. Soc. Media Soc . https://doi.org/10.1177/20563051211035697 (2021).

Mari, S. et al. Conspiracy theories and institutional trust: examining the role of uncertainty avoidance and active social media use. Polit. Psychol. 43 , 277–296 (2022).

Kopacheva, E. How the Internet has changed participation: exploring distinctive preconditions of online activism. Commun. Soc. 34 , 67–85 (2021).

Gavazza, A., Nardotto, M. & Valletti, T. Internet and politics: evidence from U.K. local elections and local government policies. Rev. Econ. Stud. 86 , 2092–2135 (2019).

De Coninck, D. et al. Beliefs in conspiracy theories and misinformation about COVID-19: comparative perspectives on the role of anxiety, depression and exposure to and trust in information sources. Front. Psychol. 12 , 646394 (2021).

Wang, L. Race, social media news use, and political participation. J. Inf. Technol. Polit. 19 , 83–97 (2022).

Basch, C. E., Basch, C. H., Hillyer, G. C., Meleo-Erwin, Z. C. & Zagnit, E. A. YouTube videos and informed decision-making about COVID-19 vaccination: successive sampling study. JMIR Public Health Surveill. 7 , e28352 (2021).

Weeks, B. E., Lane, D. S. & Hahn, L. B. Online incidental exposure to news can minimize interest-based political knowledge gaps: evidence from two U.S. elections. Int. J. Press Polit. 27 , 243–262 (2022).

Valenzuela, S., Halpern, D., Katz, J. E. & Miranda, J. P. The paradox of participation versus misinformation: social media, political engagement, and the spread of misinformation. Digit. Journal. 7 , 802–823 (2019).

Onuch, O., Mateo, E. & Waller, J. G. Mobilization, mass perceptions, and (dis)information: new and old media consumption patterns and protest. Soc. Media Soc. https://doi.org/10.1177/2056305121999656 (2021).

Hernandez, R. G., Hagen, L., Walker, K., OLeary, H. & Lengacher, C. The COVID-19 vaccine social media infodemic: healthcare providers missed dose in addressing misinformation and vaccine hesitancy. Hum. Vaccin. Immunother. 17 , 2962–2964 (2021).

Ruijgrok, K. Illusion of control: how internet use generates anti-regime sentiment in authoritarian regimes. Contemp. Polit. 27 , 247–270 (2021).

Brandtzaeg, P. B. Facebook is no “Great equalizer”: a big data approach to gender differences in civic engagement across countries. Soc. Sci. Comput. Rev. 35 , 103–125 (2017).

Rodrguez-Virgili, J., Serrano-Puche, J. & Fernández, C. B. Digital disinformation and preventive actions: perceptions of users from Argentina, Chile, and Spain. Media Commun. 9 , 323–337 (2021).

Suhay, E., Blackwell, A., Roche, C. & Bruggeman, L. Forging bonds and burning bridges: polarization and incivility in blog discussions about occupy wall street. Am. Polit. Res. 43 , 643–679 (2015).

Lilleker, D., Koc-Michalska, K. & Bimber, B. Women learn while men talk? Revisiting gender differences in political engagement in online environments. Inf. Commun. Soc. 24 , 2037–2053 (2021).

Schulz, W. Spiraleffekte in der neuen Medienwelt: Wählermobilisierung und die Nutzung politischer Online- und Offline-Information im Bundestagswahlkampf 2013. Stud. Commun. Media 8 , 77–114 (2019).

Tolbert, C. J. & McNeal, R. S. Unraveling the effects of the internet on political participation? Polit. Res. Q. 56 , 175–185 (2003).

Valenzuela, S., Halpern, D. & Araneda, F. A downward spiral? A panel study of misinformation and media trust in Chile. Int. J. Press Polit. 27 , 353–373 (2022).

Choi, J. & Lee, J. Enthusiasm toward the other side matters: emotion and willingness to express disagreement in social media political conversation. Soc. Sci. J. 1-17 https://doi.org/10.1080/03623319.2021.1949548 (2021). https://www.tandfonline.com/doi/full/10.1080/03623319.2021.1949548

Jennings, W. et al. Lack of trust, conspiracy beliefs, and social media use predict COVID-19 vaccine hesitancy. Vaccines 9 , 593 (2021).

Nagayoshi, K. The political orientation of Papanese online right-wingers. Pac. Aff. 94 , 5–32 (2021).

Kahne, J. & Bowyer, B. The political significance of social media activity and social networks. Polit. Commun. 35 , 470–493 (2018).

Liu, W., Chen, N.-T. N., Ognyanova, K., Nah, S. & Ball-Rokeach, S. Connecting with hyperlocal news website: cause or effect of civic participation?. Am. Behav. Sci. 62 , 1022–1041 (2018).

Housholder, E., Watson, B. R. & LoRusso, S. Does political advertising lead to online information seeking? a real-world test using Google search data. J. Broadcast. Electron. Media 62 , 337–353 (2018).

Shahin, S., Saldaña, M. & Gil de Zúñiga, H. Peripheral elaboration model: the impact of incidental news exposure on political participation. J. Inf. Technol. Polit. 18 , 148–163 (2021).

Haenschen, K. & Jennings, J. Mobilizing millennial voters with targeted internet advertisements: a field experiment. Polit. Commun. 36 , 357–375 (2019).

Sakya, S. M. et al. The impact of COVID-19-related changes in media consumption on public knowledge: results of a cross-sectional survey of Pennsylvania adults. Curr. Med. Res. Opin. 37 , 911–915 (2021).

Sridhar, D. & Getoor, L. Estimating causal effects of tone in online debates. In Proc. 28th International Joint Conference on Artificial Intelligence 1872–1878 (International Joint Conferences on Artificial Intelligence Organization, 2019).

Patra, R. K. & Pandey, N. Disinformation on novel coronavirus (COVID 19): a content analysis of news published on fact checking sites in India. DESIDOC J. Libr. Inf. Technol. 41 , 275–283 (2021).

Shin, J. How do partisans consume news on social media? A comparison of self-reports with digital trace measures among Twitter users. Soc. Media Soc. https://doi.org/10.1177/2056305120981039 (2020).

Praprotnik, K., Perlot, F., Ingruber, D. & Filzmaier, P. Soziale Medien als politischer Informationskanal. Austrian J. Polit. Sci. https://webapp.uibk.ac.at/ojs/index.php/OEZP/article/viewFile/2726/2291 (2019).

Bail, C. A. et al. Assessing the Russian Internet Research Agency’s impact on the political attitudes and behaviors of American Twitter users in late 2017. Proc. Natl Acad. Sci. USA 117 , 243–250 (2020).

Batool, S. H., Ahmed, W., Mahmood, K. & Saeed, H. Twitter dialogue: an analysis of Pakistani politicians’ information sharing. Inf. Discov. Deliv. 50 , 64–74 (2022).

Siongers, J., Keppens, G., Spruyt, B. & Van Droogenbroeck, F. On the digital lane to citizenship? Patterns of internet use and civic engagement amongst Flemish adolescents and young adults. J. Soc. Sci. Educ. https://doi.org/10.4119/jsse-901 (2019).

Linvill, D. L., Boatwright, B. C., Grant, W. J. & Warren, P. L. THE RUSSIANS ARE HACKING MY BRAIN! investigating Russia’s internet research agency twitter tactics during the 2016 United States presidential campaign. Comput. Hum. Behav. 99 , 292–300 (2019).

Chan, M.-pS. et al. Legacy and social media respectively influence risk perceptions and protective behaviors during emerging health threats: a multi-wave analysis of communications on Zika virus cases. Soc. Sci. Med. 212 , 50–59 (2018).

Koivula, A., Kaakinen, M., Oksanen, A. & Räsänen, P. The role of political activity in the formation of online identity bubbles. Policy Internet 11 , 396–417 (2019).

Kleinnijenhuis, J., van Hoof, A. M. J. & van Atteveldt, W. The combined effects of mass media and social media on political perceptions and preferences. J. Commun. 69 , 650–673 (2019).

Bowman, W. M. & Bowman, J. D. Censorship or self-control? Hate speech, the state and the voter in the Kenyan election of 2013. J. Mod. Afr. 54 , 495–531 (2016).

Kim, H. H. & Lim, C. From virtual space to public space: the role of online political activism in protest participation during the Arab Spring. Int. J. Comp. Sociol. 60 , 409–434 (2019).

Matuszewski, P. & Szabó, G. Are echo chambers based on partisanship? Twitter and political polarity in Poland and Hungary. Soc. Media Soc . https://doi.org/10.1177/2056305119837671 (2019).

Lee, F. L. F., Lee, P. S. N., So, C. Y., Leung, L. & Chan, M. C. Conditional impact of Facebook as an information source on political opinions: the case of political reform in Hong Kong. Asian J. Polit. Sci. 25 , 365–382 (2017).

Casas, A. & Williams, N. W. Images that matter: online protests and the mobilizing role of pictures. Polit. Res. Q. 72 , 360–375 (2019).

Xiong, J., Feng, X. & Tang, Z. Understanding user-to-user interaction on government microblogs: an exponential random graph model with the homophily and emotional effect. Inf. Process. Manage. 57 , 102229 (2020).

Nguyen, T. T. et al. Evaluating associations between area-level Twitter-expressed negative racial sentiment, hate crimes, and residents’ racial prejudice in the United States. SSM Popul. Health 13 , 100750 (2021).

Gondal, M. T., Munir, A., Shabir, G. & Naz, A. Facebook and propaganda: following politics on Facebook and its impact on political behaviors of youth. Clin. Soc. Work Health Interv. 10 , 27–33 (2019).

Zang, L., Xiong, F. & Gao, Y. Reversing the U: new evidence on the internet and democracy relationship. Soc. Sci. Comput. Rev. 37 , 295–314 (2019).

Ohme, J. Updating citizenship? The effects of digital media use on citizenship understanding and political participation. Inf. Commun. Soc. 22 , 1903–1928 (2019).

Corbu, N., Oprea, D.-A., Negrea-Busuioc, E. & Radu, L. They can’t fool me, but they can fool the others! Third person effect and fake news detection. Eur. J. Commun. 35 , 165–180 (2020).

Germann, M. & Gemenis, K. Getting out the vote with voting advice applications. Polit. Commun. 36 , 149–170 (2019).

Neely, S., Eldredge, C. & Sanders, R. Health information seeking behaviors on social media during the COVID-19 pandemic among American social networking site users: survey study. J. Med. Internet Res. 23 , e29802 (2021).

Inguanzo, I., Zhang, B. & Gil de Zúñiga, H. Online cultural backlash? Sexism and political user-generated content. Inf. Commun. Soc. 24 , 2133–2152 (2021).

Chekol, M. A., Moges, M. A. & Nigatu, B. A. Social media hate speech in the walk of Ethiopian political reform: analysis of hate speech prevalence, severity, and natures. Inf. Commun. Soc. https://doi.org/10.1080/1369118X.2021.1942955 (2021).

Valenzuela, S., Bachmann, I. & Bargsted, M. The personal is the political? What do WhatsApp users share and how it matters for news knowledge, polarization and participation in Chile. Digit. Journal. 9 , 155–175 (2021).

Van Duyn, E., Peacock, C. & Stroud, N. J. The gender gap in online news comment sections. Soc. Sci. Comput. Rev. 39 , 181–196 (2021).

Wiedlitzka, S., Prati, G., Brown, R., Smith, J. & Walters, M. A. Hate in word and deed: the temporal association between online and offline islamophobia. J. Quant. Criminol. https://doi.org/10.1007/s10940-021-09530-9 (2021).

Oh, H. J., Lor, Z. & Choi, J. News repertoires and political information efficacy: focusing on the mediating role of perceived news overload. SAGE Open https://doi.org/10.1177/2158244020988685 (2021).

Siegel, A. A. et al. Trumping hate on Twitter? Online hate speech in the 2016 U.S. election campaign and its aftermath. Q. J. Polit. Sci. 16 , 71–104 (2021).

Kruikemeier, S. How political candidates use Twitter and the impact on votes. Comput. Hum. Behav. 34 , 131–139 (2014).

van Erkel, P. F. A. & Van Aelst, P. Why dont we learn from social media? Studying effects of and mechanisms behind social media news use on general surveillance political knowledge. Polit. Commun. 38 , 407–425 (2021).

Chan, N. K. Political inequality in the digital world: the puzzle of Asian American political participation online. Polit. Res. Q. 74 , 882–898 (2021).

Wang, X. & Kobayashi, T. Nationalism and political system justification in China: differential effects of traditional and new media. Chin. J. Commun. 14 , 139–156 (2021).

Freudenthaler, R. & Wessler, H. Mapping emerging and legacy outlets online by their democratic functions—agonistic, deliberative, or corrosive?. Int. J. Press Polit. 27 , 417–438 (2022).

Zhou, C., Xiu, H., Wang, Y. & Yu, X. Characterizing the dissemination of misinformation on social media in health emergencies: an empirical study based on COVID-19. Inf. Process. Manage. 58 , 102554 (2021).

Hajj, N., McEwan, P. J. & Turkington, R. Women, information ecology, and political protest in the Middle East. Mediterr. Polit. 24 , 62–83 (2019).

Zhu, A. Y. F., Chan, A. L. S. & Chou, K. L. Creative social media use and political participation in young people: the moderation and mediation role of online political expression. J. Adolesc. 77 , 108–117 (2019).

Towner, T. L. & Muñoz, C. L. Baby boom or bust? The new media effect on political participation. J. Polit. Market. 17 , 32–61 (2018).

Licari, P. R. Sharp as a fox: are foxnews.com visitors less politically knowledgeable?. Am. Polit. Res. 48 , 792–806 (2020).

Krueger, B. S. A comparison of conventional and internet political mobilization. Am. Polit. Res. 34 , 759–776 (2006).

Makhortykh, M., de Vreese, C., Helberger, N., Harambam, J. & Bountouridis, D. We are what we click: understanding time and content-based habits of online news readers. New Media Soc. 23 , 2773–2800 (2021).

Sydnor, E. Platforms for incivility: examining perceptions across different media formats. Polit. Commun. 35 , 97–116 (2018).

Piazza, J. A. & Guler, A. The online caliphate: internet usage and ISIS support in the Arab World. Terror. Polit. Violence 33 , 1256–1275 (2021).

Dahlgren, P. M., Shehata, A. & Strömbäck, J. Reinforcing spirals at work? Mutual influences between selective news exposure and ideological leaning. Eur. J. Commun. 34 , 159–174 (2019).

Garrett, R. K. Politically motivated reinforcement seeking: reframing the selective exposure debate. J. Commun. 59 , 676–699 (2009).

Milani, E., Weitkamp, E. & Webb, P. The visual vaccine debate on Twitter: a social network analysis. Media Commun. 8 , 364–375 (2020).

Nguyen, A. & Western, M. Socio-structural correlates of online news and information adoption/use: implications for the digital divide. J. Sociol. 43 , 167–185 (2007).

Kane, B. & Luo, J. Do the communities we choose shape our political beliefs? a study of the politicization of topics in online social groups. In 2018 IEEE Int. Conference on Big Data (Big Data) 3665–3671 (IEEE, 2018). https://ieeexplore.ieee.org/document/8622535/

Wollebæ k, D., Karlsen, R., Steen-Johnsen, K. & Enjolras, B. Anger, fear, and echo chambers: the emotional basis for online behavior. Soc. Media Soc . https://doi.org/10.1177/2056305119829859 (2019).

Scheffauer, R., Goyanes, M. & Gil de Zúñiga, H. Beyond social media news use algorithms: how political discussion and network heterogeneity clarify incidental news exposure. Online Inf. Rev. 45 , 633–650 (2021).

Chan, M., Chen, H.-T. & Lee, F. L. F. Examining the roles of political social network and internal efficacy on social media news engagement: a comparative study of six Asian countries. Int. J. Press Polit. 24 , 127–145 (2019).

Miao, H. Media use and political participation in China: taking three national large-n surveys as examples. Asian J. Public Opin. Res. 7 , 1–22 (2019).

Peterson, E., Goel, S. & Iyengar, S. Partisan selective exposure in online news consumption: evidence from the 2016 presidential campaign. Polit. Sci. Res. Methods 9 , 242–258 (2021).

Hjorth, F. & Adler-Nissen, R. Ideological asymmetry in the reach of pro-Russian digital disinformation to United States audiences. J. Commun. 69 , 168–192 (2019).

Dozier, D. M., Shen, H., Sweetser, K. D. & Barker, V. Demographics and Internet behaviors as predictors of active publics. Public Relat. Rev. 42 , 82–90 (2016).

Asker, D. & Dinas, E. Thinking fast and furious: emotional intensity and opinion polarization in online media. Public Opin. Q. 83 , 487–509 (2019).

Sugihartati, R., Suyanto, B. & Sirry, M. The shift from consumers to prosumers: susceptibility of young adults to radicalization. Soc. Sci. 9 , 40 (2020).

Johnson, T. J., Kaye, B. K. & Lee, A. M. Blinded by the spite? Path model of political attitudes, selectivity, and social media. Atlantic J. Commun. 25 , 181–196 (2017).

Yang, J. & Grabe, M. E. Knowledge acquisition gaps: a comparison of print versus online news sources. New Media Soc. 13 , 1211–1227 (2011).

Bode, L., Vraga, E. K., Borah, P. & Shah, D. V. A new space for political behavior: political social networking and its democratic consequences. J. Comput. Mediat. Commun. 19 , 414–429 (2014).

Shim, K. & Oh, S.-K. K. Who creates the bandwagon? The dynamics of fear of isolation, opinion congruency and anonymity-preference on social media in the 2017 South Korean presidential election. Comput. Hum. Behav. 86 , 181–189 (2018).

Rosenbusch, H., Evans, A. M. & Zeelenberg, M. Multilevel emotion transfer on Youtube: disentangling the effects of emotional contagion and homophily on video audiences. Soc. Psychol. Personal. Sci. 10 , 1028–1035 (2019).

Munger, K., Luca, M., Nagler, J. & Tucker, J. The (null) effects of clickbait headlines on polarization, trust, and learning. Public Opin. Q. 84 , 49–73 (2020).

Eady, G., Nagler, J., Guess, A., Zilinsky, J. & Tucker, J. A. How many people live in political bubbles on social media? Evidence from linked survey and Twitter data. SAGE Open https://doi.org/10.1177/2158244019832705 (2019).

Guerrero-Solé, F. Interactive behavior in political discussions on Twitter: politicians, media, and citizens patterns of interaction in the 2015 and 2016 electoral campaigns in Spain. Soc. Media Soc . https://doi.org/10.1177/2056305118808776 (2018).

Theocharis, Y., Moor, J. & Deth, J. W. Digitally networked participation and lifestyle politics as new modes of political participation. Policy Internet 13 , 30–53 (2021).

Robles, J. M., Velez, D., De Marco, S., Rodríguez, J. T. & Gomez, D. Affective homogeneity in the Spanish general election debate. A comparative analysis of social networks political agents. Inf. Commun. Soc. 23 , 216–233 (2020).

Costello, M. & Hawdon, J. Who are the online extremists among us? Sociodemographic characteristics, social networking, and online experiences of those who produce online hate materials. Violence Gend. 5 , 55–60 (2018).

Vaccari, C. & Valeriani, A. Digital political talk and political participation: comparing established and third wave democracies. SAGE Open https://doi.org/10.1177/2158244018784986 (2018).

Park, C. S. & Karan, K. Unraveling the relationships between smartphone use, exposure to heterogeneity, political efficacy, and political participation: a mediation model approach. Asian J. Commun. 24 , 370–389 (2014).

David, C. C., San Pascual, M. R. S. & Torres, M. E. S. Reliance on Facebook for news and its influence on political engagement. PLoS ONE 14 , e0212263 (2019).

Arshad, S. & Khurram, S. Can governments presence on social media stimulate citizens online political participation? Investigating the influence of transparency, trust, and responsiveness. Gov. Inf. Q. 37 , 101486 (2020).

Machackova, H. & Šerek, J. Does clicking matter? The role of online participation in adolescents’ civic development. Cyberpsychology https://doi.org/10.5817/CP2017-4-5 (2017). https://cyberpsychology.eu/article/view/8741

Jeroense, T., Luimers, J., Jacobs, K. & Spierings, N. Political social media use and its linkage to populist and postmaterialist attitudes and vote intention in the Netherlands. Eur. Polit. Sci. 21 , 193–215 (2022).

Bosi, L., Lavizzari, A. & Portos, M. The impact of intolerance on young peoples online political participation. Politics 42 , 95–127 (2022).

Sommariva, S., Vamos, C., Mantzarlis, A., Uyên-Loan Ɖào, L. U.-L. & Martinez Tyson, D. Spreading the (fake) news: exploring health messages on social media and the implications for health professionals using a case study. Am. J. Health. Educ. 49 , 246–255 (2018).

Kim, D. H., Jones-Jang, S. M. & Kenski, K. Why do people share political information on social media? Digit. Journal. 9 , 1123–1140 (2021).

Kulshrestha, J. et al. Search bias quantification: investigating political bias in social media and web search. Inf. Retr. J. 22 , 188–227 (2019).

Eddington, S. M. The communicative constitution of hate organizations online: a semantic network analysis of “Make America Great Again”. Soc. Media Soc . https://doi.org/10.1177/2056305118790763 (2018).

Feezell, J. T. & Ortiz, B. I saw it on Facebook: an experimental analysis of political learning through social media. Inf. Commun. Soc. 24 , 1283–1302 (2021).

Ejaz, W., Ittefaq, M., Seo, H. & Naz, F. Factors associated with the belief in COVID-19 related conspiracy theories in Pakistan. Health Risk Soc. 23 , 162–178 (2021).

Omotayo, F. & Folorunso, M. B. Use of social media for political participation by youths. JeDEM 12 , 132–157 (2020).

Hasangani, S. Religious identification on Facebook visuals and (online) out-group intolerance: experimenting the Sri Lankan case. J. Asian Afr. Stud. 57 , 247–268 (2022).

Mueller-Herbst, J. M., Xenos, M. A., Scheufele, D. A. & Brossard, D. Saw it on Facebook: the role of social media in facilitating science issue awareness. Soc. Media Soc . https://doi.org/10.1177/2056305120930412 (2020).

Lu, Y., Lee, J. K. & Kim, E. Network characteristics matter in politics on Facebook: evidence from a US national survey. Online Inf. Rev. 42 , 372–386 (2018).

Forati, A. M. & Ghose, R. Geospatial analysis of misinformation in COVID-19 related tweets. Appl. Geogr. 133 , 102473 (2021).

Hong, S. & Kim, S. H. Political polarization on Twitter: implications for the use of social media in digital governments. Gov. Inf. Q. 33 , 777–782 (2016).

Westerwick, A., Sude, D., Robinson, M. & Knobloch-Westerwick, S. Peers versus pros: confirmation bias in selective exposure to user-generated versus professional media messages and its consequences. Mass Commun. Soc. 23 , 510–536 (2020).

Stoica, A.-A., Riederer, C. & Chaintreau, A. Algorithmic glass ceiling in social networks: the effects of social recommendations on network diversity. In Proc. 2018 World Wide Web Conference 923–932 (ACM Press, 2018).

Vissenberg, J., Coninck, D. D. & dHaenens, L. Relating adolescents’ exposure to legacy and digital news media and intergroup contact to their attitudes towards immigrants. Communications 46 , 373–393 (2021).

Yu, R. P. & Oh, Y. W. Social media and expressive citizenship: understanding the relationships between social and entertainment expression on Facebook and political participation. Telemat. Inform. 35 , 2299–2311 (2018).

Lake, J. S., Alston, A. T. & Kahn, K. B. How social networking use and beliefs about inequality affect engagement with racial justice movements. Race Justice 11 , 500–519 (2021).

Barnidge, M., Huber, B., Gil de Zúñiga, H. & Liu, J. H. Social media as a sphere for risky political expression: a twenty-country multilevel comparative analysis. Int. J. Press Polit. 23 , 161–182 (2018).

Mustapha, L. K. & Omar, B. Do social media matter? Examining social media use and youths political participation during the 2019 Nigerian general elections. Round Table 109 , 441–457 (2020).

Qin, A. Y. Judging them by my media use: exploring the cause and consequences of perceived selective exposure. Mass Commun. Soc. 25 , 237–259 (2022).

Blank, G. & Lutz, C. Benefits and harms from Internet use: a differentiated analysis of Great Britain. New Media Soc. 20 , 618–640 (2018).

Bail, C. A., Merhout, F. & Ding, P. Using Internet search data to examine the relationship between anti-Muslim and pro-ISIS sentiment in U.S. counties. Sci. Adv. 4 , eaao5948 (2018).

Hendriks Vettehen, P., Troost, J., Boerboom, L., Steijaert, M. & Scheepers, P. The relationship between media content preferences and political participation in 25 European countries: the moderating role of broadband penetration and broadband access. Commun. Res. 47 , 967–987 (2020).

Arlt, D. & Wolling, J. Bias wanted! Examining peoples information exposure, quality expectations and bias perceptions in the context of the refugees debate among different segments of the German population. Communications 43 , 75–99 (2018).

Cardenal, A. S., Aguilar-Paredes, C., Cristancho, C. & Majó-Vázquez, S. Echo-chambers in online news consumption: evidence from survey and navigation data in Spain. Eur. J. Commun. 34 , 360–376 (2019).

Gallego, J., Martínez, J. D., Munger, K. & Vásquez-Cortés, M. Tweeting for peace: experimental evidence from the 2016 Colombian plebiscite. Elect. Stud. 62 , 102072 (2019).

Allington, D., McAndrew, S., Moxham-Hall, V. L. & Duffy, B. Media usage predicts intention to be vaccinated against SARS-CoV-2 in the US and the UK. Vaccine 39 , 2595–2603 (2021).

Wagner, K. M., Gainous, J. & Abbott, J. P. Gender differences in critical digital political engagement in China: the consequences for protest attitudes. 39 , 211–225 (2021).

Foos, F., Kostadinov, L., Marinov, N. & Schimmelfennig, F. Does social media promote civic activism? A field experiment with a civic campaign. Polit. Sci. Res. Methods 9 , 500–518 (2021).

Erdem, R. & Ozejder, I. Use of social media for political purposes: the case of Diyarbakir. Rev. Cercet. Interv. Soc. 72 , 187–209 (2021).

Bimber, B. Information and political engagement in America: the search for effects of information technology at the individual level. Polit. Res. Q. 54 , 53 (2001).

Sindermann, C., Elhai, J. D., Moshagen, M. & Montag, C. Age, gender, personality, ideological attitudes and individual differences in a person’s news spectrum: how many and who might be prone to filter bubbles and echo chambers online? Heliyon 6 , e03214 (2020).

Guntuku, S. C., Buttenheim, A. M., Sherman, G. & Merchant, R. M. Twitter discourse reveals geographical and temporal variation in concerns about COVID-19 vaccines in the United States. Vaccine 39 , 4034–4038 (2021).

Kim, H. & Joshanloo, M. Internet access and voicing opinions: the moderating roles of age and the national economy. Soc. Indic. Res. 150 , 121–141 (2020).

Heinsohn, T., Fatke, M., Israel, J., Marschall, S. & Schultze, M. Effects of voting advice applications during election campaigns. Evidence from a panel study at the 2014 European elections. J. Inf. Technol. Polit. 16 , 250–264 (2019).

Shmargad, Y. & Klar, S. Sorting the news: how ranking by popularity polarizes our politics. Polit. Commun. 37 , 423–446 (2020).

Chen, H.-T. Spiral of silence on social media and the moderating role of disagreement and publicness in the network: analyzing expressive and withdrawal behaviors. New Media Soc. 20 , 3917–3936 (2018).

Romer, D. & Jamieson, K. H. Patterns of media use, strength of belief in COVID-19 conspiracy theories, and the prevention of COVID-19 from March to July 2020 in the United States: survey study. J. Med. Internet Res. 23 , e25215 (2021).

Klein, E. & Robison, J. Like, post, and distrust? How social media use affects trust in government. Polit. Commun. 37 , 46–64 (2020).

Kim, C. & Lee, S. Does social media type matter to politics? Investigating the difference in political participation depending on preferred social media sites. Soc. Sci. Q. 102 , 2942–2954 (2021).

Garrett, R. K. & Bond, R. M. Conservatives susceptibility to political misperceptions. Sci. Adv. 7 , eabf1234 (2021).

van Tubergen, F., Cinjee, T., Menshikova, A. & Veldkamp, J. Online activity of mosques and Muslims in the Netherlands: a study of Facebook, Instagram, YouTube and Twitter. PLoS ONE 16 , e0254881 (2021).

Gherghina, S. & Rusu, E. Begin again: election campaign and own opinions among first-time voters in Romania. Soc. Sci. Q. 102 , 1311–1329 (2021).

Karakaya, S. & Glazier, R. A. Media, information, and political participation: the importance of online news sources in the absence of a free press. J. Inf. Technol. Polit. 16 , 290–306 (2019).

Heatherly, K. A., Lu, Y. & Lee, J. K. Filtering out the other side? Cross-cutting and like-minded discussions on social networking sites. New Media Soc. 19 , 1271–1289 (2017).

Kaakinen, M., Oksanen, A. & Räsänen, P. Did the risk of exposure to online hate increase after the November 2015 Paris attacks? A group relations approach. Comput. Hum. Behav. 78 , 90–97 (2018).

Kurfi, M. Y., Msughter, M. E. & Mohamed, I. Digital images on social media and proliferation of fake news on Covid-19 in Kano, Nigeria. Galactica Media J. Media Stud. 3 , 103–124 (2021).

Zhang, X. & Lin, W.-Y. Stoking the fires of participation: extending the gamson hypothesis on social media use and elite-challenging political engagement. Comput. Hum. Behav. 79 , 217–226 (2018).

Costello, M., Barrett-Fox, R., Bernatzky, C., Hawdon, J. & Mendes, K. Predictors of viewing online extremism among America’s youth. Youth. Soc. 52 , 710–727 (2020).

Dvir-Gvirsman, S., Tsfati, Y. & Menchen-Trevino, E. The extent and nature of ideological selective exposure online: combining survey responses with actual web log data from the 2013 Israeli elections. New Media Soc. 18 , 857–877 (2016).

Chang, K. & Park, J. Social media use and participation in dueling protests: the case of the 2016-2017 presidential corruption scandal in South Korea. Int. J. Press Polit. 26 , 547–567 (2021).

Lee, H. & Hahn, K. S. Partisan selective following on Twitter over time: polarization or depolarization?. Asian J. Commun. 28 , 227–246 (2018).

Boxell, L., Gentzkow, M. & Shapiro, J. M. Greater Internet use is not associated with faster growth in political polarization among US demographic groups. Proc. Natl Acad. Sci. USA 114 , 10612–10617 (2017).

Stier, S., Kirkizh, N., Froio, C. & Schroeder, R. Populist attitudes and selective exposure to online news: a cross-country analysis combining web tracking and surveys. Int. J. Press Polit. 25 , 426–446 (2020).

Garrett, R. K. et al. Implications of pro- and counterattitudinal information exposure for affective polarization: partisan media exposure and affective polarization. Hum. Commun. Res. 40 , 309–332 (2014).

Sharma, I., Jain, K. & Singh, G. Effect of online political incivility on partisan attitude: role of issue involvement, moral identity and incivility accountability. Online Inf. Rev. 44 , 1421–1441 (2020).

Stella, M., Ferrara, E. & De Domenico, M. Bots increase exposure to negative and inflammatory content in online social systems. Proc. Natl Acad. Sci. USA 115 , 12435–12440 (2018).

Ackland, R., ONeil, M. & Park, S. Engagement with news on Twitter: insights from Australia and Korea. Asian J. Commun. 29 , 235–251 (2019).

Mothes, C. & Ohme, J. Partisan selective exposure in times of political and technological upheaval: a social media field experiment. Media Commun. 7 , 42–53 (2019).

Popa, S. A., Theocharis, Y. & Schnaudt, C. From seeing the writing on the wall, to getting together for a bowl: direct and compensating effects of Facebook use on offline associational membership. J. Inf. Technol. Polit. 13 , 222–238 (2016).

Karlsen, R., Beyer, A. & Steen-Johnsen, K. Do high-choice media environments facilitate news avoidance? A longitudinal study 1997–2016. J. Broadcast. Electron. Media 64 , 794–814 (2020).

Goyanes, M. Antecedents of incidental news exposure: the role of media preference, use and trust. Journali. Pract. 14 , 714–729 (2020).

Siegel, A. A., Nagler, J., Bonneau, R. & Tucker, J. A. Tweeting beyond tahrir: ideological diversity and political intolerance in Egyptian Twitter networks. World Polit. 73 , 243–274 (2021).

Mashuri, A. et al. The socio-psychological predictors of support for post-truth collective action. J. Soc. Psychol. 162 , 504–522 (2022).

Alshareef, M. & Alotiby, A. Prevalence and perception among Saudi Arabian population about resharing of information on social media regarding natural remedies as protective measures against COVID-19. Int. J. Gen. Med. 14 , 5127–5137 (2021).

Florio, K., Basile, V., Lai, M. & Patti, V. Leveraging hate speech detection to investigate immigration-related phenomena in Italy. In 2019 8th International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW) 1–7 (IEEE, 2019). https://ieeexplore.ieee.org/document/8925079/

Dilliplane, S. All the news you want to hear: the impact of partisan news exposure on political participation. Public Opin. Q. 75 , 287–316 (2011).

Wasisto, A. Electoral volatility of the 2019 presidential election: a study in Jakarta and Depok, Indonesia. Masy. Kebud. Pol. 34 , 281–292 (2021).

Sturm Wilkerson, H., Riedl, M. J. & Whipple, K. N. Affective affordances: exploring Facebook reactions as emotional responses to hyperpartisan political news. Digit. Journal. 9 , 1040–1061 (2021).

Zunino, E. Medios digitales y COVID-19: sobreinformación, polarización y desinformación. Universitas 34 , 127–146 (2021). https://universitas.ups.edu.ec/index.php/universitas/article/view/34.2021.06

Strauß, N., Huber, B. & Gil de Zúñiga, H. Structural influences on the news finds me perception: why people believe they dont have to actively seek news anymore. Soc. Media Soc . https://doi.org/10.1177/20563051211024966 (2021).

Tai, K.-T., Porumbescu, G. & Shon, J. Can e-participation stimulate offline citizen participation: an empirical test with practical implications. Public Manage. Rev. 22 , 278–296 (2020).

Kim, H. The mere exposure effect of tweets on vote choice. J. Inf. Technol. Polit. 18 , 455–465 (2021).

Marozzo, F. & Bessi, A. Analyzing polarization of social media users and news sites during political campaigns. Soc. Netw. Anal. Min. 8 , 1 (2018).

Granberg-Rademacker, J. S. & Parsneau, K. Lets get ready to tweet! An analysis of Twitter use by 2018 senate candidates. Congr. Pres. 48 , 78–100 (2021).

Chan, M. Media use and the social identity model of collective action: examining the roles of online alternative news and social media news. Journal. Mass Commun. Q. 94 , 663–681 (2017).

Kitchens, B., Johnson, S. L. & Gray, P. Understanding echo chambers and filter bubbles: the impact of social media on diversification and partisan shifts in news consumption. MIS Q. 44 , 1619–1649 (2020).

Amit, S., Mannan, S. & Islam, A. Bangladesh: time spent online, conflict and radicalization. Confl. Stud. Q. 2020 , 3–21 (2020). http://www.csq.ro/wp-content/uploads/Sajid-AMIT-et-al.pdf

Casero-Ripollés, A. Influencia de los medios de comunicación en la conversación política en Twitter. Rev. ICONO14 18 , 33–57 (2020).

Saud, M., El Hariri, D. B. & Ashfaq, A. The role of social media in promoting political participation: the Lebanon experience. Masy. Kebud. Pol. 33 , 248–255 (2020).

Sismeiro, C. & Mahmood, A. Competitive vs. complementary effects in online social networks and news consumption: a natural experiment. Manage. Sci. 64 , 5014–5037 (2018).

Johannesson, M. P. & Knudsen, E. Disentangling the influence of recommender attributes and news-story attributes: a conjoint experiment on exposure and sharing decisions on social networking sites. Digit. Journal. 9 , 1141–1161 (2021).

Yamamoto, M. & Nah, S. Mobile information seeking and political participation: a differential gains approach with offline and online discussion attributes. New Media Soc. 20 , 2070–2090 (2018).

Lee, J. M., Park, Y. & Kim, G. D. Social media and regionalism in South Korean voting behavior: The case of the 19th South Korean presidential election. Issues Stud. 54 , 1840006 (2018).

Allcott, H. & Gentzkow, M. Social media and fake news in the 2016 election. J. Econ. Perspect. 31 , 211–236 (2017).

Theocharis, Y., Barberá, P., Fazekas, Z. & Popa, S. A. The dynamics of political incivility on Twitter. SAGE Open https://doi.org/10.1177/2158244020919447 (2020).

Nah, S. & Yamamoto, M. The integrated media effect: rethinking the effect of media use on civic participation in the networked digital media environment. Am. Behav. Sci. 62 , 1061–1078 (2018).

Pang, H. Can microblogs motivate involvement in civic and political life? Examining uses, gratifications and social outcomes among Chinese youth. Online Inf. Rev. 42 , 663–680 (2018).

Moeller, J., Kühne, R. & De Vreese, C. Mobilizing youth in the 21st century: how digital media use fosters civic duty, information efficacy, and political participation. J. Broadcast. Electron. Media 62 , 445–460 (2018).

Cinelli, M. et al. Selective exposure shapes the Facebook news diet. PLoS ONE 15 , e0229129 (2020).

Yamamoto, M. & Morey, A. C. Incidental news exposure on social media: a campaign communication mediation approach. Soc. Media Soc . https://doi.org/10.1177/2056305119843619 (2019).

Levy, R. Social media, news consumption, and polarization: evidence from a field experiment. Am. Econ. Rev 111 , 831–870 (2021).

Vozab, D. Generational patterns of digital news consumption: from traditionalists to millennial minimalists. Medijske Stud. 10 , 107–126 (2020).

Jamal, A., Kizgin, H., Rana, N. P., Laroche, M. & Dwivedi, Y. K. Impact of acculturation, online participation and involvement on voting intentions. Gov. Inf. Q. 36 , 510–519 (2019).

Shen, F., Xia, C. & Skoric, M. Examining the roles of social media and alternative media in social movement participation: a study of Hong Kong’s umbrella movement. Telemat. Inform. 47 , 101303 (2020).

Dohle, M., Bernhard, U. & Kelm, O. Presumed media influences and demands for restrictions: using panel data to examine the causal direction. Mass Commun. Soc. 20 , 595–613 (2017).

Marquart, F., Goldberg, A. C. & de Vreese, C. H. This time I’m (not) voting: a comprehensive overview of campaign factors influencing turnout at European Parliament elections. Eur. Union Polit. 21 , 680–705 (2020).

Hoffmann, C. P. & Lutz, C. Digital divides in political participation: the mediating role of social media self-efficacy and privacy concerns. Policy Internet 13 , 6–29 (2021).

Mitts, T. From isolation to radicalization: anti-muslim hostility and support for ISIS in the West. Am. Polit. Sci. Rev 113 , 173–194 (2019).

Lukito, J. Coordinating a multi-platform disinformation campaign: internet research agency activity on three U.S. social media platforms, 2015 to 2017. Polit. Commun. 37 , 238–255 (2020).

Hong, S., Choi, H. & Kim, T. K. Why do politicians tweet? extremists, underdogs, and opposing parties as political tweeters. Policy Internet 11 , 305–323 (2019).

Chadwick, A. et al. Online social endorsement and Covid-19 vaccine hesitancy in the United Kingdom. Soc. Media Soc. 7 , 205630512110088 (2021).

Zumárraga-Espinosa, M. Redes sociales y protesta política: Un análisis del rol moderador del estatus socioeconmico y la pertenencia a grupos políticos. Doxa Comun. 30 , 55–77 (2020). https://revistascientificas.uspceu.com/doxacomunicacion/article/view/500

Fletcher, R. & Nielsen, R. K. Automated serendipity: the effect of using search engines on news repertoire balance and diversity. Digit. Journal. 6 , 976–989 (2018).

Sell, T. K., Hosangadi, D. & Trotochaud, M. Misinformation and the US Ebola communication crisis: analyzing the veracity and content of social media messages related to a fear-inducing infectious disease outbreak. BMC Public Health 20 , 550 (2020).

Brugnoli, E., Cinelli, M., Quattrociocchi, W. & Scala, A. Recursive patterns in online echo chambers. Sci. Rep. 9 , 20118 (2019).

Barnidge, M., Kim, B., Sherrill, L. A., Luknar, Z. & Zhang, J. Perceived exposure to and avoidance of hate speech in various communication settings. Telemat. Inform. 44 , 101263 (2019).

Kim, B. & Hoewe, J. Developing contemporary factors of political participation. Soc. Sci. J. https://doi.org/10.1080/03623319.2020.1782641 (2020). https://www.tandfonline.com/doi/full/10.1080/03623319.2020.1782641

Barnidge, M., Sayre, B. & Rojas, H. Perceptions of the media and the public and their effects on political participation in Colombia. Mass Commun. Soc. 18 , 259–280 (2015).

Petrova, M., Sen, A. & Yildirim, P. Social media and political contributions: the impact of new technology on political competition. Manage. Sci. 67 , 2997–3021 (2021).

Bryson, B. P. Polarizing the middle: internet exposure and public opinion. Int. J. Sociol. Soc. Policy 40 , 99–113 (2020).

Bovet, A. & Makse, H. A. Influence of fake news in Twitter during the 2016 US presidential election. Nat. Commun. 10 , 7 (2019).

Germani, F. & Biller-Andorno, N. The anti-vaccination infodemic on social media: a behavioral analysis. PLoS ONE 16 , e0247642 (2021).

Wei, R. & Lo, V.-h. News media use and knowledge about the 2006 U.S. midterm elections: why exposure matters in voter learning. Int. J. Public Opin. Res. 20 , 347–362 (2008).

Zhu, A. Y. F., Chan, A. L. S. & Chou, K. L. The pathway toward radical political participation among young people in Hong Kong: a communication mediation approach. East Asia 37 , 45–62 (2020).

Balcells, J. & Padró-Solanet, A. Crossing lines in the Twitter debate on Catalonia’s independence. Int. J. Press Polit. 25 , 28–52 (2020).

Grover, P., Kar, A. K., Dwivedi, Y. K. & Janssen, M. Polarization and acculturation in US Election 2016 outcomes - can Twitter analytics predict changes in voting preferences. Technol. Forecast. Soc. Change 145 , 438–460 (2019).

Choi, D.-H. & Shin, D.-H. A dialectic perspective on the interactive relationship between social media and civic participation: the moderating role of social capital. Inf. Commun. Soc. 20 , 151–166 (2017).

Thorson, K., Cotter, K., Medeiros, M. & Pak, C. Algorithmic inference, political interest, and exposure to news and politics on Facebook. Inf. Commun. Soc. 24 , 183–200 (2021).

Kim, Y., Hsu, S.-H. & Gil de Zúñiga, H. Influence of social media use on discussion network heterogeneity and civic engagement: the moderating role of personality traits. J. Commun. 63 , 498–516 (2013).

Guess, A. M., Nyhan, B. & Reifler, J. Exposure to untrustworthy websites in the 2016 US election. Nat. Hum. Behav. 4 , 472–480 (2020).

Hokka, J. & Nelimarkka, M. Affective economy of national-populist images: investigating national and transnational online networks through visual big data. New Media Soc. 22 , 770–792 (2020).

Schumann, S., Boer, D., Hanke, K. & Liu, J. Social media use and support for populist radical right parties: assessing exposure and selection effects in a two-wave panel study. Inf. Commun. Soc. 24 , 921–940 (2021).

Kim, M. How does Facebook news use lead to actions in South Korea? The role of Facebook discussion network heterogeneity, political interest, and conflict avoidance in predicting political participation. Telemat. Inform. 35 , 1373–1381 (2018).

Mosca, L. & Quaranta, M. Are digital platforms potential drivers of the populist vote? A comparative analysis of France, Germany and Italy. Inf. Commun. Soc. 24 , 1441–1459 (2021).

Justwan, F., Baumgaertner, B., Carlisle, J. E., Clark, A. K. & Clark, M. Social media echo chambers and satisfaction with democracy among Democrats and Republicans in the aftermath of the 2016 US elections. J. Elect. Public Opin. Parties 28 , 424–442 (2018).

Smith, S. T., Kao, E. K., Shah, D. C., Simek, O. & Rubin, D. B. Influence estimation on social media networks using causal inference. In 2018 IEEE Statistical Signal Processing Workshop (SSP) 328–332 (IEEE, 2018). https://ieeexplore.ieee.org/document/8450823/

Bode, L. et al. Participation in contentious politics: rethinking the roles of news, social media, and conversation amid divisiveness. J. Inf. Technol. Polit. 15 , 215–229 (2018).

Cinelli, M., Cresci, S., Galeazzi, A., Quattrociocchi, W. & Tesconi, M. The limited reach of fake news on Twitter during 2019 European elections. PLoS ONE 15 , e0234689 (2020).

Guenther, L., Ruhrmann, G., Bischoff, J., Penzel, T. & Weber, A. Strategic framing and social media engagement: analyzing memes posted by the German identitarian movement on Facebook. Soc. Media Soc . https://doi.org/10.1177/2056305119898777 (2020).

Song, H., Cho, J. & Benefield, G. A. The dynamics of message selection in online political discussion forums: self-segregation or diverse exposure?. Commun. Res. 47 , 125–152 (2020).

Rice, L. L. & Moffett, K. W. Snapchat and civic engagement among college students. J. Inf. Technol. Polit. 16 , 87–104 (2019).

Beam, M. A., Hmielowski, J. D. & Hutchens, M. J. Democratic digital inequalities: threat and opportunity in online citizenship from motivation and ability. Am. Behav. Sci. 62 , 1079–1096 (2018).

Hermann, E., Eisend, M. & Bayón, T. Facebook and the cultivation of ethnic diversity perceptions and attitudes. Internet Res. 30 , 1123–1141 (2020).

Powell, A., Scott, A. J. & Henry, N. Digital harassment and abuse: experiences of sexuality and gender minority adults. Eur. J. Criminol. 17 , 199–223 (2020).

Li, L., Chen, J. & Raghunathan, S. Informative role of recommender systems in electronic marketplaces: a boon or a bane for competing sellers. MIS Q. 44 , 1957–1985 (2020).

Choi, D.-H., Yoo, W., Noh, G.-Y. & Park, K. The impact of social media on risk perceptions during the MERS outbreak in South Korea. Comput. Hum. Behav. 72 , 422–431 (2017).

Samuel-Azran, T. & Hayat, T. Online news recommendations credibility: the tie is mightier than the source. Comunicar 27 , 71–80 (2019).

Kushin, M. J., Yamamoto, M. & Dalisay, F. Societal majority, Facebook, and the spiral of silence in the 2016 US presidential election. Soc. Media Soc . https://doi.org/10.1177/2056305119855139 (2019).

Ardi, R. Partisan selective exposure to fake news content. Makara Hum. Behav. Stud. Asia. 23 , 3 (2019).

Feld, S. L. & McGail, A. Egonets as systematically biased windows on society. Netw. Sci. 8 , 399–417 (2020).

Hedayatifar, L., Rigg, R. A., Bar-Yam, Y. & Morales, A. J. US social fragmentation at multiple scales. J. R. Soc. Interface 16 , 20190509 (2019).

Bale, T., Webb, P. & Poletti, M. Participating locally and nationally: explaining the offline and online activism of British Party Members. Polit. Stud. 67 , 658–675 (2019).

Celik, S. Experiences of internet users regarding cyberhate. Inf. Technol. People 32 , 1446–1471 (2019).

Sainudiin, R., Yogeeswaran, K., Nash, K. & Sahioun, R. Characterizing the Twitter network of prominent politicians and SPLC-defined hate groups in the 2016 US presidential election. Soc. Netw. Anal. Min. 9 , 34 (2019).

Lee, S. & Xenos, M. Incidental news exposure via social media and political participation: evidence of reciprocal effects. New Media Soc. 24 , 178–201 (2022).

Casero-Ripollés, A., Micó-Sanz, J.-L. & Díez-Bosch, M. Digital public sphere and geography: the influence of physical location on Twitter’s political conversation. Media Commun. 8 , 96–106 (2020).

Kim, Y., Chen, H.-T. & Wang, Y. Living in the smartphone age: examining the conditional indirect effects of mobile phone use on political participation. J. Broadcast. Electron. Media. 60 , 694–713 (2016).

Elvestad, E., Phillips, A. & Feuerstein, M. Can trust in traditional news media explain cross-national differences in news exposure of young people online? A comparative study of Israel, Norway and the United Kingdom. Digit. Journal. 6 , 216–235 (2018).

Steffan, D. & Venema, N. New medium, old strategies? Comparing online and traditional campaign posters for German Bundestag elections, 2013–2017. Eur. J. Commun. 35 , 370–388 (2020).

Vraga, E. K. & Tully, M. News literacy, social media behaviors, and skepticism toward information on social media. Inf. Commun. Soc. 24 , 150–166 (2021).

Kofi Frimpong, A. N., Li, P., Nyame, G. & Hossin, M. A. The impact of social media political activists on voting patterns. Polit. Behav. 44 , 599–652 (2022).

Ribeiro, M. H., Ottoni, R., West, R., Almeida, V. A. F. & Meira, W. Auditing radicalization pathways on YouTube. In Proc. 2020 Conference on Fairness, Accountability, and Transparency 131–141 (ACM, 2020). https://dl.acm.org/doi/10.1145/3351095.3372879

López-Rabadán, P. & Doménech-Fabregat, H. Nuevas funciones de Instagram en el avance de la política espectáculo. Claves profesionales y estrategia visual de Vox en su despegue electoral. Prof. Inf. https://doi.org/10.3145/epi.2021.mar.20 (2021). https://revista.profesionaldelainformacion.com/index.php/EPI/article/view/85530

Gainous, J., Abbott, J. P. & Wagner, K. M. Active vs. passive social media engagement with critical information: protest behavior in two Asian countries. Int. J. Press Polit. 26 , 464–483 (2021).

Skoric, M. M., Zhu, Q. & Lin, J.-H. T. What predicts selective avoidance on social media? A study of political unfriending in Hong Kong and Taiwan. Am. Behav. Sci. 62 , 1097–1115 (2018).

Zannettou, S. et al. Disinformation warfare: understanding state-sponsored trolls on Twitter and their influence on the Web. In Companion Proc. 2019 World Wide Web Conference 218–226 (ACM, 2019). https://dl.acm.org/doi/10.1145/3308560.3316495

Alsaad, A., Taamneh, A. & Al-Jedaiah, M. N. Does social media increase racist behavior? An examination of confirmation bias theory. Technol. Soc. 55 , 41–46 (2018).

Nanz, A., Heiss, R. & Matthes, J. Antecedents of intentional and incidental exposure modes on social media and consequences for political participation: a panel study. Acta Politica 57 , 235–253 (2022).

Ardèvol-Abreu, A., Hooker, C. M. & Gil de Zúñiga, H. Online news creation, trust in the media, and political participation: direct and moderating effects over time. Journalism 19 , 611–631 (2018).

Lu, Y. & Pan, J. Capturing clicks: how the Chinese government uses clickbait to compete for visibility. Polit. Commun. 38 , 23–54 (2021).

Davidson, B. I., Jones, S. L., Joinson, A. N. & Hinds, J. The evolution of online ideological communities. PLoS ONE 14 , e0216932 (2019).

Mahmood, Q. K., Bhutta, M. H. & Haq, M. A. U. Effects of sociodemographic variables and Facebook group membership on students’ political participation. Educ. Inf. Technol. 23 , 2235–2247 (2018).

Min, H. & Yun, S. Selective exposure and political polarization of public opinion on the presidential impeachment in South Korea: Facebook vs. KakaoTalk. Korea Obs. 49 , 137–159 (2018).

Yamamoto, M., Kushin, M. J. & Dalisay, F. How informed are messaging app users about politics? A linkage of messaging app use and political knowledge and participation. Telemat. Inform. 35 , 2376–2386 (2018).

Martínez-Torres, H. & Gámez, C. Is internet access bad news for media-capturing incumbents?. J. Appl. Econ. 22 , 527–553 (2019).

Imran, M. S., Fatima, M. & Kosar, G. Connectivism: e-learning of democratic values on social media public spheres. In 2017 International Conference on Information and Communication Technologies (ICICT) 82–89 (IEEE, 2017).

Scharkow, M., Mangold, F., Stier, S. & Breuer, J. How social network sites and other online intermediaries increase exposure to news. Proc. Natl Acad. Sci. USA 117 , 2761–2763 (2020).

Kwak, N. et al. Perceptions of social media for politics: testing the slacktivism hypothesis. Hum. Commun. Res. 44 , 197–221 (2018).

Schmidt, A. L., Zollo, F., Scala, A., Betsch, C. & Quattrociocchi, W. Polarization of the vaccination debate on Facebook. Vaccine 36 , 3606–3612 (2018).

Vaccari, C. & Valeriani, A. Dual screening, public service broadcasting, and political participation in eight Western democracies. Int. J. Press Polit. 23 , 367–388 (2018).

Sharar, B. & Abd-El-Barr, M. Citizens’ perspective on the impact of social media on politics in Kuwait. In 2018 International Conference on Computing Sciences and Engineering (ICCSE) 1–6 (IEEE, 2018). https://ieeexplore.ieee.org/document/8374207/

Lee, S. H. & Fu, K.-w. Internet use and protest politics in South Korea and Taiwan. J. East Asian Stud. 19 , 89–109 (2019).

Piatak, J. & Mikkelsen, I. Does social media engagement translate to civic engagement offline? Nonprofit Volunt. Sect. Q. 50 , 1079–1101 (2021).

Baek, Y. M. Political mobilization through social network sites: the mobilizing power of political messages received from SNS friends. Comput. Hum. Behav. 44 , 12–19 (2015).

Ferrucci, P., Hopp, T. & Vargo, C. J. Civic engagement, social capital, and ideological extremity: exploring online political engagement and political expression on Facebook. New Media Soc. 22 , 1095–1115 (2020).

Corrigall-Brown, C. & Wilkes, R. Media exposure and the engaged citizen: how the media shape political participation. Soc. Sci. J. 51 , 408–421 (2014).

Valenzuela, S., Correa, T. & Gil de Zúñiga, H. Ties, likes, and tweets: using strong and weak ties to explain differences in protest participation across Facebook and Twitter use. Polit. Commun. 35 , 117–134 (2018).

Popan, J. R., Coursey, L., Acosta, J. & Kenworthy, J. Testing the effects of incivility during internet political discussion on perceptions of rational argument and evaluations of a political outgroup. Comput. Hum. Behav. 96 , 123–132 (2019).

Štětka, V., Mazák, J. & Vochocová, L. Nobody tells us what to write about: the disinformation media ecosystem and its consumers in the Czech Republic. Javnost 28 , 90–109 (2021).

Kobayashi, T. Depolarization through social media use: evidence from dual identifiers in Hong Kong. New Media Soc. 22 , 1339–1358 (2020).

Talwar, S., Dhir, A., Kaur, P., Zafar, N. & Alrasheedy, M. Why do people share fake news? Associations between the dark side of social media use and fake news sharing behavior. J. Retail. Consum. Serv. 51 , 72–82 (2019).

Yarchi, M., Baden, C. & Kligler-Vilenchik, N. Political polarization on the digital sphere: a cross-platform, over-time analysis of interactional, positional, and affective polarization on social media. Polit. Commun. 38 , 98–139 (2021).

Bobba, G. Social media populism: features and likeability of Lega Nord communication on Facebook. Eur. Polit. Sci. 18 , 11–23 (2019).

Tewksbury, D. & Riles, J. M. Polarization as a function of citizen predispositions and exposure to news on the internet. J. Broadcast. Electron. Media. 59 , 381–398 (2015).

Giglietto, F., Righetti, N., Rossi, L. & Marino, G. It takes a village to manipulate the media: coordinated link sharing behavior during 2018 and 2019 Italian elections. Inf. Commun. Soc. 23 , 867–891 (2020).

Marcinkowski, F. & Dosenovic, P. From incidental exposure to intentional avoidance: psychological reactance to political communication during the 2017 German national election campaign. New Media Soc. 23 , 457–478 (2021).

Bastien, F., Koop, R., Small, T. A., Giasson, T. & Jansen, H. The role of online technologies and digital skills in the political participation of citizens with disabilities. J. Inf. Technol. Polit. 17 , 218–231 (2020).

Nelson, J. L. & Taneja, H. The small, disloyal fake news audience: the role of audience availability in fake news consumption. New Media Soc. 20 , 3720–3737 (2018).

Lelkes, Y., Sood, G. & Iyengar, S. The hostile audience: the effect of access to broadband internet on partisan affect. Am. J. Polit. Sci. 61 , 5–20 (2017).

Hunter, L. Y., Griffith, C. E. & Warren, T. Internet connectivity and domestic terrorism in democracies. Int. J. Sociol. 50 , 201–219 (2020).

Heiss, R. & Matthes, J. Does incidental exposure on social media equalize or reinforce participatory gaps? Evidence from a panel study. New Media Soc. 21 , 2463–2482 (2019).

Bobba, G., Cremonesi, C., Mancosu, M. & Seddone, A. Populism and the gender gap: comparing digital engagement with populist and non-populist Facebook pages in France, Italy, and Spain. Int. J. Press Polit. 23 , 458–475 (2018).

Troian, J., Arciszewski, T. & Apostolidis, T. The dynamics of public opinion following terror attacks: evidence for a decrease in equalitarian values from Internet Search Volume Indices. Cyberpsychology 13 , 4 (2019). https://cyberpsychology.eu/article/view/12015

Frissen, T. Internet, the great radicalizer? Exploring relationships between seeking for online extremist materials and cognitive radicalization in young adults. Comput. Hum. Behav. 114 , 106549 (2021).

Park, B., Kang, M. Y. & Lee, J. Sustainable political social media marketing: effects of structural features in plain text messages. Sustainability 12 , 5997 (2020).

Yun, G. W., Park, S.-Y., Holody, K., Yoon, K. S. & Xie, S. Selective moderation, selective responding, and balkanization of the blogosphere: a field experiment. Media Psychol. 16 , 295–317 (2013).

Allen, J., Howland, B., Mobius, M., Rothschild, D. & Watts, D. J. Evaluating the fake news problem at the scale of the information ecosystem. Sci. Adv. 6 , eaay3539 (2020).

Gainous, J., Abbott, J. P. & Wagner, K. M. Traditional versus internet media in a restricted information environment: how trust in the medium matters. Polit. Behav. 41 , 401–422 (2019).

Boulianne, S., Koc-Michalska, K. & Bimber, B. Mobilizing media: comparing TV and social media effects on protest mobilization. Inf. Commun. Soc. 23 , 642–664 (2020).

Waechter, N. The participative role of social media for the disadvantaged young generation in the Arab Spring. Österreich. Z. Soziol. 44 , 217–236 (2019).

Wang, T. & Shen, F. Perceived party polarization, news attentiveness, and political participation: a mediated moderation model. Asian J. Commun. 28 , 620–637 (2018).

Guess, A., Nagler, J. & Tucker, J. Less than you think: prevalence and predictors of fake news dissemination on Facebook. Sci. Adv. 5 , eaau4586 (2019).

Kaur, M. & Verma, R. Demographics, social media usage, and political engagement in Punjab. Indian J. Market. 48 , 43 (2018).

Castillo-Díaz, A. & Castillo-Esparcia, A. Relación entre la participación en foros y blogs de debate político en internet y el seguimiento de información política en medios profesionales: análisis evolutivo 2011-2016. Prof. Inf. 27 , 1248 (2018).

Heiss, R., Knoll, J. & Matthes, J. Pathways to political (dis-)engagement: motivations behind social media use and the role of incidental and intentional exposure modes in adolescents political engagement. Communications 45 , 671–693 (2020).

Enjolras, B., Steen-Johnsen, K. & Wollebæk, D. Social media and mobilization to offline demonstrations: transcending participatory divides? New Media Soc. 15 , 890–908 (2013).

Bhat, S. I., Arif, T., Malik, M. B. & Sheikh, A. A. Browser simulation-based crawler for online social network profile extraction. Int. J. Web Based Communities 16 , 321 (2020).

Aruguete, N., Calvo, E. & Ventura, T. News sharing, gatekeeping, and polarization: a study of the #Bolsonaro election. Digit. Journal. 9 , 1–23 (2021).

Bae, S. Y. The social mediation of political rumors: examining the dynamics in social media and belief in political rumors. Journalism 21 , 1522–1538 (2020).

Nikolov, D., Lalmas, M., Flammini, A. & Menczer, F. Quantifying biases in online information exposure. J. Assoc. Inf. Sci. Technol. 70 , 218–229 (2019).

Akpan, I. J. et al. Association between what people learned about COVID-19. J. Med. Internet Res. 23 , e28975 (2021).

Yamamoto, M., Nah, S. & Bae, S. Y. Social media prosumption and online political participation: an examination of online communication processes. New Media Soc. 22 , 1885–1902 (2020).

Song, T., Tang, Q. & Huang, J. Triadic closure, homophily, and reciprocation: an empirical investigation of social ties between content providers. Inf. Syst. Res. 30 , 912–926 (2019).

Strauß, N., Huber, B. & Gil de Zúñiga, H. “Yes, I saw it – but didn’t read it…” a cross-country study, exploring relationships between incidental news exposure and news use across platforms. Digit. Journal. 8 , 1181–1205 (2020).

Ahmad, S. Political behavior in virtual environment: role of social media intensity, internet connectivity, and political affiliation in online political persuasion among university students. J. Hum. Behav. Soc. Environ. 30 , 457–473 (2020).

Lee, H. Voters involvement, attitude, and confidence in the era of new media. Palgrave Commun. 6 , 1 (2020).

Čábelková, I., Smutka, L. & Strielkowski, W. Public support for sustainable development and environmental policy: a case of the Czech Republic. Sustain. Dev. 30 , 110–126 (2022).

Lu, J. & Yu, X. Does the internet make us more intolerant? A contextual analysis in 33 countries. Inf. Commun. Soc. 23 , 252–266 (2020).

David, Y. Public opinion, media and activism: the differentiating role of media use and perceptions of public opinion on political behaviour. Soc. Mov. Stud. 21 , 334–354 (2022).

Back, E. A., Back, H., Fredén, A. & Gustafsson, N. A social safety net? Rejection sensitivity and political opinion sharing among young people in social media. New Media Soc. 21 , 298–316 (2019).

Ye, Y., Xu, P. & Zhang, M. Social media, public discourse and civic engagement in modern China. Telemat. Inform. 34 , 705–714 (2017).

Salman, A., Yusoff, M. A., Mohamad Salleh, M. A. & Hj Abdullah, M. Y. Pengunaan media sosial untuk sokongan politik di Malaysia (The use of social media for political support in Malaysia). J. Nusantara Stud. 3 , 51–63 (2018).

Mueller, K. & Schwarz, C. Fanning the flames of hate: social media and hate crime. J. Eur. Econ. Assoc. 19 , 2131–2167 (2021).

Bursztyn, L., Egorov, G., Enikolopov, R. & Petrova, M. Social media and xenophobia: evidence from Russia. Tech. Rep. w26567, National Bureau of Economic Research, Cambridge, MA (NBER, 2019). http://www.nber.org/papers/w26567.pdf

Download references

Acknowledgements

We thank S. Munzert for providing his perspective on causal inference and issues specific to political science, D. Ain for editing the manuscript and F. Stock for help in the literature comparison. P.L.-S., S.L. and R.H. acknowledge financial support from the Volkswagen Foundation (grant ‘Reclaiming individual autonomy and democratic discourse online: How to rebalance human and algorithmic decision-making’). S.L. acknowledges support from the Humboldt Foundation through a research award and partial support by an ERC Advanced Grant (PRODEMINFO) during completion of this paper. L.O. acknowledges financial support by the German National Academic Foundation in the form of a PhD scholarship. The authors received no specific funding for this work. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Open access funding provided by Max Planck Society.

Author information

These authors contributed equally: Philipp Lorenz-Spreen and Lisa Oswald.

Authors and Affiliations

Center for Adaptive Rationality, Max Planck Institute for Human Development, Berlin, Germany

Philipp Lorenz-Spreen & Ralph Hertwig

Hertie School, Berlin, Germany

Lisa Oswald

School of Psychological Science and Cabot Institute, University of Bristol, Bristol, UK

• Stephan Lewandowsky

School of Psychological Science, University of Western Australia, Perth, Australia

You can also search for this author in PubMed   Google Scholar

Contributions

All authors designed the study. P.L.-S. and L.O. selected and coded the literature. P.L.-S. and L.O. evaluated the coded tables. All authors analysed the results and wrote the manuscript.

Corresponding author

Correspondence to Philipp Lorenz-Spreen .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Peer review

Peer review information.

Nature Human Behaviour thanks Christian von Sikorski and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary information.

Supplementary Figs. 1–6, Tables 1 and 2, Info box and references.

Reporting Summary

Peer review file, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Lorenz-Spreen, P., Oswald, L., Lewandowsky, S. et al. A systematic review of worldwide causal and correlational evidence on digital media and democracy. Nat Hum Behav 7 , 74–101 (2023). https://doi.org/10.1038/s41562-022-01460-1

Download citation

Received : 01 December 2021

Accepted : 16 September 2022

Published : 07 November 2022

Issue Date : January 2023

DOI : https://doi.org/10.1038/s41562-022-01460-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

Thinking clearly about misinformation.

• Li Qian Tay
• Ullrich K. H. Ecker

Communications Psychology (2024)

A framework for promoting online prosocial behavior via digital interventions

• David J. Grüning
• Julia Kamin
• Philipp Lorenz-Spreen

Invasiveness, Intrusiveness and Influence: three new metrics to measure communication between political digital echo chambers

• Eglantina Kalluçi
• Ridvan Peshkopia

Social Network Analysis and Mining (2024)

Skewed perspectives: examining the influence of engagement maximization on content diversity in social media feeds

• Paul Bouchaud

Journal of Computational Social Science (2024)

Media Reform as Transformation Tool: A Hegemonic Gap in Environmental Research and Policy

• Myanna Lahsen

International Journal of Politics, Culture, and Society (2024)

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Articles on Correlation

Displaying all articles.

People don’t mate randomly – but the flawed assumption that they do is an essential part of many studies linking genes to diseases and traits

Richard Border , University of California, Los Angeles and Noah Zaitlen , University of California, Los Angeles

A recent study suggests red wine may protect you from COVID. But I wouldn’t drink to this yet

Hassan Vally , Deakin University

Is when you lose your virginity and have your first kid really written in your genes? Not quite

Andrew Shelling , University of Auckland, Waipapa Taumata Rau

6 tips to help you detect fake science news

Marc Zimmer , Connecticut College

Is this study legit? 5 questions to ask when reading news stories of medical research

Hassan Vally , La Trobe University

There’s no evidence caesarean sections cause autism or ADHD

Andrew Whitehouse , The University of Western Australia

Numbers in the news? Make sure you don’t fall for these 3 statistical tricks

Liberty Vittert , Washington University in St. Louis

How Cambridge Analytica’s Facebook targeting model really worked – according to the person who built it

Matthew Hindman , George Washington University

The seven deadly sins of statistical misinterpretation, and how to avoid them

Winnifred Louis , The University of Queensland and Cassandra Chapman , The University of Queensland

Why so many baseball experts whiffed with last year’s predictions

James Walker , Saint Xavier University and Robert Bellamy , Duquesne University

Clearing up confusion between correlation and causation

Jonathan Borwein (Jon) , University of Newcastle and Michael Rose , University of Newcastle

Related Topics

• Bad science
• Cambridge Analytica
• Epidemiology
• Genetic correlation
• Science journalism

Top contributors

Associate Professor, Epidemiology, Deakin University

Professor, Social Psychology, The University of Queensland

Bennett Chair of Autism, Telethon Kids Institute, The University of Western Australia

PhD Candidate, School of Mathematical and Physical Sciences, University of Newcastle

Distinguished Professor of Genetics and Vice Chancellor's Fellow, La Trobe University

Emeritus Professor of Communication, Saint Xavier University

Professor; Department of Journalism and Multimedia Arts [JMA], Duquesne University

Professor of Chemistry, Connecticut College

Lecturer in Marketing & ARC DECRA Fellow, The University of Queensland

Professor of Media and Public Affairs, George Washington University

Professor of the Practice of Data Science, Washington University in St. Louis

Professor and Associate Dean (Research), University of Auckland, Waipapa Taumata Rau

Professor of Neurology and Human Genetics, University of California, Los Angeles

Laureate Professor of Mathematics, University of Newcastle

Postdoctoral Researcher in Statistical Genetics, University of California, Los Angeles

• X (Twitter)
• Unfollow topic Follow topic

• Primary research
• Open access
• Published: 28 August 2013

Correlational study: illness representations and coping styles in caregivers for individuals with schizophrenia

• Shyhrete Rexhaj 1 , 2 ,
• Nataly Viens Python 2 ,
• Diane Morin 3 ,
• Charles Bonsack 1 &
• Jérôme Favrod 1 , 2  

Annals of General Psychiatry volume  12 , Article number:  27 ( 2013 ) Cite this article

30k Accesses

25 Citations

2 Altmetric

Metrics details

Caring for individuals with schizophrenia can create distress for caregivers which can, in turn, have a harmful impact on patient progress. There could be a better understanding of the connections between caregivers’ representations of schizophrenia and coping styles. This study aims at exploring those connections.

This correlational descriptive study was conducted with 92 caregivers of individuals suffering from schizophrenia. The participants completed three questionnaires translated and validated in French: (a) a socio-demographic questionnaire, (b) the Illness Perception Questionnaire for Schizophrenia and (c) the Family Coping Questionnaire.

Our results show that illness representations are slightly correlated with coping styles. More specifically, emotional representations are correlated to an emotion-focused coping style centred on coercion, avoidance and resignation.

Our results are coherent with the Commonsense Model of Self-Regulation of Health and Illness and should enable to develop new interventions for caregivers.

Problem statement

Between 30% and 91% of individuals with schizophrenia live in a family setting [ 1 – 3 ]. The decreased length of stay in hospital and restrictions on involuntary treatments mean that family-based caregivers provide an important support during periods of psychological instability. This support implies that caregivers rely on a variety of strategies to confront the consequences resulting from the psychological instability of the schizophrenia patient.

Burden, distress, illness representations and coping strategies

The literature indicates that there are interactions between the different concepts of family burden, distress, illness representations, the expressed emotion (EE) and coping strategies. It has been demonstrated that caregivers’ representations of negative consequences of the illness for the patient are positively correlated with their objective burden, while representations of the negative consequences of schizophrenia for the caregiver are positively correlated to a subjective burden [ 4 , 5 ]. The caregivers’ subjective burden is also associated with negative emotional responses to the illness [ 5 ]. Also, caregivers having an elevated and hostile type of EE usually consider the patient responsible for the causes of illness [ 6 , 7 ]. Furthermore, in comparison with caregivers having a low EE, those having a high, critical-type EE tend to underestimate their possibilities of controlling problems themselves and perceive the illness as unlikely to be controlled by treatment as well as attribute more negative consequences of the illness for the patient and themselves [ 4 ].

Distress experienced by caregivers is also correlated with illness representations such as the following: (1) illness outbreaks are chronic, (2) the feeling that treatment does not help control the illness, (3) the perception that the patient can have greater personal control over the illness, (4) the perception that the illness brings about negative consequences for the patient and (5) that the illness brings about negative consequences for the caregiver and (6) representations that the illness elicits painful emotions such as anxiety and fear [ 4 , 5 , 8 ].

Coping strategies by caregivers of patients with schizophrenia have been explored by several authors who mainly based their observations using models or frameworks derived from Lazarus and Folkman’s theory of stress coping [ 9 – 12 ]. A review of the literature showed that there are strategies that are more or less efficient for confronting stress. More specifically for caregivers, strategies like coercion, avoidance and resignation are associated with suffering and patient relapse [ 13 , 14 ].

Birchwood and Cochrane [ 14 ] explored coping strategies used by caregivers of schizophrenia patients. Their results detail eight essential coping categories: coercion, avoidance, ignorance/acceptance, constructive, resignation, reassurance, disorganization and collusion. They also showed that coercion is the first predictor in patient relapses. As concerns specificities of different coping styles, Magliano et al. [ 15 ] conducted a study which explored coping strategies in relation to physical and somatic symptoms of caregivers as well as the association between these two variables. The Family Coping Questionnaire (FCQ) was used [ 10 ]. Their results indicated that emotion-focused coping (coercion, avoidance and resignation) is positively correlated with participant anxiety and depression [ 15 ]. In another study, Knudson and Coyle came up with a procedure which incited certain caregivers to modify their coping strategies [ 11 ]. In their qualitative study, the caregivers of individuals with schizophrenia were invited to describe their coping strategy. At the onset of the illness, the caregivers tended to use a problem-focused coping strategy. However, if the symptoms became persistent, they progressively opted for emotion-focused strategies, which enabled them to attain a position of acceptance and, ultimately, of well-being. Nevertheless, the conceptualization of emotion-focused coping done according to factor analysis in the study of Magliano et al. only included strategies such as resignation, avoidance and coercion [ 10 ]. Acceptance of the illness has seldom been studied and could be a potential functional human response in some instances.

Magliano et al. reported that the high level of burden is associated with a reduction in social interests, a reduction of social support network as well as a resignation and an avoidance of contact with the family member suffering from schizophrenia. It also appears that in the absence of any specific intervention with these caregivers, this result can remain unchanged for as long as a year [ 15 ]. Specific to the subjective burden, it appears that emotional and cognitive reactions are associated with the use of coping strategies that are specifically emotional, such as avoidance and isolation [ 16 , 17 ]. These strategies have been identified as being linked to an increase in caregiver distress [ 18 ]. Increased distress can result in a deterioration of the family atmosphere through elevated EE levels [ 12 ]. These elements lead to a further increase in burden [ 19 ].

Theoretical framework

In terms of theory, in the 1960s, Leventhal and his colleagues developed the Commonsense Model of Self-Regulation of Health and Illness (SRM) [ 20 ]. The results from their initial research indicated that modifying the personal representations of health and the development of an action plan were the two determining factors for the creation of health-promoting actions. Their conclusions led to the development of a series of other studies meant to define the different characteristics of health and illness representations. According to the model, representations are cognitive and emotional constructions of a health problem. The SRM includes eight concepts or dimensions: (1) internal and external stimuli, (2) treatment system, (3) representation of the illness and the treatment, (4) coping procedures, (5) evaluation of the cognitive treatment of the information, (6) emotional representation, (7) coping responses and (8) evaluation of the emotional treatment of the information.

Coping involves procedures which enable an individual to collect information and control the problem, as well as different responses such as distraction or relaxation. Coping can be described as having two primary functions. The first is centred on problem solving and the second on the immediate regulation of the emotion elicited by the problem. According to the SRM, the choice of a coping strategy used by caregivers is associated with the kind of representation of the illness that the caregiver has developed. In this way, the caregiver’s explicative model seems to have an influence on their choice of coping strategy. This choice can then, in turn, lighten or increase caregiver burden [ 17 ]. Effectively, negative representations of the illness can lead to the use of unsuitable coping strategies [ 21 , 22 ].

Goal of the study

The goal of this study was to explore the associations between illness representations and three forms of coping styles—(1) problem-focused coping, (2) emotion-focused coping and (3) social support-focused coping—for caregivers of individuals with schizophrenia.

Design and recruitment

This correlational descriptive study was conducted with 92 caregivers of individuals with schizophrenia. Participants were members of French-speaking social support organizations, were recruited using a convenience sampling strategy and met the following criteria: (1) being 18 years or older, (2) living in Switzerland or France, (3) being able to speak French, (4) acting as carer for a family member or close friend who suffers from schizophrenia and (5) having had at least a 1-h contact with this person in the last month. As the first step of the recruiting strategy, social support organizations were met at the time of meeting with members in order to provide a complete presentation of the study. Then, the members who were present received an envelope containing an informed consent sheet, an information letter and the necessary questionnaires, as well as a pre-addressed and pre-stamped envelope for the return of all completed documents.

A second recruiting strategy was conducted using an electronic survey. The presidents of the social support organizations sent to all members of their respective groups the survey link leading to a dedicated website that included the same documentation as the paper version.

Instruments

Participants filled out three self-administered forms: (a) a socio-demographic questionnaire, (b) the Illness Perception Questionnaire for Schizophrenia : Relatives’ version and (c) the Family Coping Questionnaire. Authorization to translate these questionnaires into French using independent backward translation was granted by the authors.

The socio-demographic questionnaire

According to the literature, the most significant socio-demographic variables which influence illness representations and coping strategies are the following: (1) caregiver gender [ 23 ], (2) caregiver age, (3) caregiver education level [ 13 ], (4) the length of contact with the patient [ 19 , 24 ], (5) professional and social support [ 25 , 26 ] such as involvement in a Profamille program [ 27 ], (6) the nature of their connection with the patient [ 28 ], (7) the length of the illness [ 29 ], (8) whether the caregiver lives with the patient or not [ 30 ], (9) age of the patient, (10) patient gender and (11) the patient’s professional support. With the further goal of comparing our results with other studies already reported in the literature, we collected data on the number of people living in a household and the civil status of the caregivers.

The Illness Perception Questionnaire for Schizophrenia: Relatives’ version

Illness representations were measured using a self-administered questionnaire entitled ‘Illness Perception Questionnaire for Schizophrenia: Relatives’ version (IPQS: Relatives)’ [ 5 ]. It involves 13 sub-scales (150 items): identity, timeline (acute/chronic), timeline (cyclic), negative consequences for the individual, negative consequences for the caregiver, personal control (feeling of powerlessness—patient), personal control (feeling of powerlessness—caregiver), personal blame (patient responsibility), personal blame (caregiver responsibility), therapeutic control, mental health problem coherence, emotional representations and causes. For the current study, the identity and the causes of the illness were not considered. The responses provided by the caregivers regarding their perceptions of the illness were measured using a 5-level Likert scale: 0 = do not agree at all, 1 = do not agree, 2 = more or less agree = 3 = agree and 4 = completely agree. Results show a dimensional reliability situated between α = 0.63 and α = 0.83 [ 5 ].

The Family Coping Questionnaire

Coping styles were measured using a self-administered questionnaire entitled ‘Family Coping Questionnaire (FCQ)’. The version of the FCQ, used in this study, included 27 items measuring seven dimensions (information gathering, positive communication, social involvement, coercion, avoidance, resignation, the patient’s social involvement). A factor analysis enabled the authors to identify three main factors. Looking at these three factors and at the conceptual definition of coping, it is clear that the seven dimensions can be regrouped into three coping modes: (1) problem-focused coping (the patient’s social involvement, positive communication and information gathering are positively correlated with each other and negatively with avoidance, (2) emotion-focused coping (coercion, avoidance and resignation are positively correlated) and (3) social support-focused coping (social involvement is associated with avoidance) [ 10 ]. Our study catalogued these strategies. Caregiver responses to different situations were measured using a 5-level Likert scale: 1 = never, 2 = rarely, 3 = sometimes, 4 = very often and 5 = not applicable. Its validity was demonstrated during the BIOMED 1 study, conducted in five European countries [ 15 ]. Results showed a dimensional reliability between α = 0.68 and α = 0.83 [ 13 ].

Data analysis

Data were treated using the computing program ‘IBM SPSS Statistics® version 20’. Descriptive statistics were used to describe socio-demographic characteristics. To test associations between illness representations and coping styles, bivariate correlational analyses were used. The p value threshold used was set to <0.05.

Ethical considerations

All participants were required to sign an informed consent form for the paper version of the survey or confirm their consent in order to access the electronic version of the survey. The research protocol received full authorization by the Canton of Vaud’s Ethics Committee for human-based research.

Results and discussion

While Table  1 presents the socio-demographic characteristics of participants, Table  2 presents the main characteristics of the individuals with schizophrenia cared for. As showed, the final sample involved 92 middle-aged individuals (mean age = 56 years, standard deviation (SD) = 12.6) including 68% of women. Most participants were married or lived in a household as a married couple (70.3%). The majority of respondents were either mothers or fathers (66.3%) of the individual with schizophrenia. All participants had completed some level of education; vocational school, apprenticeship and university-level education were the most often cited. Most participants reported that they lived with one or several other individuals. Most of them (62%) participated to the Profamille program, a psycho-educational program for family members and friends of individuals with schizophrenia [ 31 ]. Among the 92 participants, 37 (40.2%) reported that they lived with the schizophrenia patient.

As concerns the individual with schizophrenia, results show that the duration of the illness is long (mean = 15.1 years, SD = 9.7). The mean age of the individual with schizophrenia taken care for by the participants of this study was 34 years and 77.2% were male. Individuals with schizophrenia call upon community services in a variety of ways. In 88% of all cases, a psychiatrist was involved in patient follow-up. In 34.8% of all cases, a general medical practitioner was involved in patient follow-up. In our sample, 33.7% of the schizophrenia patients received nursing care. Also, a significant number (33.7%) of the individuals with schizophrenia in our sample were involved in sheltered workshops. Other community services were also used at more than 10% (social worker, 20.7%; psychologist, 13%; day centre, 17.4%).

Results presented in Table  3 show that caregivers have a great conception of the illness as being chronic and as having recurring symptoms per cycle. They have a strong perception that the illness brings about negative consequences for the individual with schizophrenia. While at the same time, they perceive fewer negative consequences for themselves caused by the patient’s illness. They have a feeling of control over the illness and that the patient can also have control. They do not think that the patient is responsible for his or her illness. They do not see themselves as responsible for the appearance of the illness. Treatment is perceived as helpful for better controlling the illness. For these caregivers, the illness has meaning and coherence. They more or less agree that the illness brings about negative emotions like sorrow or anxiety, measured here under the category of emotional representations. Participants in this study mostly used problem-focused and social support-focused coping styles. Emotion-focused coping was used less compared to the two other styles (median = 15, min to max = 7.00–28.00).

Correlations between illness representations and coping styles are presented in Table  4 . It can be seen that all the statistically significant correlations are somewhat low ranging from r = 0.23 to r = 0.41. The statistically significant correlations are the following:

Representations that the illness brings about negative consequences for the caregiver have a moderate positive correlation with problem-focused coping ( r = 0.31, p = 0.006) and emotion-focused coping ( r = 0.35, p = 0.002). There was also a slight negative correlation between these representations and social support-focused coping ( r = −0.29, p = 0.012).

Representations that the illness brings about negative consequences for the patient have a slight positive correlation with emotion-focused coping ( r = 0.23, p = 0.037).

The presence of a feeling of control by the caregiver is moderately positively correlated with problem-focused coping ( r = 0.31, p = 0.006).

Representations that the caregiver might be responsible for the onset of the illness have a moderate positive correlation with emotion-focused coping ( r = 0.34, p = 0.001).

Representations that the patient is responsible for the onset of the illness have a slight positive correlation with emotion-focused coping ( r = 0.25, p = 0.020).

Representations that treatment helps to control the illness have a slight positive correlation with problem-focused coping ( r = 0.23, p = 0.040).

A lack of meaning attributed to the illness has a moderate positive correlation with emotion-focused coping ( r = 0.31, p = 0.005).

Representations that the illness brings about negative emotions have a moderate positive correlation with emotion-focused coping ( r = 0.41, p = 0.000) and a moderate negative correlation with social support-focused coping ( r = −0.39, p = 0.000).

It is timely to note that this study is the first known to use French-validated tools to measure correlations between representations of schizophrenia and coping styles within a caregiver sample. For most variables, our results align with several other studies in this field [ 8 , 9 , 13 , 15 , 32 ]. Nevertheless, the caregivers in this study were less likely to have the perception that the patient was at fault or that they themselves were responsible for the onset of the illness than compared to the results published by Lobban and collaborators [ 5 ]. Our results also show that emotion-focused coping is moderately correlated with (1) negative consequences for the caregivers, (2) the feeling of being at fault, (3) the feeling that the mental health problem is not coherent and (4) the overall score of the emotional representation scale. Problem-focused coping is itself moderately correlated with (1) the perception of negative consequences for caregivers and (2) the feeling of control. Social support-focused coping has a moderate negative correlation with the overall scores of the emotional representation scale. The other correlations are non-significant or less than 0.30 with a slight scaling effect [ 33 ].

The SRM predicts that illness representations influence coping procedures and that emotional representations influence coping responses. These different factors also influence each other. More specifically, when a health threat occurs, it is handled by the processing system , the second concept of SRM. This system consists of two types of threat management: (1) the representations of illness and treatment and (2) the emotional representations. Both types can influence each other. Each type of threat management includes three steps to process information: representation (illness and treatment), coping procedures and evaluation. Regarding the first type of threat management, the first stage of data processing is the development of a representation of the disease and the treatment. This concept determines the goals and coping procedures to achieve them [ 34 ]. Coping procedures are the second stage of information processing. They are a wide range of cognitive and behavioural measures undertaken in response to the cognitive representation, such as problem-focused coping [ 22 , 34 ]. The third step of information processing is the evaluation of the cognitive treatment of information. Regarding the second type of threat management, the first step of information processing is the formation of an emotional representation [ 22 ]. Based on this emotional representation, goals are set and coping responses are determined to achieve them. Coping responses are the second step of information processing. They are a wide range of strategies focused on managing emotions that are applied in response to the emotional representation. The last step of information processing for this second type is the evaluation of the emotional treatment of the information. The results of the evaluation step produce a feedback at the preceding steps .

Our results are quite coherent with this predictive model, especially for emotion-focused coping and the score of emotional representations, consequences for the caregiver and the feeling of being at fault. Consequences for the patient and the feeling of control are associated with problem-focused coping. The fact that the consequences for the caregiver are associated as much with problem-focused coping as with emotion-focused coping underscores the presence of interactions between the emotional and cognitive variables.

A study similar to ours [ 8 ] did not establish a positive correlation between the representations of the illness that bring about negative consequences for the caregiver and problem-focused coping. It is important to note that these authors did not conceptualize coping as we did, and therefore, comparing our results is a limited exercise. However, the abovementioned correlation leads one to think that the Profamille psycho-educational program followed by a significant portion of our participants had an impact on this coping style without actually modifying the representations of the illness as conceived by the participants.

Our sample pool was particular in that it was composed of individuals belonging to family-type assistance programs and that most of the participants had followed some kind of psycho-educational program. On average, our participants had been caring for their patient for 15 years. This could explain the weak level of reliance on emotion-focused coping. Nevertheless, the correlations observed suggest that the SRM model is still valid over time.

In terms of intervention, the psycho-educational programs available for the caregivers tend to focus on knowledge acquisition of the illness and treatment and positive communication skills training to reduce stress within the family. Our data suggest that it would be of interest to take better into account the emotional representations of the illness. In everyday practice, first-line health care professionals have to care for caregivers’ painful emotions such as anger, guilt, sadness and fear. However, the required skills to help caregivers to better manage their painful emotions are difficult to define precisely. The development and evaluation of the best professional strategies to help caregivers to better manage their painful emotions appear to be an interesting line of research for the future. Refining the best professionals’ skills to help caregivers to deal with negative emotions may influence more directly emotional representations and coping skills in caregivers.

Given the scarcity of studies to compare to, further studies could be useful to better identify coping styles that improve emotional regulation without negative consequences for the patient, like acceptance or changing value systems. In this way, a holistic understanding of care, as suggested by several authors [ 35 , 36 ], could be further promoted. Thus, interventions could be focused on the development of alternatives to less effective or even painful strategies like resignation, coercion and avoidance. Furthermore, the tools used to measure emotion-focused coping tend to highlight more negative emotional coping methods like coercion, avoidance or resignation. It is important to develop a measuring instrument that also takes into account the positive strategies focused on emotion.

Limitations of the study

This was the first study using the IPQS: Relatives and the FCQ in French. Participant recruitment took place within the context of social support organizations according to a convenience sampling method. Therefore, our results could be cautiously generalized to other programs having a similar context to ours, in terms of culture and health care systems.

In the empirical literature, as well as in the Commonsense Model of Self-Regulation of Health and Illness, problem-focused and social support-focused coping styles are recognized as efficient for decreasing caregiver loads and increasing a patient’s chances of better coping with schizophrenia. The results of this study also show that illness representations influence the choice of coping styles. The coping style approach is useful to develop further targeted and feasible interventions. Further studies in this field should focus on the evaluation of interventions based on illness representations in order to prevent the use of unsuitable coping styles. Also, it will be necessary to identify this phenomenon in a caregiver population of individuals who are not involved into any mutual support group in order to adapt appropriate care and support procedures.

Abbreviations

Expressed emotion

Family coping questionnaire

Relatives, illness perception questionnaire for schizophrenia: relatives’ version

Standard deviation

Commonsense model of self-regulation of health and illness.

van Wijngaarden B, Schene A, Koeter M, Becker T, Knapp M, Knudsen HC, Tansella M, Thornicroft G, Vazquez-Barquero JL, Lasalvia A, Leese M, EPSILON Study Group: People with schizophrenia in five countries: conceptual similarities and intercultural differences in family caregiving. Schizophrenia Bulletin. 2003, 29 (3): 573-586. 10.1093/oxfordjournals.schbul.a007029.

Article   PubMed   Google Scholar  

Heider D, Angermeyer MC, Winkler I, Schomerus G, Bebbington PE, Brugha T, Azorin JM, Toumi M: A prospective study of quality of life in schizophrenia in three European countries. Schizophrenia Research. 2007, 93 (1–3): 194-202.

Lauber C, Eichenberger A, Luginbuhl P, Keller C, Rossler W: Determinants of burden in caregivers of patients with exacerbating schizophrenia. European Psychiatry: The Journal of the Association of European Psychiatrists. 2003, 18 (6): 285-289. 10.1016/j.eurpsy.2003.06.004.

Article   Google Scholar  

Barrowclough C, Lobban F, Hatton C, Quinn J: An investigation of models of illness in carers of schizophrenia patients using the illness perception questionnaire. The British Journal of Clinical Psychology/The British Psychological Society. 2001, 40 (Pt 4): 371-385.

Article   CAS   PubMed   Google Scholar  

Lobban F, Barrowclough C, Jones S: Assessing cognitive representations of mental health problems. II. The illness perception questionnaire for schizophrenia: relatives’ version. The British Journal of Clinical Psychology/The British Psychological Society. 2005, 44 (Pt 2): 163-179.

Barrowclough C, Tarrier N, Johnston M: Distress, expressed emotion, and attributions in relatives of schizophrenia patients. Schizophrenia Bulletin. 1996, 22 (4): 691-702. 10.1093/schbul/22.4.691.

Lobban F, Barrowclough C, Jones S: Does expressed emotion need to be understood within a more systemic framework? An examination of discrepancies in appraisals between patients diagnosed with schizophrenia and their relatives. Social Psychiatry and Psychiatric Epidemiology. 2006, 41 (1): 50-55. 10.1007/s00127-005-0993-z.

Fortune DG, Smith JV, Garvey K: Perceptions of psychosis, coping, appraisals, and psychological distress in the relatives of patients with schizophrenia: an exploration using self-regulation theory. The British Journal of Clinical Psychology/The British Psychological Society. 2005, 44 (Pt 3): 319-331.

Moller-Leimkuhler AM: Burden of relatives and predictors of burden. Baseline results from the Munich 5-year-follow-up study on relatives of first hospitalized patients with schizophrenia or depression. European Archives of Psychiatry and Clinical Neuroscience. 2005, 255 (4): 223-231. 10.1007/s00406-004-0550-x.

Magliano L, Guarneri M, Marasco C, Tosini P, Morosini PL, Maj M: A new questionnaire assessing coping strategies in relatives of patients with schizophrenia: development and factor analysis. Acta Psychiatrica Scandinavica. 1996, 94 (4): 224-228. 10.1111/j.1600-0447.1996.tb09853.x.

Knudson B, Coyle A: Parents’ experiences of caring for sons and daughters with schizophrenia: a qualitative analysis of coping. European Journal of Psychotherapy & Counselling. 2002, 5 (5): 169-183.

Barrowclough C, Parle M: Appraisal, psychological adjustment and expressed emotion in relatives of patients suffering from schizophrenia. The British Journal of Psychiatry: The Journal of Mental Science. 1997, 171: 26-30. 10.1192/bjp.171.1.26.

Article   CAS   Google Scholar  

Magliano L, Fadden G, Economou M, Xavier M, Held T, Guarneri M, Marasco C, Tosini P, Maj M: Social and clinical factors influencing the choice of coping strategies in relatives of patients with schizophrenia: results of the BIOMED I study. Social Psychiatry and Psychiatric Epidemiology. 1998, 33 (9): 413-419. 10.1007/s001270050074.

Birchwood M, Cochrane R: Families coping with schizophrenia: coping styles, their origins and correlates. Psychological Medicine. 1990, 20 (4): 857-865. 10.1017/S0033291700036552.

Magliano L, Fadden G, Economou M, Held T, Xavier M, Guarneri M, Malangone C, Marasco C, Maj M: Family burden and coping strategies in schizophrenia: 1-year follow-up data from the BIOMED I study. Social Psychiatry and Psychiatric Epidemiology. 2000, 35 (3): 109-115. 10.1007/s001270050192.

Gonzalez-Torres MA, Oraa R, Aristegui M, Fernandez-Rivas A, Guimon J: Stigma and discrimination towards people with schizophrenia and their family members. A qualitative study with focus groups. Social Psychiatry and Psychiatric Epidemiology. 2007, 42 (1): 14-23. 10.1007/s00127-006-0126-3.

Magliano L, Fadden G, Madianos M, de Almeida JM, Held T, Guarneri M, Marasco C, Tosini P, Maj M: Burden on the families of patients with schizophrenia: results of the BIOMED I study. Social Psychiatry and Psychiatric Epidemiology. 1998, 33 (9): 405-412. 10.1007/s001270050073.

Magliano L, Veltro F, Guarneri M, Marasco C: Clinical and socio-demographic correlates of coping strategies in relatives of schizophrenic patients. European Psychiatry: The Journal of the Association of European Psychiatrists. 1995, 10 (3): 155-158. 10.1016/0767-399X(96)80104-0.

Madianos M, Economou M, Dafni O, Koukia E, Palli A, Rogakou E: Family disruption, economic hardship and psychological distress in schizophrenia: can they be measured?. European Psychiatry: The Journal of the Association of European Psychiatrists. 2004, 19 (7): 408-414. 10.1016/j.eurpsy.2004.06.028.

Leventhal H, Leventhal EA, Schafer P: Vigilant coping and health behavior: a life span problem. Aging, Health, and Behavior. Edited by: IMORA. 1991, Baltimore: Johns Hopkins, 109-140.

Google Scholar  

Magliano L, Guarneri M, Fiorillo A, Marasco C, Malangone C, Maj M: A multicenter Italian study of patients’ relatives’ beliefs about schizophrenia. Psychiatric Services. 2001, 52 (11): 1528-1530. 10.1176/appi.ps.52.11.1528.

Leventhal H, Leventhal EA, Cameron LD: Representations, procedures, and affect in illness self-regulation: a perceptual-cognitive approach. Handbook of Health Psychology. Edited by: Baum AS, Revenson TA, Singer JE. 2001, New York: Erbaum, 19-48.

Mitsonis C, Voussoura E, Dimopoulos N, Psarra V, Kararizou E, Latzouraki E, Zervas I, Katsanou MN: Factors associated with caregiver psychological distress in chronic schizophrenia. Social Psychiatry and Psychiatric Epidemiology. 2010, 47 (2): 331-337.

Koukia E, Madianos MG: Is psychosocial rehabilitation of schizophrenic patients preventing family burden? A comparative study. Journal of Psychiatric and Mental Health Nursing. 2005, 12 (4): 415-422. 10.1111/j.1365-2850.2005.00852.x.

Magliano L, Fiorillo A, Malangone C, Marasco C, Guarneri M, Maj M, National Mental Health Project Working Group: The effect of social network on burden and pessimism in relatives of patients with schizophrenia. The American Journal of Orthopsychiatry. 2003, 73 (3): 302-309.

Magliano L, Fiorillo A, Malangone C, De Rosa C, Maj M, Family Intervention Working Group: Implementing psychoeducational interventions in Italy for patients with schizophrenia and their families. Psychiatric Services. 2006, 57 (2): 266-269. 10.1176/appi.ps.57.2.266.

Hodé Y, Krychowski R, Beck N, Vonthron R, Rouyèr N, Huentz M, Risacher P, Fattah S, Pasche M, Pasche P: Effet d’un programme psychoéducatif sur l’humeur des familles des malades soufrant de schizophrénie. Journal de thérapie comportementale et cognitive. 2008, 18: 104-107. 10.1016/j.jtcc.2008.06.003.

Stalberg G, Ekerwald H, Hultman CM: At issue: siblings of patients with schizophrenia: sibling bond, coping patterns, and fear of possible schizophrenia heredity. Schizophrenia Bulletin. 2004, 30 (2): 445-458. 10.1093/oxfordjournals.schbul.a007091.

Moller-Leimkuhler AM: Multivariate prediction of relatives’ stress outcome one year after first hospitalization of schizophrenic and depressed patients. European Archives of Psychiatry and Clinical Neuroscience. 2006, 256 (2): 122-130. 10.1007/s00406-005-0619-1.

Schene AH, van Wijngaarden B: A survey of an organization for families of patients with serious mental illness in The Netherlands. Psychiatric Services. 1995, 46 (8): 807-813.

Cromier H, Guimond G, Joncas J, Leblanc G, Morin R, Vaillancourt S: Profamille, programme d’intervention de groupe auprès des familles de personnes atteintes de schizophrénie. 1991, Québec: Unité de psychiatrie sociale et préventive du CHUL et Centre de recherche de l’Université Laval Robert Giffard

Kartalova-O’Doherty Y, Doherty DT: Coping strategies and styles of family carers of persons with enduring mental illness: a mixed methods analysis. Scandinavian Journal of Caring Sciences. 2008, 22 (1): 19-28. 10.1111/j.1471-6712.2007.00583.x.

Article   PubMed Central   PubMed   Google Scholar  

Field AP: Discovering Statistics Using SPSS: (and Sex and Drugs and Rock 'n' Roll). 2009, Thousand Oaks: Sage, 3

Leventhal H, Leventhal EA, Contrada RJ: Self regulation, health, and behavior: a perceptual-cognitive approach. Psychology and Health. 1998, 13: 717-733. 10.1080/08870449808407425.

Dallaire C, Toupin I: Les théories à spectre modéré, les théories propres aux situations et les interventions infirmières. Le savoir infirmier: au coeur de la discipline et de la profession infirmière. Edited by: Dallaire IC. 2008, Boucherville: Gaëtan Morin, 239-264.

Pépin J, Kérouac S, Ducharme F: La pensée infirmière. 2010, Montréal: Chenelière Éducation, 3

Download references

Acknowledgements

This work has been supported by a donation from Dr. Alexander Engelhorn.

Author information

Authors and affiliations.

Community Psychiatry Service, Department of Psychiatry, University Hospital Centre of Vaud, Site de Cery, Prilly, 1008, Switzerland

Shyhrete Rexhaj, Charles Bonsack & Jérôme Favrod

School of Nursing Sciences, La Source, University of Applied Sciences of Western Switzerland, Av. Vinet 30, Lausanne, 1004, Switzerland

Shyhrete Rexhaj, Nataly Viens Python & Jérôme Favrod

Institut Universitaire de Formation et de Recherche en Soins, University of Lausanne, Biopôle 2, Rue de la Corniche 10, Lausanne, 1010, Switzerland

Diane Morin

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Shyhrete Rexhaj .

Additional information

Competing interests.

The authors declare that they have no competing interests.

Authors’ contributions

SR, DM and JF contributed to the conception and design of the study. SR contributed to the acquisition of data. SR and JF performed the statistical analysis and drafted the first manuscript. NVP and CB critically reviewed and revised the manuscript. All authors read and approved the final manuscript.

Authors’ original submitted files for images

Below are the links to the authors’ original submitted files for images.

Authors’ original file for figure 1

Authors’ original file for figure 2, authors’ original file for figure 3, authors’ original file for figure 4, rights and permissions.

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( http://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and permissions

About this article

Cite this article.

Rexhaj, S., Python, N.V., Morin, D. et al. Correlational study: illness representations and coping styles in caregivers for individuals with schizophrenia. Ann Gen Psychiatry 12 , 27 (2013). https://doi.org/10.1186/1744-859X-12-27

Download citation

Received : 12 June 2013

Accepted : 20 August 2013

Published : 28 August 2013

DOI : https://doi.org/10.1186/1744-859X-12-27

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Representations of schizophrenia
• Nursing care

Annals of General Psychiatry

ISSN: 1744-859X

• Submission enquiries: [email protected]

Social Sciences

© 2024 Inquiries Journal/Student Pulse LLC . All rights reserved. ISSN: 2153-5760.

Disclaimer: content on this website is for informational purposes only. It is not intended to provide medical or other professional advice. Moreover, the views expressed here do not necessarily represent the views of Inquiries Journal or Student Pulse, its owners, staff, contributors, or affiliates.

Home | Current Issue | Blog | Archives | About The Journal | Submissions Terms of Use :: Privacy Policy :: Contact

Need an Account?

Forgot password? Reset your password »

• Research Note
• Open access
• Published: 01 April 2024

Resilience and mindfulness among radiological personnel in Norway, their relationship and their impact on quality and safety– a questionnaire study

• Ann Mari Gransjøen 1 , 2  

BMC Research Notes volume  17 , Article number:  96 ( 2024 ) Cite this article

Metrics details

Stress and burnout are widespread problems among radiological personnel Individual and organizational resilience and mindfulness offer protection against burnout.

To investigate the level of resilience and mindfulness among radiological personnel, the associations between organizational resilience, individual resilience, and mindfulness, and how these factors impact the quality of care provided in radiological departments.

An online questionnaire consisting of the Connor-Davidson Resilience Scale, the Mindful Attention Awareness Scale, the Benchmark Resilience Tool, and questions regarding burnout, and quality and safety was used. Data analysis consisted of descriptive statistics, bivariate correlation and standard multiple regression.

Results and Conclusion

Few participants considered burnout a significant challenge. Individual and organizational resilience were low (30.40 ± 4.92 and 63.21 ± 13.63 respectively), and mindfulness was high (4.29 ± 0.88). There was a significant correlation between individual and organizational resilience ( p  = 0.004), between individual resilience and mindfulness ( p  = 0.03), and between organizational resilience and mindfulness ( p  = 0.02). Individual and organizational resilience affect each other. However; neither significantly affect quality and safety, nor mindfulness

Peer Review reports

Introduction

Stress and burnout are widespread problems among radiological personnel [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 ]. Different forms of mindfulness, ranging from formal meditation to more informal attention to day-to-day tasks, can prevent and reduce burnout among radiological personnel [ 3 ]. Another strategy for reducing stress and burnout among health professionals is promoting individual resilience [ 4 , 9 , 12 ]. Mindfulness and resilience could also affect the quality of care, due to their effectiveness in reducing stress and burnout [ 2 , 5 , 13 ].

Organizational resilience regards an organization’s ability to manage change, bounce back from setbacks and maintain desirable functions and outcomes under pressure. This is influenced by for example leadership practices and human capital [ 14 ]. Some studies show a link between individual resilience and organizational resilience, and that these two types of resilience affect each other [ 15 , 16 , 17 ].

The objective of this study is to investigate the level of resilience and mindfulness among radiological personnel, the associations between organizational resilience, individual resilience, and mindfulness, and how these factors impact the quality of care provided in radiological departments.

Materials and methods

Design and setting.

This study utilized a cross-sectional design to collect data on resilience, mindfulness, and the quality and safety of care among healthcare workers and their departments. The study is set within radiological departments in Norway, which encompasses both public hospitals and private institutions.

Population, study size and recruitment

The study population consisted of radiologists, registrars, radiographers, and radiation therapists. Participants were selected based on the following eligibility criteria; (a) they had a valid authorizations and (b) they currently worked in a clinical setting. According to an online sample size calculator (surveymonkey.com) the estimated sample size needed for this study, based on population size, 95% CI and 5% margin of error, was approximately 356 participants, which was not reached.

Participants were recruited in collaboration with the Norwegian Society of Radiographers and the Norwegian Radiological Association. These associations posted the link to a digital, online questionnaire on social media and their newsletter, resulting in probability sampling. Recruitment lasted from July 18th to October 5th 2022 and included a total of 4783 members.

Variables, data sources and measurement

The variables of interest in this study were individual resilience, mindfulness, organizational resilience, and quality and safety. Background variables that were used were public vs. private setting, and how leaders address burnout. All these variables were measured through a questionnaire consisting of six parts.

Not all parts had a Norwegian version available. The researcher, following the steps described by the Norwegian Directory of Health, translated this from English to Norwegian. Too see the interview guide used in the validation of the translated questionnaire, see supplementary file 2 .

Part 1 was designed by the researcher to collect demographic data about the respondent. This included profession, workplace (public vs. private), department size and whether their position included personnel management.

Part 2 is the Norwegian Connor-Davidson Resilience Scale (CD-RISC-10), which is used to assess the ability to respond and adapt to life adversity, trauma, tragedy, threats or other major life stressors [ 18 ].

Part 3 is the five item Mindfulness Attention Awareness Scale (MAAS) translated to Norwegian by Smith et al. This scale measures the extent to which an individual can attend to, and remain aware of, experiences in the present moment [ 19 ]..

Part 4 is the short version of the Benhmark Resilience Tool (BRT 13), which assesses behavioral traits and perceptions linked to the organization’s ability to plan for, respond to, and recover from emergencies and crises (organizational resilience) [ 20 ].

Parts 5 and 6 are aimed at specific groups. Part 5 was intended for respondents with personnel management roles and was only made available for the respondents who answered they had such roles. These questions were inspired by the questionnaire developed by Parikh et al. (2020) to evaluate a leader’s effectiveness in detecting burnout among employees, and the tools used to measure burnout among employees [ 21 ].

Part 6 was intended for radiographers and radiotherapists and only made available for those listing these as their profession. The researcher designed the questions to evaluate the aspects of quality and safety in radiology that may be affected by stress and mindfulness. To see the questionnaire in its entirety, see supplementary file 1 .

Statistical analysis

All analyses were performed using IBM SPSS version 26.0. Cronbach’s α was measured to further validate the translated parts of the questionnaire. A low value could indicate poor translation.

Demographic data, the score for individual and organizational resilience and mindfulness are described using frequencies and means. See Figs.  1 , 2 , 3 , 4 and 5 for tests of normality performed for all main variables. Bivariate correlation using Spearman’s rho was used for correlation analysis, and standard multiple regression was used to further explore the relationships between the variables.

Tests of normality. Table produced by SPSS describing the tests of normality that were performed on all main variables: individual resilience (CDRS1 to CDRS10), mindfulness (MAAS1– MAAS5), organizational resilience (BRT1 to BRT13), and quality and safety (QS1 to QS8). This includes the Kolmogorov - Smirnov and Shapiro - Wilk tests. The significance value (Sig.) under 0.05 indicates that the variables individual resilience, organizational resilience and quality and safety are not normally distributed. This does not necessarily indicate a problem with the scale used, but rather reflects the underlying nature of the construct being measured. In the case of resilience previous studies have shown this to be low among radiological personnel, which can explain why this variable is somewhat skewed. Low organizational resilience can explain why this variable is skewed, and high quality and safety can explain why this variable is skewed even if there are no problems with the scales themselves. Further inspections of normality are shown in figures 2 , 3 , 4 and 5

Histogram, boxplot, and Q-Q Plots for the variable individual resilience. The histogram (labeled a in the figure) shows that the data are not entirely normally distributed but have a peak to the left. However, the data are not severely skewed. The boxplot (labeled b in the figure) shows no outliers. The Normal Q-Q Plot (labeled c in the figure) shows a reasonably straight line, indicating that the data are not entirely normally distributed, but are not severely skewed. Last, the Detrended Normal Q-Q Plot (labeled d in the figure) show no clustering of points, indicating that the data are not severely skewed for this variable

Histogram, boxplot and Q-Q Plots for the variable organizational resilience. The histogram (labeled a in the figure) shows that the data are not entirely normally distributed but are somewhat skewed to the left. However, the data are not severely skewed. The boxplot (labeled b in the figure) shows no outliers. The Normal Q-Q Plot (labeled c in the figure) shows a reasonably straight line, indicating that the data are not entirely normally distributed, but are not severely skewed. Last, the Detrended Normal Q-Q Plot (labeled d in the figure) show no clustering of points, indicating that the data are not severely skewed for this variable

Histogram, boxplot and Q-Q Plots for the variable mindfulness. The histogram (labeled a in the figure) shows that the data are reasonably normally distributed. The boxplot (labeled b in the figure) shows no outliers. The Normal Q-Q Plot (labeled c in the figure) is showing a reasonably straight line, indicating that the data is normally distributed. Last, the Detrended Normal Q-Q Plot (labeled d in the figure) shows no clustering of points, indicating that the data are not skewed for this variable.

Histogram, boxplot and Q-Q Plots for the variable quality and safety. The histogram (labeled a in the figure) shows that the data are not entirely normally distributed but have a peak to the right. However, the data are not severely skewed. The boxplot (labeled b in the figure) shows no outliers. The Normal Q-Q Plot (labeled c in the figure) shows a reasonably straight line, indicating that the data are not entirely normally distributed, but are not severely skewed. Last, the Detrended Normal Q-Q Plot (labeled d in the figure) show no clustering of points, indicating that the data are not severely skewed for this variable.

Three models for multiple linear regression were used. In the first model, individual resilience was used as the dependent variable (is individual resilience affected by organizational resilience and mindfulness?). In the second model, organizational resilience was used (is organizational resilience affected by individual resilience and mindfulness), and in the third, quality and safety were used as the dependent variable (is quality and safety affected by both types of resilience and mindfulness?)

The model building supports the use of these models. Even if mindfulness might be a confounding factor with individual resilience (see Fig.  6 and limitations for the discussion of its effect), there are no obvious interacting variables (see Fig.  7 ), and bivariate correlation shows some relationship between most of the variables (see Fig.  8 ), in addition to the literature indicating that these variables have some effect on each other.

Tests for confounding factors in the models. To check for confounding factors the models were built by adding in one independent variable at a time. In model 1 (labeled a in the figure), where individual resilience is the dependent variable and mindfulness and organizational resilience are the independent variables, mindfulness might be a confounding variable. This is indicated by a change in the β-value (and standardized β-value) that is rather large. However, the large CI makes this change less worrisome. In model 2 (labeled b in the figure), where organizational resilience is the dependent variable and individual resilience and mindfulness are the independent variables a similar challenge occurred. This can indicate that the confounding might be between mindfulness and individual resilience. However, the CI is still large enough that the change in value in mindfulness is not worrisome. In the third and last model (labeled c in the figure), mindfulness still might be a confounding variable with individual resilience based on the change in its β-value when individual resilience is introduced which is not seen when organizational resilience is introduced to the model. The change in beta-value is the largest in this model, and the smaller CI makes this change more worrisome than in the other two models. The change in β-values and large CI can also, in part, be explained by the correlation between these factors and the relationship between them that has been established in previous studies. Since the evidence for confounding is not that strong and the indication of confounding is between two factors with a known correlation the choice was made to perform the statistical analysis as planned.

Tests for interacting variables. To check for interaction between variables the Z-scores for the variables were used, as well as moderator-variables. The Z-scores are a variable standardized to have a standard deviation of 1 and a mean of 0. The moderation-variable is the product of the independent variables in the planned regression model, which is then added to the regression model. To confirm if a variable has a moderation effect on the relationship between an independent variable and a dependent variable, the nature of this relationship must change once the moderator variable changes. In this case there does not seem to be any interacting factors, since the moderator variable is not statistically significant in either model 1 (labeled a in the figure), model 2 (labeled b in the figure) or model 3 (labeled c in the figure). This is further supported by the fact that the R Squared or adjusted R squared did not significantly change between this model and the model run with the actual variables, indicating that the relationship between the variables has not changed

Bivariate correlation using Spearman’s Rho. The correlation analysis revealed that there are statistically significant relationships between mindfulness and individual resilience (ρ = 0.27, n=62, p= 0.03), between mindfulness and organizational resilience (ρ = 0.28, n=62, p= 0.02), and between individual and organizational resilience (ρ = 0.35, n=62, p= 0.004). There are no variables that are significantly correlated with quality and safety, however. Even if it is not statistically significant, there seems to be a small, negative relationship between quality and safety and individual resilience (ρ = -0.16, n=62, p=0.21). This could indicate that there is a relationship between these variables that could be worth exploring even if their relationship is not statistically significant in this test.

The significance level was set at P  < 0.05 for all tests performed.

The Cronbach’s α scores ranged from 0.72 to 0.89, indicating internal consistency in all parts of the questionnaire. Thirty-one radiologists, 8 registrars, 24 radiographers and 5 radiotherapists completed the questionnaire (total = 68). Most respondents worked in a public setting (88%), and 67% worked in moderate to large departments. Eleven respondents (16%) had a personnel management role. Of those 11 respondents, 12.9% considered burnout a significant challenge among their employees. Approximately 1.7% of the respondents considered themselves to be very effective at detecting burnout, and 81% reported using a tool to detect employee burnout. The tools used were personal development interviews (55%), questionnaires (33%) and work environment surveys (11%).

The CD-RISC-10 total score was 30.40 ± 4.92, BRT 13 was 63.21 ± 13.63, and MAAS was 4.29 ± 0.88. The highest scores were for those working in the private sector. The total score for quality and safety was 17.79 ± 3.31. The public sector scored slightly lower than the private sector (17.83 vs. 18.20), and departments with the fewest labs (> 5) had the lowest score (16.00 ± 0.44), indicating higher quality.

The relationship between individual resilience, organizational resilience, mindfulness and quality and safety was investigated using bivariate correlation (Spearman’s rho is reported). This was chosen when preliminary analysis indicated some violations of normality (see Figs.  1 , 2 , 3 , 4 and 5 ). There was a small, positive correlation between mindfulness and individual resilience (ρ = 0.27, n  = 62, p  = 0.03), and between mindfulness and organizational resilience (ρ = 0.28, n  = 62, p  = 0.02). There was also a moderate, positive correlation between individual and organizational resilience (ρ = 0.35, n  = 62, p  = 0.004). See Fig.  8 for more information obtained from the bivariate correlation.

Standard multiple regression was performed to further explore the relationship between these variables, as described in the statistical analysis. The models revealed no strong violations of normality, linearity, or multicollinearity (Figs.  9 , 10 and 11 ), and residual analysis showed model fit (Fig.  12 ). Model 1 showed that 13.8% of the variance in individual resilience could be explained by organizational resilience and mindfulness (adjusted R squared 0.138, intercept = 18.79, F = 6.37, p  = 0.003, VIF = 1.08), with organizational resilience providing the largest unique contribution (β = 0.31, p  = 0.01) (see Fig.  9 for more information).

Summary of model 1. There do not appear to be any problems with multicollinearity in this model (tolerance <0.10, VIF-values >10 in the table labeled a in the figure, only one dimension with a variance proportion <0.90 in the table labeled b in the figure, and small correlation between the independent variables, the Pearson Correlation being 0.27, as seen in the table labeled c in the figure). There do not seem to be any outliers in the model, and the reasonably straight line in the Normal P-P Plot (labeled d in the figure) indicates normality of the data. The Adjusted R Square of the model is 0.138 (13.8% of the variance in individual resilience can be explained by the independent variables), which is statistically significant (F=6.38, p= 0.003). Organizational resilience contributed the largest, and statistically significant, unique contribution to the equation (Beta=0.31, p=0.01, as seen in the table labeled a in the figure)

Summary of model 2There could be a small challenge with multicollinearity in this model. Tolerance <0.10, and VIF-values >10 in the table labeled a in the figure, does not indicate any problems, but there are two dimensions with a variance proportion <0.90 in the table labeled b in the figure, which can indicate some problems with multicollinearity. However, the correlation between the independent variables is low enough (Pearson Correlation =0.27) that it is not worrisome. There do not seem to be any outliers in the model, and the reasonably straight line in the Normal P-P Plot (labeled d in the figure) indicates normality of the data. The models Adjusted R Square is 0.139 (13.9% of the variance in organizational resilience can be explained by the independent variables), which is statistically significant (F=6.39, p= 0.003). Individual resilience contributed the largest, and statistically significant, unique contribution to the equation (Standardized β=0.31, p=0.01, as seen in the table labeled a in the figure)

Summary of model 3There do not appear to be any problems with multicollinearity in this model (tolerance <0.10, VIF-values >10 in the table labeled a in the figure, no dimension with a variance proportion <0.90 in the table labeled b in the figure, and small correlation between the independent variables, the Pearson Correlation ranging from -0.12 to 0.07, as seen in the table labeled c in the figure). There do not seem to be any outliers in the model, and the reasonably straight line in the Normal P-P Plot (labeled d in the figure) indicates normality of the data. The models Adjusted R Square is -0.018 indicating that the independent variables do not have enough predictive value. The model is not statistically significant (F=0.64, p= 0.59).

Residual analysis for model fit. Based on the residual analysis all three models have a reasonably good fit. All residuals are somewhere between -3 and 3 in all models (model 1 is labeled a in the figure, model 2 is labeled b, and model 3 is labeled c in the figure), indicating a reasonably good fit. In model 3 (labeled c), all residuals are somewhere between -2 and 2, indicating that this model might have the best fit out of the three. The residuals are also reasonably normally distributed for models 1 and 3 (labeled a and c), further supporting that the models have a good fit. For model 2 (labeled b in the figure) the residuals seem to be somewhat skewed to the left; however, they are not skewed enough that they indicate a problem with the fit of the model.

Similar results were seen for model 2, where 13.9% of the variance in organizational resilience could be explained by individual resilience and mindfulness (adjusted R squared 0.139, intercept = 24.49, F = 6.39, p  = 0.003, VIF = 1.08). Individual resilience provided the largest unique contribution (β = 0.31, p  = 0.01) (see Fig.  10 ) Model 3 showed no statistically significant findings (adjusted R squared 0.03, intercept = 19.62, F = 0.63, p  = 0.59, VIF = 1.12) (see Fig.  11 ).

Only a minority of respondents (12.9%) considered burnout a significant challenge among their employees, and a majority (81%) reported having a tool in place for detecting burnout. This contradicts a previous study indicating that most leaders in the radiological field consider burnout a significant challenge among their employees, with only a minority having tools available to detect burnout [ 21 ]. This difference in results could be explained by differences in what is considered a tool for detecting burnout. In this questionnaire, the respondents consider development interviews, questionnaires, and work environment surveys as tools for detecting burnout, whereas respondents in previous studies might utilize these tools, but not consider them tools for detecting burnout.

The total CD-RISC-10 and BRT 13 scores indicate relatively low individual and organizational resilience among the respondents, which is consistent with previous studies [ 22 , 23 ]. This has been attributed to stress, frustration, lack of stress buffers, increased complexity of tasks, less resources, time constraints and worrying about the effect of diagnostic error on patient care [ 22 , 23 ]. At the same time, these studies demonstrated a high degree of optimism, indicating confidence in respondents’ ability to overcome the difficulties at hand [ 22 , 23 ].

Based on the correlation analysis there is a small, but positive relationship between the two types of resilience. This relationship is further validated through the standard multiple regression. The similar effects of individual and organizational resilience contradict a previous study showing that organizational resilience enables the resilient behavior of employees, and the capability to cope and learn at the individual level [ 24 ].

The correlation analysis further supports the claim that these are closely linked, and that it is important to take both in consideration when applying interventions to improve occupational health among healthcare workers. The need for not only individual, but also systematic, change has been demonstrated in previous studies [ 3 , 4 , 5 , 8 ].

Although this study shows indications of relatively high mindfulness, the results regarding quality and safety demonstrate that small mistakes that can be made under stress and time constraints are still somewhat frequent. This contradicts previous studies indicating that higher mindfulness and resilience increase the quality and safety of care [ 3 , 13 ]. The discrepancy may be attributable to variations in how different studies measure the quality of care. It is also possible that different studies measured mindfulness with different tools.

In conclusion: both individual and organizational resilience are somewhat low in Norwegian radiological departments, and mindfulness is somewhat high. There is a positive relationship between both types of resilience and mindfulness; however, resilience affects each other more than mindfulness. Quality and safety do not seem to be affected by either resilience or mindfulness.

Limitations

Variables such as gender, age, and seniority (which were not included in this study) could have an effect that is not demonstrated in this study and could account for some of the differences between this and previous studies.

Another limitation of this study is the small sample size, which did not reach the suggested number of participants needed. Small sample sizes can have a negative effect on linear regression analysis, mainly affecting the validity of the results, and to some extent, the transferability of the results to other contexts.

However, both the correlation and the linear regression showed the same relationship between individual and organizational resilience, indicating that the results regrading that correlation are valid. The findings are also still transferable for quality improvement projects and future research.

Last, there were some indications of multicollinearity in model 2 (dependent variable = organizational resilience), and mindfulness might be a confounding factor with individual resilience. However, there were no strong indications for this, so the analysis was performed as planned. Due to the indications of multicollinearity and confounding being very weak, any effects of this were also expected to be minimal.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Abbreviations

Benchmark Resilience Tool

Connor-Davidson Resilience Scale

Mindfulness Attention Awareness Scale

Norwegian Social Science Data Services

Statistical Package for Social Sciences

Ayyala RS, Ahmed FS, Ruzal-Shapiro C, Taylor GA. Prevalence of burnout among pediatric radiologists. J Am Coll Radiol. 2019;16(4):518–22.

Article   PubMed   Google Scholar  

Bundy JJ, Hage AN, Srinivasa RN, Gemmete JJ, Lee E, Gross JS, et al. Burnout among interventional radiologists. J Vasc Interv Radiol. 2020;31(4):607–13. e1.

Spieler B, Baum N. Burnout: a mindful framework for the radiologist. Curr Probl Diagn Radiol. 2022;51(2):155–61.

Kalantarova S, Mickinac N, Santhosh S, Malik S, Surovitsky M, Madsen L, et al. Preventing physician burnout in breast imaging: scope of the Problem and Keys to Success. Curr Probl Diagn Radiol. 2021;50(5):734–7.

Ganeshan D, Wei W, Yang W. Burnout in chairs of academic radiology departments in the United States. Acad Radiol. 2019;26(10):1378–84.

Ferguson C, Low G, Shiau G. Burnout in Canadian radiology residency: a national assessment of prevalence and underlying contributory factors. Can Assoc Radiol J. 2020;71(1):40–7.

Huang HL, Chen RC, Teo I, Chaudhry I, Heng AL, Zhuang KD, et al. A survey of anxiety and burnout in the radiology workforce of a tertiary hospital during the COVID-19 pandemic. J Med Imaging Radiat Oncol. 2021;65(2):139–45.

Article   CAS   PubMed   PubMed Central   Google Scholar  

Harolds JA, Parikh JR, Bluth EI, Dutton SC, Recht MP. Burnout of radiologists: frequency, risk factors, and remedies: a report of the ACR Commission on Human resources. J Am Coll Radiol. 2016;13(4):411–6.

Giess CS, Ip IK, Cochon LR, Gupte A, Dudley JC, Boland GW, et al. Predictors of self-reported burnout among radiology faculty at a large academic medical center. J Am Coll Radiol. 2020;17(12):1684–91.

Giess CS, Ip IK, Gupte A, Dudley JC, Healey MJ, Boland GW, et al. Self-reported burnout: comparison of radiologists to nonradiologist peers at a large academic medical center. Acad Radiol. 2022;29(2):277–83.

Shields M, James D, McCormack L, Warren-Forward H. Burnout in the disciplines of medical radiation science: a systematic review. J Med Imaging Radiation Sci. 2021;52(2):295–304.

Article   Google Scholar  

Fennessy FM, Mandell JC, Boland GW, Seltzer SE, Giess CS. Strategies to increase resilience, team building, and productivity among radiologists during the COVID-19 era. J Am Coll Radiol. 2021;18(5):675–8.

Melo JACd, Gelbcke FL, Amadigi FR, Huhn A, Silva Cd, Ribeiro G. Psychological exhaustion of radiological nursing workers in nuclear medicine services. Revista Brasileira De Enfermagem. 2021;73.

Serrat O. On resilient organizations. 2013.

Liang F, Cao L. Linking employee resilience with Organizational Resilience: the roles of coping mechanism and managerial resilience. Psychol Res Behav Manage. 2021;14:1063–75.

Southwick FS, Martini BL, Charney DS, Southwick SM. Leadership and Resilience. In: Marques J, Dhiman S, editors. Leadership Today: practices for Personal and Professional Performance. Cham: Springer International Publishing; 2017. pp. 315–33.

Chapter   Google Scholar  

Patriarca R, Di Gravio G, Costantino F, Falegnami A, Bilotta F. An Analytic Framework to assess Organizational Resilience. Saf Health Work. 2018;9(3):265–76.

Campbell-Sills L, Stein MB. Psychometric analysis and refinement of the connor–davidson resilience scale (CD‐RISC): validation of a 10‐item measure of resilience. J Trauma Stress: Official Publication Int Soc Trauma Stress Stud. 2007;20(6):1019–28.

Brown KW, Ryan RM. The benefits of being present: mindfulness and its role in psychological well-being. J Personal Soc Psychol. 2003;84(4):822.

Whitman R, Kachali Z, Roger H, Vargo D, Seville J. Short-form version of the Benchmark Resilience Tool (BRT-53). Measuring Bus Excellence. 2013;17(3):3–14.

Parikh JR, Bender CE. How Radiology leaders can address Burnout. J Am Coll Radiol. 2021;18(5):679–84.

Huang L, Wang Y, Liu J, Ye P, Cheng B, Xu H, et al. Factors associated with resilience among medical staff in radiology departments during the outbreak of 2019 novel coronavirus disease (COVID-19): a cross-sectional study. Med Sci Monitor: Int Med J Experimental Clin Res. 2020;26:e925669–1.

Article   CAS   Google Scholar  

Sood A, Sharma V, Schroeder DR, Gorman B. Stress management and resiliency training (SMART) program among Department of Radiology faculty: a pilot randomized clinical trial. Explore. 2014;10(6):358–63.

Gröschke D, Hofmann E, Müller ND, Wolf J. Individual and organizational resilience—insights from healthcare providers in Germany during the COVID-19 pandemic. Front Psychol. 2022;13.

Download references

Acknowledgements

Not applicable.

Funding information is not applicable.

Author information

Authors and affiliations.

Department of Health Sciences in Gjøvik, Norwegian University of Science and Technology in Gjøvik (NTNU), Teknologiveien 22, 2815, Gjøvik, Norway

Ann Mari Gransjøen

SHARE-Centre for Resilience in Healthcare, Faculty of Health Sciences, University of Stavanger, Kjell Arholmsgate 41, 4036, Stavanger, Norway

You can also search for this author in PubMed   Google Scholar

Contributions

AMG was responsible for the conception and design of the study and this paper, collection and analysis of data, and the writing of the manuscript.

Corresponding author

Correspondence to Ann Mari Gransjøen .

Ethics declarations

Ethics approval and consent to participate.

The study was submitted to, and approved by, the Norwegian Social Science Data Services (NSD) (Ref. 616162, 28 March 2022). Ethical approval was given based on how data were going to be archived and processed in regard to privacy / data protection. NSD does not have an ethics committee / IRB but is a national data protection agency. However, NSD did confirm that no further ethical approval from a Norwegian Regional Ethical Committee (REK) was necessary based on the type of study being conducted and the data being collected. All participation was voluntary and consented. Participants were informed about the study during recruitment, as well as when the link to the questionnaire was opened. Here more detailed information regarding what the study entailed for participants was provided. Consent was given when participants chose to proceed and answer the survey. Informed consent was obtained from all subjects participating in the study. The questionnaire and all other methods used in this project were carried out in accordance with relevant guidelines and regulations, as provided by the Norwegian Social Science Data Services (NSD) (Ref. 616162, 28 March 2022).

Consent for publication

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary material 2, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Gransjøen, A.M. Resilience and mindfulness among radiological personnel in Norway, their relationship and their impact on quality and safety– a questionnaire study. BMC Res Notes 17 , 96 (2024). https://doi.org/10.1186/s13104-024-06748-1

Download citation

Received : 15 August 2023

Accepted : 18 March 2024

Published : 01 April 2024

DOI : https://doi.org/10.1186/s13104-024-06748-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Individual resilience
• Organizational resilience
• Mindfulness
• Quality and safety

BMC Research Notes

ISSN: 1756-0500

• Submission enquiries: [email protected]
• General enquiries: [email protected]