There is a positive relationship between intelligence and survival; intelligence is a protective factor for reaching upper-middle age, thereafter survival depends less on intelligence and more on other factors

Intelligence and life expectancy in late adulthood: A meta-analysis- Macarena Sánchez-Izquierdo et al. Intelligence, Volume 98, May–June 2023, 101738. Mar 2023. https://doi.org/10.1016/j.intell.2023.101738

Highlights

• There is a positive relationship between intelligence and survival.

• The most robust moderator is years of follow-up.

• Intelligence is a protective factor for reaching upper-middle age, thereafter survival depends less on intelligence and more on other factors.

Abstract: In an aging society, it is crucial to understand why some people live long and others do not. There has been a proliferation of studies in recent years that highlight the importance of psycho-behavioural factors in the ways of aging, one of those psychological components is intelligence. In this meta-analysis, the association between intelligence and life expectancy in late adulthood is analysed through the Hazard Ratio (HR). Our objectives are: (i) to update Calvin's meta-analysis, especially the estimate of the association between survival and intelligence; and (ii) to evaluate the role of some moderators, especially the age of the participants, to explore intelligence–mortality throughout adulthood and old age. The results show a positive relationship between intelligence and survival (HR•: 0.79; 95% CI: 0.81–0.76). This association is significantly moderated by the years of follow-up, the effect size being smaller the more years elapse between the intelligence assessment and the recording of the outcome. Intelligence is a protective factor to reach middle-high age, but from then on survival depends less and less on intelligence and more on other factors.

Keywords: IntelligenceMortalitymeta-analysisSystematic review

4. Discussion

>20 longitudinal studies from several countries (Australia, Denmark, Luxembourg, Sweden, United Kingdom, and USA) have demonstrated the link between higher intelligence and longer life. This gave rise to the field of cognitive epidemiology, which focuses on understanding the relationship between cognitive functioning and health.

This study aimed to update Calvin's meta-analysis, confirm the quantification of this association and also analyse whether the intelligence–mortality association varies across adulthood and old age. We found evidence that having intelligence of at least 1-SD above the mean seems to reduce the mortality rate, although our rate was a little lower (21.6%) than that of Calvin's meta-analysis (24%).

Another objective of the study was to analyse the influence of several factors as possible moderators, specifically bias, age, and sex. Our results showed that recent studies tend to find a weaker association between intelligence and mortality than older studies. Along this line, Calvin and colleagues have shown a trend for larger cohorts accumulating in more recent years (Calvin et al., 2011).

The average life expectancy of women exceeds that of men, however, sex does not moderate the association between intelligence and mortality, being the same for men and women, as previously reported in twin longitudinal studies (Arden et al., 2016) and Calvin et al. (2011).

The question “What causes the relationship between intelligence and longevity/ mortality?” remains unsolved and crucial. Factors such as childhood environment, family income, schooling, and healthy/unhealthy lifestyle habits (diet, exercise, tobacco use, alcohol, illnesses), have been studied (Deary, Weiss, & Batty, 2010; Whalley & Deary, 2001). As Deary et al. (2021) suggested, there seems to be a reciprocal dynamic association between intelligence and health throughout life, and although there are several constructs associated with health/ illness and death (e.g., parental social class, intelligence in youth, more education, higher health literacy, healthy behaviors, and more affluent social class) shared genetic differences are likely to account for only a small proportion of these associations.

In this study we wanted to re-evaluate the influence of Socio-economic status as a predictor of mortality; our results showed that childhood SES did not moderate the potential of intelligence for predicting mortality. Although several studies (Batty et al., 2007; Hemmingsson, Melin, Allebeck, & Lundberg, 2009) suggested that intelligence had effects on the risk of mortality independent from those of early socio-economic influences, other studies suggested that SES was not a confounder of the intelligence–mortality association (Calvin et al., 2011, Calvin et al., 2017). Furthermore, a study of over 900 Scottish participants (Hart et al., 2003) found that statistically controlling for economic class and a measure of “deprivation” reflecting unemployment, overcrowding, and other adverse living conditions accounted for only about 30% of the IQ-mortality correlation.

Along this line, Gottfredson (Gottfredson, 2004) argued that underlying IQ differences explained social inequalities in health and that these were not necessarily mediated via adult/person's-own SES. This idea was tested by Batty, Der, Macintyre, and Deary (2006) who found that IQ does not completely explain socioeconomic inequalities in health, however, it might contribute to them through a variety of processes.

Another line of research suggested that genes may contribute to the link between IQ and mortality. Arden and colleagues (Arden et al., 2016) analysed three twin studies (from the U.S., Denmark, and Sweden) and found a small positive phenotypic correlation between intelligence and lifespan, furthermore, in the combined sample, the genetic contribution to covariance was 95%; in the US study, 84%; in the Swedish study, 86%, and in the Danish study, 85%. As the authors highlighted, any genetic factors that contribute to intelligence and mortality may operate indirectly via good health choices or higher income which leads to better healthcare. Deary, Harris, and Hill (2019); Deary et al. (2021)) reviewed the genetics through genome-wide association studies (GWASs), Genome-wide complex trait analysis (GREML), and LD regression studies, which allowed them to estimate genetic correlations between phenotypes (intelligence and health).

The second question of this study: “Does the relationship between intelligence and mortality change in the older adults?” Yes, that relationship changes when the most long-lived studies were compared with the youngest studies. As several studies have suggested (Arden et al., 2016; Hart et al., 2005), the causes of the association between intelligence and lifespan may vary between ages. Childhood IQ has been related to mortality in Scottish populations: Hart et al., (2005) showed that childhood IQ was significantly related to deaths occurring up to age 65, but not to deaths occurring after age 65, whereas the Aberdeen study found that people with a lower IQ were less likely to be alive at age 76 (Whalley & Deary, 2001).

Analysing whether the relationship between intelligence and mortality changes in the older adults, our results showed a small but significant positive slope for FUY, which reflected that the association was slightly smaller as more years elapsed between time 1 (intelligence assessment) and time 2 (check for survival). This means that the relationship between intelligence and survival is dampened. Our findings confirm results from the Midspan studies (Hart et al., 2005). As suggested by several studies, one possible reason might be that higher IQ might be associated with better healthcare and engaging in healthier behaviors (Deary et al., 2019; Hart et al., 2005; Wraw, Der, Gale, & Deary, 2018), which is associated to a lower mortality risk (Gottfredson, 2004; Gottfredson & Deary, 2004). IQ might also predispose to conditions of adult life (Marmot & Kivimäki, 2009), quitting smoking in later life (Batty et al., 2007; Daly & Egan, 2017), and entry into safer environments (Whalley & Deary, 2001), which promote staying healthier and living longer.

One way of discovering why intelligence and mortality are related and why this association seems to be smaller at higher ages might be to review the specific causes of death to which intelligence relates from childhood and adulthood. Along this line, several studies have shown its association with most of the major causes of death. The main literature has reported inverse patterns of the association between childhood intelligence and respiratory disease (Batty, Deary, & Zaninotto, 2016; Calvin et al., 2017), coronary heart disease (e.g., Calvin et al., 2017; Christensen, Mortensen, Christensen, & Osler, 2016; Hart et al., 2004; Lawlor, Ronalds, Clark, Davey Smith, & Leon, 2005;), stroke (Calvin et al., 2017), total cardiovascular disease (Batty et al., 2016; Calvin et al., 2017; Christensen et al., 2016; Hart et al., 2003, Hart et al., 2004; Hemmingsson, Melin, Allebeck, & Lundberg, 2006; Leon, Lawlor, Clark, Batty, & Macintyre, 2009), digestive disease (Calvin et al., 2017), cancer (Batty et al., 2009, Batty et al., 2016; Leon et al., 2009), specifically with smoking-related diseases (Calvin et al., 2017), dementia (Calvin et al., 2017; Russ et al., 2013), and suicide (Hemmingsson et al., 2006,

Deary et al. (2021) presented consistent results showing intelligence associated with several causes of death (cardiovascular disease, coronary heart disease, stroke, respiratory disease, diabetes, digestive disease, dementia, non-smoking-related cancers, accidents and suicide), illnesses (hypertension, metabolic syndrome, diabetes, schizophrenia, and major depression), health biomarkers (e.g. systolic and diastolic blood pressure, heart rate, triglycerides and cholesterol, body mass index), and health behaviors (smoking and physical inactivity). As the authors highlight, intelligence's long-term association with health is mediated via adult social factors and health behaviors.

5. Limitations

The present meta-analysis includes large published studies representing in total >47,000 average sample size. However, it includes 22 studies: 16 studies were already included in the previous meta-analysis (Calvin et al., 2011) and six new articles.

Although all studies were adjusted for multiple potential moderators, there are likely to remain different factors, such as SES in adulthood, cause of death in the intelligence–mortality association, etc., that could substantially affect the results.

For those reasons, the combined models are not strictly comparable, since other moderators are frequently added to childhood SES and it is not possible to disentangle their effects in the meta-analysis. Although we are aware of this potential weakness, we have preferred to perform and report this analysis. If we had found an effect, it would have been difficult to interpret, but we did not any effect of this moderator. As might be expected, the inclusion of moderators leads to a significant increase in heterogeneity, which we may interpret to mean that some of the moderators increase the effect and others reduce it, but we cannot identify the role each of them plays at the meta-analytic level.

Also, future research should explore mediating effects on a pathway from premorbid intelligence to the risk of mortality, taking into account common genetic effects (e.g. with GWAS) and the role of socioeconomic status, health literacy, and adult environments and behaviors.

It should also be important to include other countries and cultures in the studies.

Liberalizing prostitution leads to a significant decrease in rape rates, while prohibiting it leads to a significant increase (much greater than liberalization's effect)

Do Prostitution Laws Affect Rape Rates? Evidence from Europe. Huasheng Gao and Vanya Petrova. The Journal of Law and Economics, Volume 65, Number 4, Mar 2023. https://www.journals.uchicago.edu/doi/epdf/10.1086/720583

Abstract: We identify a causal effect of the liberalization and prohibition of commercial sex on rape rates, using staggered legislative changes in European countries. Liberalizing prostitution leads to a significant decrease in rape rates, while prohibiting it leads to a significant increase. The results are stronger when rape is less severely underreported and when it is more difficult for men to obtain sex via marriage or partnership. We also provide the first evidence for the asymmetric effect of prostitution regulation on rape rates: the magnitude of prostitution prohibition is much larger than that of prostitution liberalization. Placebo tests show that prostitution laws have no impact on nonsexual crimes. Overall, our results indicate that prostitution is a substitute for sexual violence and that the recent global trend of prohibiting commercial sex (especially the Nordic model) could have the unforeseen consequence of proliferating sexual violence.

        If you expel prostitutes from society, you will unsettle everything on account of lusts. (St. Augustine, in Richards 1995, p. 118)1

Personality disorders reveal much stronger sex differences than normal personality traits, with men leaning much more towards the "dark" side

Sex in the dark: Sex differences on three measures of dark side personality. Adrian Furnham, George Horne. Acta Psychologica, Volume 234, April 2023, 103876. https://doi.org/10.1016/j.actpsy.2023.103876

Abstract: This study examined sex differences in the scores on three different measures of the personality disorders (PDs) all derived from on-line surveys. Two groups (total N = 871) completed the Coolidge Axis-II Inventory which assessed 14 PDs; two groups (total N = 732) completed the Short Dark Tetrad which assessed 4 PDs; four groups (total N = 1558) completed the Personality Inventory for DSM-5—Brief Form which assessed 5 PD dimensions. Cohen's d after ANOVAs, and binary regression analysis revealed consistent findings. In this study we calculated 63 d statistics of which 5 were d > 0.50 and 28 were d > 0.20. In two samples, each using two different instruments, men scored higher than women on Anti-Social, Narcissistic and Sadistic PD which is a consistent finding in the literature. Speculations are made about the origin of these differences. Limitations are acknowledged.

Keywords: SexPersonalityTraitsDisordersEffect sizeBright/dark side

4. Discussion

The results of this study are largely consistent with previous research in this area, and confirms they hypotheses. This paper raises a number of points. First, the consistency of the PD sex differences across samples who took the same test, and second across PDs measured between different tests. With regard to the consistency between samples there seemed “reasonable” agreement particularly with those that were most and least significantly different. In all, we had eight participant groups with an 232 < N < 506 who were recruited on-line. In no instance did analyses show opposite results with the exception of one group tested on the DSM-5 where men scored higher than women on the Negative Affectively scale in contrast to the other three groups. Thus, we have demonstrated the generalisability of results across very different measures, in eight different samples.

A major question concerns sex differences in “bright-” as opposed to “dark-side” traits. Furnham and Treglown (2021) who looked at six tests found the Cohen's d statistic showed very few (3 out of 130) differences >0.50. In a study of dark-side traits, Furnham and Grover, 2022a, Furnham and Grover, 2022b found a Cohen's d statistic showed very few (5 out of 44) differences >0.20. In this study however we calculated 63 d statistics of which 5 were d > 0.50 and 28 were d > 0.20. Thus, it appears there are more differences on dark-, as opposed to bright-side measures. This finding requires an explanation and further investigations. However, we have to acknowledge that overall, there are both relatively few and small sex differences, an observation made by many in this area.

We were also able to compare sex differences on different measures of the same trait as the SCATI and the Dark Tetrad both measured Anti-Social, Narcissistic and Sadistic PD. This was consistent between the samples and the instruments showing the following d scores: Anti-Social: 0.30, 0.28, 63, 0.69; Narcissistic: 0.26, 0.84, 0.32, 0.83 and Sadistic PD 0.41; 0.25, 0.80, 0.49. These results confirm the previous literature on Anti-Social and Narcissistic PD but highlight the role of Sadistic PD which, admittedly does not appear as a PD in any of the DSM manuals (American Psychiatric Association, 2000, American Psychiatric Association, 2015). It explains also why so many studies on powerful derailed individuals nearly always highlight men rather than women (Babiak & Hare, 2006).

An examination of the Binary Logistic Regressions showed that the Exp(B) varied mainly between 0.80 and 1.20 the lowest being 0.63 for Antagonistic for Group 1 and the highest being 1.68 for Negative Effect for Group 1. Again, depending on cut-interpretations these could be considered high or low.

However, it does appear from this data that having a PD is predominantly a “male problem” in that on all four Tetrad traits, and four of the five DSM-5 dimensions males scored significantly higher than females across all samples. The SCATI did show that where there were consistent findings across the two samples and a d > 0.10 women did score higher on Borderline, Dependent, and Schizotypal, which has been established in previous studies.

There appears to be relatively little theoretical development in the PD literature about the “causes” of the different PDs that may lead to very clear hypothesis testing. Whilst it would not be difficult to develop an evolutionary-based theory explaining why men might be higher on Anti-Social and Narcissistic PD it seems much more difficult to explain why women might score more highly on other PDs like Borderline or Schizotypal. In this sense few of the sex difference studies in PD are theoretically, rather than psychometrically, driven.

5. Conclusion

The strengths of this paper was to report sex difference in the PDs using multiple measures (three) and multiple samples (eight). The results suggest that compared to studies of sex differences in bright-side (normal) personality where sex differences are common but small, sex differences in (some) dark-side traits are consistently larger.

This study has implications for the theory, measurement and indeed treatment of the PDs. From an evolutionary psychology perspective it seems possible to explain some of these differences: for instance, why boldness, fearlessness and self-confidence maybe beneficial to males, though in excess a disadvantage. Equally it may be possible to explain some sex differences in the traditional socialisation of children into established sex roles.

As the movement of PD researchers from a categorical to a dimensional perspective progresses we should be able to inspect sex differences seeking to establish consistency of findings and following that explanations.

6. Limitations

There are frequent critiques that online survey data is often problematic with participants being perfunctory in their responses. In each study we included IQ items as well as other checks to be able to inspect the quality of the responses. In most studies we removed a small number of participants before the analysis with concerns about the quality of their data.

This study explored the data bank of a research group. Nearly all the participants were functioning working adults and not a student or clinical sample, though it is possible that a small number were present in each study. Although we had a lot of data on each participant it was not consistent between samples. Furthermore, it would have been desirable to have a lot more data on each person such as their education and general mental health.

Of the two categorical measures the SCATI has been used in a number of studies and appears to have adequate psychometric properties but is not a particularly well-known measure. The Dark Tetrad measure on the other hand is relatively new though attracting a good deal of attention (Alavi et al., 2022; Fernández-del-Río et al., 2022; Jain et al., 2022). However, the DSM-5 is now 10 years old and has been used in many studies. Studies such as this serve to describe sex differences but give no indication in their cause or consequence. Thus, they can show that differences exist but not why.

Income and emotional well-being

Income and emotional well-being: A conflict resolved. Matthew A. Killingsworth, Daniel Kahneman, and Barbara Mellers. Proc. Natl. Acad. Sci., March 1 2023, 120 (10) e2208661120. https://doi.org/10.1073/pnas.2208661120

Significance: Measures of well-being have often been found to rise with log (income). Kahneman and Deaton [Proc. Natl. Acad. Sci. U.S.A. 107, 16489–93 (2010)] reported an exception; a measure of emotional well-being (happiness) increased but then flattened somewhere between $60,000 and $90,000. In contrast, Killingsworth [Proc. Natl. Acad. Sci. U.S.A. 118, e2016976118 (2021)] observed a linear relation between happiness and log(income) in an experience-sampling study. We discovered in a joint reanalysis of the experience sampling data that the flattening pattern exists but is restricted to the least happy 20% of the population, and that complementary nonlinearities contribute to the overall linear-log relationship between happiness and income. We trace the discrepant results to the authors’ reliance on standard practices and assumptions of data analysis that should be questioned more often, although they are standard in social science.

Abstract: Do larger incomes make people happier? Two authors of the present paper have published contradictory answers. Using dichotomous questions about the preceding day, [Kahneman and Deaton, Proc. Natl. Acad. Sci. U.S.A. 107, 16489–16493 (2010)] reported a flattening pattern: happiness increased steadily with log(income) up to a threshold and then plateaued. Using experience sampling with a continuous scale, [Killingsworth, Proc. Natl. Acad. Sci. U.S.A. 118, e2016976118 (2021)] reported a linear-log pattern in which average happiness rose consistently with log(income). We engaged in an adversarial collaboration to search for a coherent interpretation of both studies. A reanalysis of Killingsworth’s experienced sampling data confirmed the flattening pattern only for the least happy people. Happiness increases steadily with log(income) among happier people, and even accelerates in the happiest group. Complementary nonlinearities contribute to the overall linear-log relationship. We then explain why Kahneman and Deaton overstated the flattening pattern and why Killingsworth failed to find it. We suggest that Kahneman and Deaton might have reached the correct conclusion if they had described their results in terms of unhappiness rather than happiness; their measures could not discriminate among degrees of happiness because of a ceiling effect. The authors of both studies failed to anticipate that increased income is associated with systematic changes in the shape of the happiness distribution. The mislabeling of the dependent variable and the incorrect assumption of homogeneity were consequences of practices that are standard in social science but should be questioned more often. We flag the benefits of adversarial collaboration.

Deceptive affection is strategically expressed under relational threat—but not towards partners with low mate value

Deceptive affection is strategically expressed under relational threat—but not towards partners with low mate value. Neil R. Caton and Sean M. Horan. Journal of Social and Personal Relationships, March 6, 2023. https://doi.org/10.1177/02654075231152909

Abstract: Individuals sometimes express affection that they do not feel. This describes deceptive affectionate messages and occurs when communicators express affectionate messages that are not consistent with their internal feelings of affection in the moment. They are commonly expressed in romantic relationships (about 3 times per week) and are argued to function as relational maintenance and retention. The present work (N = 1993) demonstrated that deceptive affectionate messages are the behavioral output of an evolved psychological system that strategically operates to maintain significant pair bonds (i.e., high mate value partners) but not non-significant pair bonds (i.e., low mate value partners). This system is uniquely and nonrandomly designed to increasingly generate deceptive affectionate messages when the individual’s highly valued partnership is perceived to be under relational threat and decreasingly deploy deceptive affectionate messages when the highly valued partnership is not under threat, but the system does not apply this relational strategy in low-valued partnerships. This supports evolutionary psychological reasoning that affectionate communication should be predicated on a cost–benefit ratio, such that deceptive affectionate messages are expressed to high value mates because the substantial costs of losing a highly valued partner outweigh the smaller risks of enacting them (e.g., discovered deception, temporary relational conflict). By establishing that deceptive affection is predicated on a cost–benefit ratio, the present work better solidifies deceptive affection, and affection exchange theory more broadly, in the human evolutionary sciences.

Rather than generating the filter bubbles that pundits keep hallucinating, googling for political information drives individuals toward sources that are different to their routinary visits

Search engine effects on news consumption: Ranking and representativeness outweigh familiarity in news selection. Roberto Ulloa and Celina Sylwia Kacperski. New Media & Society, Mar 6 2023. https://doi.org/10.1177/14614448231154926

Abstract: While individuals’ trust in search engine results is well-supported, little is known about their preferences when selecting news. We use web-tracked behavioral data across a 2-month period (280 participants) and we analyze three competing factors, two algorithmic (ranking and representativeness) and one psychological (familiarity), that could influence the selection of search results. We use news engagement as a proxy for familiarity and investigate news articles presented on Google search pages (n = 1221). We find a significant effect of algorithmic factors but not of familiarity. We find that ranking plays a lesser role for news compared to non-news, suggesting a more careful decision-making process. We confirm that Google Search drives individuals to unfamiliar sources, and find that it increases the diversity of the political audience of news sources. We tackle the challenge of measuring social science theories in contexts shaped by algorithms, demonstrating their leverage over the behaviors of individuals.

Discussion

We tested the hypothesis of news engagement (as a proxy for familiarity with news sources) as a predictor of news article selection in the Google search engine (RQ1). We did not find evidence supporting this. Instead, we found a significant effect for two factors that are decided by the search engine alone: the position in which the result is presented (ranking) and the number of times the news source appears (representativeness). While ranking has previously been demonstrated to play a strong role (Pan et al., 2007; Urman and Makhortykh, 2021), we show for the first time that its effect is weaker for news article selection compared to non-news selection (RQ3). This may well suggest a more careful decision-making process of individuals when selecting news (e.g. reading the titles and excerpts more attentively).

Research has indicated a higher representation of “mainstream” news sources in search results (Puschmann, 2019), while, at the same time, a positive effect in the diversity of news consumption (Fletcher et al., 2021; Fletcher and Nielsen, 2018). Our results align with this seemingly counter-intuitive evidence: representativeness reduced the likelihood of news article selection (RQ1). This might be an indication that once individuals have decided not to visit a result belonging to a specific news source, they also discard subsequent results from the same source, suggesting that the individual is actively avoiding such sources (Mukerjee and Yang, 2021).

In line with previous research (RQ2a), we found that Google Search increases the diversity of participants’ news consumption (Fletcher et al., 2021; Fletcher and Nielsen, 2018; Scharkow et al., 2020). It is possible that participants use Google Search when they are actively looking for novel news sources, though we also show that Google Search facilitates a discovery process by presenting a variety of news sources among the results. In addition, we show that Google Search increases the political audience diversity that news sources receive (RQ2b). Given that Google has its own news quality controls in place (Google Developers, 2021), the finding can explain recent research showcasing that political audience diversity can be used as a sign of news source reliability, and that it should be incorporated into ranking algorithms (Bhadani et al., 2022). Instead, our results suggest that it is Google Search (including its ranking) that drives this effect. More broadly, researchers should consider that online news browsing behavior is heavily shaped by online platforms, for example, we demonstrated that there are differences in the consistency of our familiarity metric depending on whether it is measured including traffic referred by Google or not. Considering the dominance of a few large search engines and their role in driving users to news, the situation may lead to a concentration of power and influence within the media landscape in which news organizations prioritize visibility on search engines (and similar platforms) by following SEO guidelines, instead of focusing on journalistic norms (Hackett, 1984; Muñoz-Torres, 2012).

Furthermore, we build on previous research investigating the phenomenon of mere-exposure (Montoya et al., 2017) and trust in news sources (Fletcher and Park, 2017), and investigated familiarity (RQ1), that is, the participants’ acquaintance with the news sources they are presented in the search results. We proposed news engagement as a proxy and measured it using a section of the browsing history that is independent from the analyzed news articles that are selected (and visited) in the search results. This allowed us to quantify the existent relationship between the individual preferences toward news sources through the number of visits of the individual to each news domain; thus, capturing three modes of news engagement: routinary visits, social media referrals, and intentional search (Möller et al., 2020).

To summarize, we find no evidence in favor of filter bubbles, nor do we find an effect of selective exposure toward news sources based on familiarity. We find that individuals place their trust in Google, and that in turn, Google steers them toward sources that are different to their routinary visits. In this process, we tackle the challenge of developing reliable measurement models when integrating social science theory with digital behavioral data (Wagner et al., 2021).

Limitations and future directions

We would like to point out several limitations. First, our engagement metric does not fully capture when individuals are familiar and trust a given source, that is, when they do not regularly consume news or mainly do it offline. In addition, the engagement for the time independent analysis was calculated using an arguably short period of time (~1 month). Future data collections should enable analyses across timespans of multiple months up to years, though we highlight that our results did not change when considering a time-dependent analysis (~2 months).

Second, our analysis is limited to Google, which we chose due to its market dominance (StatCounter, 2021). Our findings should not be generalized to other search engines due to differences in how search results are displayed, though it is likely that effects specifically related to ranking and representativeness will remain similar across different interfaces. Our findings should also not be generalized to contexts beyond the top-10 organic results of web searches, such as carousels and top stories, news aggregators (Google News), or recommendation systems such as Google Discover. The latter does not provide results based on prompts input by the user, but presents content (including news) personalized from individuals’ preferences, which may be extracted from previous search behavior. Since search choices are heavily influenced by ranking, the algorithmic representations of preferences might also reflect the ranking effects as individual traits. This raises questions about the extent to which preferences can be captured by these systems, or if they merely expand the influence of the search engine. It highlights a challenge for study of new media: how can we better understand interdependencies between the digital services, rather than studying them in isolation?

Third, characteristics of our sample should be kept in mind when interpreting our results. The sample size used in the above research is relatively small. Out of the 739 individuals that participated in the web tracking study overall, only 280 are represented in our subsample, due to a relatively low number of news visits (including the ones driven by Google Search) in web tracking data—consistent with previous literature (Scharkow et al., 2020; Wojcieszak et al., 2021), and the strict data quality constraints for including a search page for the analysis. Our sample is relatively uniform, only including German individuals with a Chrome or Firefox browser installed on their desktop computers; while this reduces noise, it also affects generalizability of the findings. Finally, many individuals refuse to participate in web tracking studies due to privacy concerns (Makhortykh et al., 2021), which might indicate that the sample is pre-selected based on factors that we don’t yet fully understand and cannot control for.

Despite these limitations, the logged web browsing behavior that we capture occurs in a real environment with minimal intervention; we argue that cognitive awareness of the presence of the web tracker is likely to only affect the very initial browsing behavior. Moreover, web tracking studies that include the website content remain rare, and our method is exceptional as it deterministically identifies referrals by tracing the tab activity of the browser and matching the presence of the URLs among the results.

Some promising directions emerge for new media research. Our results suggest that ranking is less important for news selection: research should further confirm if individuals are more selective when choosing news from the top-10 search results, which could for example be achieved by analyzing the duration of the interaction with the individual choices or by applying eye-tracking methodologies. In addition, we find that representativeness is negatively associated with news selection. A further inspection of the reasons is of interest, for example, this behavior might signal an active avoidance of specific news sources (Mukerjee and Yang, 2021). Finally, it is important to examine search choices within the broader context of the layout of web search pages, including features like video carousels and top stories, to gain a deeper understanding of how individuals make decisions for information acquisition.

The use of social media only stirs up political anger in people who strongly embrace like-minded content, while even heavy users without such blinders become less angry

Birds of a Feather Get Angrier Together: Social Media News Use and Social Media Political Homophily as Antecedents of Political Anger. Zicheng Cheng, Hugo Marcos-Marne & Homero Gil de Zúñiga. Political Behavior, Mar 6 2023. https://link.springer.com/article/10.1007/s11109-023-09864-z

Abstract: A significant body of literature within political communication revolves around the constructive political virtues and blighting social and democratic consequences of political anger. For the most part, studies have focused on identifying the primary causes and antecedents of political anger. However, within the context of social media, fewer efforts have been devoted to clarifying how and what infuriates people about politics. Does social media news use relate to increased or reduced levels of political anger? Do social media political homophilic networks explain political anger? And to what extent does political homophily influence the potential effect of social media news use on citizens’ political anger levels—moderating effect? Results drawing on a two-wave U.S. survey dataset show that the frequency of social media news use alone has no direct effect on people’s increased political anger, whereas interacting in homophilic discussion and information networks on social media positively associates with anger. Furthermore, the relationship between social media news use and political anger is contingent upon social media political homophily. Those who report high levels of social media news use and very low levels of social media political homophily end up being less angry over time. Limitations and steps for future research are discussed in the manuscript.

Discussion

Anger is an important political emotion that holds an ambivalent relationship with democracy. While it may serve the powerless to question the political order (Lyman, 2004) and trigger constructive politics before conflicts escalate (Tagar et al., 2011), political anger is also associated with biased assimilation, fueling ideological bias in the acceptance of political information that aligns with one’s opinion (Weeks, 2015). Anger also relates to reliance on pre-existing heuristics and stereotypes (Suhay & Erisen, 2018), increased incivility, hostility, and distrust (Hasell & Weeks, 2016), less willingness to compromise (Mackuen et al., 2010; Wollebæk et al., 2019), and even with violence (Claassen, 2016). Acknowledging the wide range of consequences associated with political anger, this study focused on its social media roots.

Our analysis shows the importance of considering social media political homophily to understand political anger in online environments. Higher levels of social media political homophily do not only associate with political anger, but they also moderate the oftentimes empirically elusive relationship between social media news use and political anger. While social media news use does not contribute to explaining political anger directly, individuals who rank low on social media political homophily will be less angry about politics the more they use social media to consume information about public affairs. More importantly, the direct and moderating effect of social media political homophily on political anger is consistent across all models tested in this study: cross-sectional, lagged, and autoregressive.

Plenty of research has considered whether social media use would produce more homogeneous (Adamic & Glance, 2005; Conover et al., 2011; Feller et al., 2011) or heterogeneous (Choi & Lee, 2015; Kim, 2011; Lee et al., 2014) online environments. Our findings suggest that these studies are of the highest importance to unravel the association between social media news use and political anger. Our findings also suggest that there is no unified answer to social media news. That is, it is not solely about whether people use social media for news or not, but rather other political and communicative predispositions making individuals connect more often with like-minded people and expose themselves to ideologically congruent news. Then, they will be more likely to ‘fall victim’ to the hyper-partisan news and discussion environments which are featured with blame-attribution, moralization, and identity politics framing (Barberá, 2020; Hameleers et al., 2018; Rydell et al., 2008), thus eliciting negative effects like anger. In other words, how people consume social media and curate their news feed exerts an influence on political anger.

Our results demonstrate that whether social media news use is associated with political anger is contingent upon how the specific informational and discussion affordances that social media also provide, more specifically, whether they actively and purposively curate homophilic or heterogeneous social media news and discussion networks. Cinelli et al. (2021) suggested that aggregation of homophilic users dominates the interaction dynamics on social media like Facebook and users tend to seek information that is consistent with his/her preexisting opinion and favor the interaction with like-minded peers, and this situation leads to the formation of polarized groups online. Alternatively, Dubois and Blank (2018) found that a diverse media diet, including news use on multiple media outlets, will direct social media news users toward more diverse information and perspectives, reducing the likelihood of getting into the echo chamber. Results by Guess (2021) are particularly important in this regard, as he demonstrates that most people, at least in the U.S., interact in relatively heterogeneous environments online. While our paper remains agnostic as to the extent to which online homophily is present, the main results provide support for the idea that social media political homophily, when present, matters not only in the context of creating political segregation (Conover et al., 2011), spreading misinformation (Del Vicario et al., 2016), or strengthening group identity (Yardi & boyd, 2010), but also in explaining political anger.

This study adds some nuance to the understanding of social media news users by connecting social media political homophily, social media news use, and political anger. Prior research has suggested that affective polarization is on the rise in the U.S. and that some predicting factors, such as selective exposure (Levendusky, 2013; Tsfati & Nir, 2017) and negative political news coverage (Schmuck et al., 2020), may explain that trend. Our study contributes to filling the research gap by looking into the link between social media use patterns and people’s emotional responses to politics (i.e., political anger). Drawing on our findings, promoting a more heterogeneous social media news use and discussion network can provide an alternative pathway to reducing political anger in the American public. Although political anger has been found to mobilize the public and stimulate political actions, it increases incivility and hostility (Hasell & Weeks, 2016), causes political violence (Claassen, 2016), and exacerbates partisanship and political polarization (Huber et al., 2015). Our results suggest that by altering how individuals engage with social media news people may become less angry, which may trigger a subsequent array of democratically beneficial outcomes such as political tolerance, and less political dogmatism, as well as affect changes in people’s political behavior (Rathnayake & Winter, 2017). From the policy-making perspective, reducing homophily in people’s social and informational networks also seems to be the key. Social media platforms shall address the disadvantages brought by algorithmic news personalization, and efforts should be made to provide social media users with diverse information content, encourage users to follow accounts with opposing views, and interact with peers and news sources that encompass dissimilar political beliefs.

Albeit important, the study is not immune to limitations. There are several shortcomings that must be acknowledged and that might ideally serve as an orientation for further research on the relationship between social media news use, social media political homophily, and political anger. First, while political anger is widely spread across countries, our survey data was collected in a single country, the U.S., and even though it is a panel dataset, it is based only on one year, 2019. In this sense, while the autoregressive models have shown that overall levels of political anger vary to a moderate extent in four months, it is important to see whether choosing a different time lag will make our variables of interest gain more importance (Eveland & Morey, 2011). We also encourage future studies to examine how the link between anger and political behavior may differ across racial groups (Phoenix, 2019), as this study controls for a non-granular white versus minorities dichotomy, and further racial effect nuances may be possible (Magee & Louie, 2016).

Similarly, additional works, comparative in nature, may shed light on the existence of different patterns between social media political homophily and political anger with various degrees of overall anger in the country. Macro, meso, and individual measurement instruments may prove useful here. For instance, there might be country-level moderators for the relationship identified such as the country’s economic condition (Rico et al., 2020), ethnically located injustice (Holmes, 2004), and/or sexism culture (Kay, 2019). Additionally, we measured social media news use and social media political homophily with a self-report survey, which is subject to recall bias and social desirability bias (Scharkow, 2016). As computational methods are increasingly integrated into political communication research, it would be interesting to measure social media news use and social media political homophily with behavioral tracking data. While future studies can adopt more unobstructive measures to palliate potential bias, recent research consistently shows that although self-reported and tracking data on social media news use encompass discrepancies, overall, they positively correlate (Ernala et al., 2020; Haenschen, 2020), which minimizes the impact of this limitation. Besides, it is worth noting that the behavioral tracking data is also subject to measurement error due to the variations of the operating system setting (Jones-Jang et al., 2020), which yields a caveat for using the tracking data as an objective benchmark. As suggested by Jürgens et al. (2020), future work can address the methodological challenge by developing a more advanced digital trace tracking tool, using source-and-issue-specific survey questions, employing longitudinal survey designs (which is done in our study), and combining different data sources. Overall, this study helps clarify the informational and network discussion antecedents of political anger, in a modest but much-needed empirical assessment for the field.