Twitter Behaviour is Related to Reflective Thinking

Post by D. Chloe Chung

What's the science?

Many of us are using at least one type of social media platform to help us connect with others and discuss important social issues. Despite these benefits, social media can also be misused to easily spread false information and fuel political polarization. Given the power of social media, several studies have looked at how personalities or demographic characteristics are related to the different ways people use social media. Adding to this line of research, this week in Nature Communications, Mosleh and colleagues examined how people’s cognitive style is associated with their behaviour on Twitter.

How did they do it?

The authors recruited approximately 2,000 people who regularly use Twitter, mostly from the United Kingdom and the United States. These participants took the Cognitive Reflection Test (CRT), which measures people’s tendency to follow their instinct and choose wrong answers. Specifically, this test measures one’s ability to perform self-reflective thinking to find correct answers while suppressing a “gut feeling” related to an incorrect response. A higher CRT score indicates that the participant is better at cognitive reflection. Next, the authors collected several pieces of information related to the Twitter activity of the participants, such as how many accounts they follow and what type of content they have recently tweeted. They also gathered demographic information about the participants including their education, political ideology, religion, and income. Based on these data, the authors created a co-follower network to examine what type of accounts are followed by participants who share similar CRT scores.

What did they find?

First, the authors focused on examining the content the participants consume on Twitter. The authors observed that Twitter users with higher CRT scores (i.e. more reflective thinking) showed a tendency to follow fewer Twitter accounts. From the co-follower network analysis, the authors found a distinct division in the types of Twitter accounts followed by people with higher and lower CRT scores, suggesting that critical thinking is reflected in an individual’s account-following behaviour on Twitter. Interestingly, there was a group of accounts followed by people with lower CRT scores (i.e. less reflective thinking), supporting the notion of “cognitive echo chambers,” in which people tend to interact with those who share similar ideologies. The authors analyzed the type of content participants tended to tweet and found that the degree of reflective thinking was associated with the quality of information shared. Specifically, people who think more reflectively were more likely to share higher-quality news from reliable sources, while people who engage less in reflective thinking shared political misinformation and scams more often. Upon analyzing individual words in participants’ tweets, the authors observed that more reflective people tended to use words related to morality, insight, and inhibition, which may indicate that these people are more likely to inhibit their instincts by engaging in analytical thinking.

What’s the impact?

This study demonstrates how different cognitive styles can be reflected in our behaviour on Twitter. In particular, this work shows what can drive the dissemination of misinformation on social media. In contrast to the “intuitionist” perspective that emphasizes the importance of intuition in everyday behaviours, findings from this study suggest that reflective or analytic thinking plays a crucial role in our day-to-day judgment on social media. It will be interesting to investigate whether these findings can be also applied to other social media platforms.

Mosleh et al. Cognitive reflection correlates with behavior on Twitter. Nature Communications (2021). Access the original scientific publication here.

Changes in Personality Traits Predict Early Career Outcomes

Post by Shireen Parimoo

What's the science?

Are conscientious people more likely to be satisfied with their career? Several studies show that career outcomes are associated with personality traits. For example, traits like emotional stability and conscientiousness are related to higher income and occupational attainment. Although these traits are largely stable aspects of our personality, they also tend to vary across different life stages, suggesting that our personality undergoes changes over the course of our lifetime. It is particularly important to understand the nature of personality growth in early adulthood because it is a critical period in career development. This week in Psychological Science, Hoff and colleagues investigated how personality changes from late adolescence to early adulthood are related to various career outcomes.

How did they do it?

The authors followed two large, nationally representative samples of adolescents in Iceland (Sample 1 N = 485, Sample 2 N = 1290), beginning when participants were students (15-17 years old) to when they were young adults (27-29 years old). During this period, they measured participants’ personality traits at three (Sample 1) and five time points (Sample 2) using the NEO Five-Factor Inventory, a personality test that provides scores on the traits of openness, conscientiousness, extraversion, agreeableness, and emotional stability. They used linear growth-curve analysis to assess participants’ baseline scores in adolescence on each personality trait and the rate at which their scores changed over time.

After approximately 12 years, the authors recorded subjective career outcomes such as job and career satisfaction, as well as objective measures like monthly income, highest degree obtained, and occupational prestige. Occupational prestige was based on the most recent job held by the participants. To rate occupations on status and prestige, they used the online tool O*NET that classified jobs into different titles (e.g., surgeon) and assigned each title a rating on various work-value dimensions like Achievement and Recognition. To investigate the relationship between personality changes and career outcomes, the authors specified a separate path model (a statistical approach) for each outcome, including personality traits as predictors and controlling for the effect of gender.

What did they find?

Agreeableness, conscientiousness, and openness increased with age, whereas emotional stability did not change, and extraversion scores decreased. Higher conscientiousness, emotional stability, and agreeableness in adolescence were associated with higher degree attainment in the future. Similarly, baseline levels of emotional stability and conscientiousness (Sample 1) as well as openness and extraversion (Sample 2) were related to greater future occupational prestige. In fact, emotional stability and conscientiousness in adolescence were positively related to nearly all career outcomes, highlighting their importance in future career development. Moreover, changes in personality traits were also predictive of career outcomes. For example, higher income and career satisfaction were associated with increases in emotional stability from adolescence to young adulthood. Similarly, although extraversion decreased in general, increases in extraversion were related to higher career and job satisfaction. Overall, these results show that personality traits in adolescence predict objective measures of career success whereas certain changes in personality traits over time are linked to subjective career outcomes.

What's the impact?

These results are exciting because they highlight both enduring and mutable aspects of personality development in the transition from adolescence to adulthood. In addition, this is the first longitudinal study to demonstrate how personality traits in adolescence and young adulthood predict objective and subjective career outcomes, respectively. Future research can further examine the direction of the relationship between personality changes and career success.

Hoff et al. Personality Changes Predict Early Career Outcomes: Discovery and Replication in 12-Year Longitudinal Studies. Psychological Science (2020). Access the original scientific publication here.

Political Views Bias Information Processing in the Brain

Post by Deborah Joye

What's the science?

With an important election fast approaching in the United States, it is clearer than ever that political polarization exists in our society. A powerful contributor to political polarization is motivated reasoning; when we make decisions based on emotional bias rather than evidence or facts. Motivated reasoning is made especially obvious when people with opposing beliefs are presented with identical information and their biased interpretation leads them to become even more entrenched in their original positions.

But how might identical information trigger such different outcomes, and can we see these different responses in the brain? It’s possible that our previously held beliefs change what we pay attention to when we absorb new information. For example, someone might view a political ad and only pay attention to information that supports their beliefs. It’s also possible that previously held beliefs change how people interpret incoming information. For example, one person may interpret actions as threatening, risky, or offensive where someone with different beliefs might not. This week in PNAS, Leong and colleagues demonstrate that when viewing political content,  conservative- and liberal-leaning people displayed differential activation of the dorsomedial prefrontal cortex which was intensified by moral-emotional and risk-related language.

How did they do it?

The authors scanned thirty-eight participants with functional magnetic resonance imaging (fMRI) while they viewed videos on immigration policies. Before scanning, participants indicated their support for each of the policies, and the authors calculated an overall “Immigration Attitude” to identify liberal- and conservative-leaning participants. Importantly, the authors controlled for age, sex, income, and education. The authors first asked if neural responses to the videos were different across participants. They calculated an intersubject correlation; a measure of how similar brain activity is across participants. This measure also allowed the authors to investigate which regions of the brain exhibited differential activity..

The authors next asked whether differential brain activity was specifically associated with certain content within the videos. The authors categorized words in the videos that were related only to emotions (e.g. fear, sad, etc.), related only to morals (e.g. ethics, etc.), or related to both morals and emotions (e.g. compassion, harm, etc.). The authors identified a total of 50 categories, including risk, religion, power, family, social, male, and female, among others. Finally, the authors asked whether brain activity that was more like either political group might predict the participant’s preference for supporting the position held by that group. To assess this, the authors averaged brain activity across both groups, creating a map of what the signal might look like for an “average” liberal or conservative. After each video, the authors asked each participant whether they’d be more or less likely to support that policy. The authors then assessed whether a particular participant’s brain activity during that video was more like that of the average conservative or the average liberal.

What did they find?

The authors found that overall neural responses to the political videos were shared across participants. Specifically, regions of the brain related to sensory processing displayed similar activity patterns regardless of the participant’s political leanings, suggesting that political attitudes do not alter sensory processing. However, activity patterns in the dorsomedial prefrontal cortex diverged between liberal and conservative participants. Previous research has demonstrated that this brain region is known to be active when participants are manipulated to have different interpretations of an ambiguous story. 

The authors also found that the difference in activity patterns in the dorsomedial prefrontal cortex was associated with the use of moral-emotional and risk-related language, suggesting that those categories of language are most likely to drive divergent neural responses between liberals and conservatives. The authors also found that connectivity between the ventral striatum (associated with emotional processing) and the dorsomedial prefrontal cortex was more similar for participants with similar political views. Lastly, the authors found that for each video, whether a participant’s neural activity was more like the average conservative or liberal predicted their attitude change. For example, a participant with brain activity patterns more like the average conservative was more likely to display attitude change further in the conservative direction.

What's the impact?

This study found that political attitudes can drive divergent neural processing of political information. One striking implication of these findings has to do with the “echo chamber” of social media, where online videos with moral-emotional content tend to spread faster within social circles sharing similar political beliefs. These findings suggest people with already formed political biases will respond to such political content with further attitude polarization. This study also demonstrates that content that features moral-emotional and risk-related language is most likely to be processed in this biased way. These findings have important implications for our political discourse moving forward and have implications for the type of language we use in a political context, as well as the impact this may have on an increasingly polarized society.

Leong et al., Conservative and liberal attitudes drive polarized neural responses to political content, PNAS (2020). Access the original scientific publication here.