Gen Z and conservatives show higher misinformation susceptibility, large-scale study finds

A large-scale international study sheds light on who is more likely to struggle with identifying false information online. The research found that individuals belonging to Generation Z, those who do not identify as male, people with lower levels of formal education, and individuals identifying as politically conservative generally showed greater susceptibility to misinformation. Additionally, the study revealed complexities in self-awareness, noting that while confidence often tracked actual ability, the accuracy of this self-perception differed among various groups. The findings have been published in the journal Personality and Individual Differences.

Misinformation has been connected to offline violence, public health crises like vaccine hesitancy, and political manipulation. International bodies have identified combating misinformation as a global priority. However, understanding how to effectively tackle this issue requires a clearer picture of the factors—cognitive, social, and demographic—that make individuals more prone to believing false information.

Previous research in this area often faced limitations. Studies frequently used inconsistent or unvalidated methods to measure susceptibility, making comparisons difficult. Research predominantly focused on a few specific countries, particularly Western nations like the United States, leaving uncertainty about whether findings applied more broadly across different cultures and populations.

“We were interested in this topic because misinformation is a growing global threat that impacts everything from public health to democratic stability. Yet much of what we know is based on narrow samples concentrated in just a few countries. We wanted to better understand who is most vulnerable to misinformation—across diverse populations—using a validated measure of misinformation susceptibility,” explained study author Jin (Hyunjin) Koo, a postdoctoral research associate at the University of British Columbia.

To overcome previous shortcomings, the current study employed a two-part strategy. First, the researchers used a standardized and validated tool called the Misinformation Susceptibility Test. This test was developed over several years and specifically designed to measure a person’s ability to distinguish real news headlines from fake ones. It presents participants with a mix of 10 genuine headlines, sourced from factually accurate and politically neutral media outlets, and 10 fake headlines generated using artificial intelligence and selected through a rigorous vetting process. This standardized approach allows for more reliable measurement and comparison across different groups and studies.

Second, the researchers created a publicly accessible website where individuals anywhere in the world could take the Misinformation Susceptibility Test. The website was promoted through press releases, social media, university communications, and word-of-mouth. This recruitment method proved highly effective, attracting a massive and diverse pool of participants.

Upon visiting the site, individuals received information about the test and could choose to participate. They were given the option to receive personalized feedback on their performance without sharing their data, or they could explicitly agree (opt-in) to allow their anonymized responses to be used for scientific research. Participants could choose between a 20-item or a 16-item version of the test; this study focused on those who completed the 20-item version.

From June 2023 to July 2024, over 114,000 submissions were received from people who agreed to share their data. For this specific analysis, the researchers included data from 66,242 participants residing in 24 different countries, selecting only countries with at least 250 submissions to ensure stable statistical comparisons.

The sample was diverse, with the largest numbers coming from the United States, the United Kingdom, Canada, and Australia. Participants provided demographic information including their gender identity (Male, Female, Non-binary/Third), birth year (used to determine generation: Generation Z, Millennials, Generation X, Baby Boomers), highest level of education attained (from High School or less up to Graduate/professional degree), and political orientation (on a 7-point scale from Extremely liberal to Extremely conservative).

The participants also answered a single question about how good they believed their own ability was to distinguish real from fake news, using a 5-point scale from very poor to very good. The researchers then used statistical techniques designed to analyze large datasets with participants nested within different countries to examine the associations between these demographic factors, self-perceived ability, and actual performance on the Misinformation Susceptibility Test.

After accounting for other factors, individuals who identified as female or non-binary/third tended to score lower on the test compared to those identifying as male, indicating greater susceptibility to misinformation.

A clear generational pattern also emerged: Generation Z participants (born 1997–2012) scored the lowest on average. Millennials, Generation X, and Baby Boomers all performed significantly better than Generation Z.

“Despite being digital natives, members of Generation Z were among the most vulnerable to misinformation,” Koo told PsyPost. “This challenges the common assumption that younger generations are more media savvy simply because they’ve grown up online.”

Education level showed a strong association with test performance; individuals with higher levels of formal education consistently scored higher than those with a high school education or less.

Political orientation was also linked to test scores. Compared to participants identifying as extremely liberal, scores generally decreased as individuals identified further towards the conservative end of the spectrum. Those identifying as conservative or extremely conservative showed the lowest average scores. These patterns generally held true even when looking at U.S. participants separately or excluding them entirely.

The study also examined the relationship between how well people thought they could spot misinformation and how well they actually performed. Generally, participants who rated their ability higher did tend to score better on the test. Those who rated their ability as “very good” scored highest, while those rating it “very poor” scored lowest.

However, the researchers explored whether this connection between perceived ability and actual performance was consistent across different groups. They found interesting variations. Women’s self-assessments were more closely aligned with their actual test scores compared to men’s. Across generations, Generation Z participants, despite having the lowest average scores, showed the most accurate self-perception – the link between their confidence and their actual ability was stronger than for older generations.

Political orientation also played a role; the connection between self-perceived ability and actual performance weakened significantly for those identifying as extremely conservative, suggesting their confidence was less indicative of their true skill level compared to other political groups. Interestingly, participants with the highest levels of education (Bachelor’s degree or graduate degree) were slightly less accurate in estimating their abilities compared to those with high school or lower education.

Finally, the researchers looked at broad country-level differences. Performance varied across the 24 countries included. Participants from nations like Belgium, Canada, Finland, the Netherlands, and Sweden performed similarly to those in the United States (which served as the reference point). Participants from the remaining countries included in the analysis generally scored lower than the U.S. group, with notable differences observed for some nations.

“Misinformation susceptibility is shaped by a range of social, psychological, and demographic factors,” Koo said. “Since everyone is vulnerable in different ways, strengthening critical thinking and digital literacy is essential for us all.”

“No matter who you are, no matter what you think you know, none of us is immune to misinformation,” added Friedrich Götz, an assistant professor of psychology at the University of British Columbia and the study’s senior author. “People should realize that all of us are exposed to misinformation on a regular basis, and all of us are likely to fall for it at some point.”

The researchers acknowledge some limitations. Although the sample was geographically diverse, the test was only available in English. This means the sample likely skewed towards participants from English-speaking countries or those with high English proficiency. The findings might not fully generalize to populations or cultures where English is not the primary language. Future research could involve translating and validating the Misinformation Susceptibility Test in other languages to better understand cultural nuances in misinformation vulnerability.

While some of the observed differences between groups might seem numerically small (e.g., the average score difference between generations), the researchers suggest that even small differences in susceptibility can have meaningful real-world consequences, given the sheer volume of information encountered online and the potential impact of misinformation on critical decisions, from personal health choices to voting.

“My long-term goal is to build a global understanding of misinformation susceptibility that can inform efforts to help individuals and organizations strengthen their resistance to fake news,” Koo explained.

The study, “Profiling misinformation susceptibility,” was authored by Yara Kyrychenko, Hyunjin J. Koo, Rakoen Maertens, Jon Roozenbeek, Sander van der Linden, and Friedrich M. Götz.