Liberal and conservative college students do not just vote differently. Smartphone data from a large Texas campus show they also move, work and relax in slightly different ways, even in activities with no obvious political content. But their classmates got the details mostly wrong, and saw a bigger divide than the data showed.
The study, “Lifestyle polarization on a college campus: Do liberals and conservatives behave differently in everyday life?,” is by Sanaz Talaifar of Imperial College London, Diana Jordan of Duke University, Samuel D. Gosling of The University of Texas at Austin and the University of Melbourne, and Gabriella M. Harari of Stanford University. It was published online in the Journal of Personality and Social Psychology on May 5, 2025, in the journal’s July 2025 issue, after first appearing as a preprint on January 30, 2024.
Two weeks of phone sensors and check-ins
Study 1 drew on undergraduates in a large introductory psychology class at the University of Texas at Austin in fall 2016 and spring 2017, with up to 1,229 students in the analyses. Students rated their politics on a seven-point scale from extremely liberal to extremely conservative.
For two weeks, an app recorded GPS location, ambient sound from the microphone, calls, texts, phone unlocks and physical activity, and asked students four times a day what they were doing, where and with whom. That produced 61 everyday behaviors in four domains: social life, movement, work and leisure.
At the daily level, 18 of the 61 behaviors (30%) were linked to political orientation, eight to conservatism and ten to liberalism. Counting different times of day and week, 29 behaviors showed a link at least once. The correlations were small, averaging about 0.09, but most held after controlling for gender, ethnicity and social class, and there were as many differences by politics as by gender.
Frat houses for conservatives, home and internet for liberals
Leisure showed the biggest split, with 7 of 15 behaviors linked to politics. Conservatives spent more time in noisy places and at fraternities or sororities, while liberals spent more time at home, did more chores and errands in the mornings and evenings, and browsed the internet and social media more.
Conservatives were also more physically active and studied or read more, although the studying gap disappeared once their higher conscientiousness was taken into account. Liberals moved around more in the mornings, visiting more places and spending more time in transit, and they worked and spent time with coworkers more. The authors sum the pattern up as conservatives making the most of where they are, and liberals exploring further afield. Social behavior, such as calls and time with friends, showed no reliable differences.
Classmates saw a bigger divide than existed
Study 2 asked 156 students from the same university in 2023 to judge whether liberals or conservatives did each of the 61 behaviors more. Their guesses were unrelated overall to the real pattern, and they classified 76% of behaviors wrongly. Of the misjudged behaviors, 72% were cases where students imagined a difference that did not exist. They did better on leisure, where the real differences were largest.
The authors caution that the sample comes from one university and that the specific behaviors may not generalize, which is why they focus on the overall pattern. They also did not measure contact between liberals and conservatives directly.
Their point is that a campus of students the same age, in the same city and the same class is a hard place to find lifestyle differences, yet they showed up anyway. In the authors’ words, political identity has penetrated everyday life, “but not to the degree that people assume.”
Study Details:
- Title: Lifestyle polarization on a college campus: Do liberals and conservatives behave differently in everyday life?
- Authors: Sanaz Talaifar, Diana Jordan, Samuel D. Gosling, Gabriella M. Harari
- Journal: Journal of Personality and Social Psychology, Volume 129, Issue 1, pages 152-180
- Publication Date: May 5, 2025
- DOI: 10.1037/pspp0000545