A sedentary lifestyle has long been linked to poor health, and a growing body of evidence suggests it may also affect personality. Previous research found associations between a lack of exercise and declines in character traits such as conscientiousness, measured four to 10 years after initial surveys. Now the largest analysis of its kind to date has used longer follow-up periods to confirm these links and show they persist up to nearly two decades.

A team led by psychologist Yannick Stephan of the University of Montpellier in France reached this conclusion after combining data from two large, survey-based studies. The Wisconsin Longitudinal Study (WLS) followed people who had graduated from that state's high schools in 1957, as well as some of their siblings. The Midlife in the United States (MIDUS) study recruited people from across the country. Participants in both had completed personality questionnaires when first recruited in the 1990s and answered questions about their exercise habits and health.

Nearly 20 years later a total of about 9,000 people took the same surveys again. Stephan and his team found that subjects who reported being less active had greater reductions on average in conscientiousness, openness, agreeableness and extroversion—four of the so-called Big Five personality traits—even after accounting for differences in baseline personality and health. No link was found with the fifth trait, neuroticism. The changes in traits were small, but the link with exercise was relatively strong. Physical activity predicted personality change better than disease burden did, for example. The findings were published in April in the Journal of Research in Personality.

Numerous mechanisms may be involved—from physiological factors such as stress response to changes in physical ability that can affect how much people socialize. “Personality is, in part, what [behaviors] we repeatedly do, and changes in habits can consolidate into changes in personality,” says epidemiologist Markus Jokela of the University of Helsinki, who was not involved in the new study.

Correlations do not prove causation, however. Additional factors, such as genetics or earlier life events, might be affecting both exercise levels and personality. The findings also need to be replicated in samples from different cultures and in studies using objective measures of an active lifestyle.

Nevertheless, the new analysis underscores the idea that personality is malleable throughout life. It also tallies with studies suggesting personality is linked to health. “These findings further emphasize the need for physical activity promotion in midlife and older age,” Stephan says.