Perceptions and the role of group exercise among New York City adults, 2010–2011: An examination of interpersonal factors and leisure-time physical activity Melanie J. Firestone a, ⁎, Stella S. Yi a,b , Katherine F. Bartley a , Donna L. Eisenhower a a NYC Department of Health and Mental Hygiene, USA b Department of Population Health, New York University School of Medicine, USA abstract article info Available online 10 January 2015 Keywords: Physical activity GPAQ Descriptive norms Group exercise LTPA Objective. To examine associations of descriptive norms (i.e., behaviors of social group members) and exercis- ing ‘with a partner’ or ‘as a part of a group’ on weekly leisure-time physical activity. Methods. T-tests and adjusted multivariable linear models were used to test the associations between de- scriptive norms and exercising with a partner or as a part of a group with self-reported leisure-time physical ac- tivity using the cross-sectional, population-based New York City Physical Activity and Transit (PAT) Survey 2010–2011 (n = 3806). Results. Overall, 70.6% of adult New Yorkers reported having physically active friends. Having active friends was associated with increased leisure-time physical activity; however, the effect varied by sex. Compared to those who did not have active friends, males with active friends reported two times more activity (56 min/week) and women reported two and a half times more activity (35 min/week) (both p-values b 0.001). Physically active males and fe- males who usually engaged in leisure-time activities as a part of a group reported 1.4 times more activity than those who exercised alone (both p-values b 0.03). Conclusions. Descriptive norms and group exercise were associated with leisure-time physical activity among adults. Based on these associations, encouraging group exercise may be an effective strategy for increasing leisure-time physical activity among certain subgroups. © 2015 Elsevier Inc. All rights reserved. Introduction Physically active people have lower rates of chronic diseases and experience higher levels of general well-being (Physical Activity Guidelines Advisory Committee, 2008; Warburton et al., 2006); yet, fewer than half (48%) of adults in the United States are sufficiently ac- tive at levels that would confer health benefits (Centers for Disease Control and Prevention, 2012). In a review of 24 qualitative studies, Allender et al.(2006) found that while most people recognize the health benefits associated with physical activity, this was not their main moti- vation for participation; factors such as weight management, social in- teraction and support, and enjoyment were more common reasons for engaging in physical activity. The importance of social influences on health-promoting behaviors is becoming increasingly recognized (McNeill et al., 2006; Sallis et al., 2008). Ecological models that include multiple levels of influence, such as the social and physical environ- ments and local policies are particularly well suited for studying physical activity behaviors (Sallis et al., 2006). As such, a clear under- standing of the interpersonal environment associated with physical ac- tivity may support the development of comprehensive interventions that lead to effective behavior change. Social networks provide an opportunity to exert social influence and offer companionship, and social support (Heaney and Israel, 2008). So- cial norms — the standards against which the appropriateness of behav- iors are evaluated — have been described as the least visible yet most powerful influencers of social control (Bettenhausen and Murnighan, 1985). However, norms have shown mixed results for predicting phys- ical activity behaviors (Sassen et al., 2010; Trost et al., 2002; Okun et al., 2002; Armitage and Conner, 2001), likely due to inconsistent terminol- ogy (Okun et al., 2002). Descriptive norms, which represent the percep- tion of what behaviors members of social groups engage in, are thought to motivate behavior by providing evidence for effective and adaptive action (Cialdini et al., 1990). There is evidence that descriptive norms, such as observing significant others engage in leisure-time activities, positively influence physical activity behaviors (Ball et al., 2010; Rivis and Sheeran, 2003; Okun et al., 2002; Priebe and Spink, 2011; Emmons et al., 2007). Having physically active friends (i.e., ‘friend de- scriptive norms’), has been shown to be more influential than having physically active family members (‘family descriptive norms’) in Preventive Medicine 72 (2015) 50–55 ⁎ Corresponding author at: Department of Health and Mental Hygiene, 125 Worth St., 6th Floor, CN 58A, New York, NY 10013, USA. Fax: +1646 632 6065. E-mail address: mfirestone@health.nyc.gov (M.J. Firestone). http://dx.doi.org/10.1016/j.ypmed.2015.01.001 0091-7435/© 2015 Elsevier Inc. All rights reserved. Contents lists available at ScienceDirect Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed