|| Checking for direct PDF access through Ovid
To examine the contribution of neighborhood, household, and individual factors to socioeconomic inequalities in sports participation in a multilevel design.Data were obtained by a large-scale postal survey among a stratified sample of the adult population (age 25-75 yr) of Eindhoven (the fifth-largest city of the Netherlands) and surrounding areas, residing in 213 neighborhoods (N = 4785; response rate 64.4%). Multilevel logistic regression analyses were performed with sports participation as a binary outcome (no vs yes); that is, respondents not doing any moderate- or high-intensity sports at least once a week were classified as nonparticipants.Unfavorable perceived neighborhood factors (e.g., feeling unsafe, small social network), household factors (material and social deprivation), and individual physical activity cognitions (e.g., negative outcome expectancies, low self-efficacy) were significantly associated with doing no sports and were reported more frequently among lower socioeconomic groups. Taking these factors into account reduced the odds ratios of doing no sports among the lowest educational group by 57%, from 3.99 (95% CI, 2.99-5.31) to 2.29 (95% CI, 1.70-3.07), and among the lowest income group by 67%, from 3.02 (95% CI, 2.36-3.86) to 1.66 (95% CI, 1.22-2.27).A combination of neighborhood, household, and individual factors can explain socioeconomic inequalities in sports participation to a large extent. Interventions and policies should focus on all three groups of factors simultaneously to yield a maximal reduction of socioeconomic inequalities in sports participation.