Today’s children are spending less of their free time outdoors in the neighborhood, with negative consequences for health. Neighborhood design has attracted attention for both its contribution to the problem and its potential as a solution. However, relatively few studies have examined the causal relationship between neighborhood design and children’s outdoor play. This study uses data from a 2003 survey of a random sample of households from eight Northern California neighborhoods. Using a quasi-longitudinal design, we apply ordered probit models to examine the effect of neighborhood characteristics on children’s outdoor play while controlling for neighborhood preferences. The results provide support for a causal relationship between neighborhood design and outdoor play and point to cul-de-sacs, larger front yards, lower crime, and increased interaction among neighbors as key characteristics that influence outdoor play by increasing parents’ perceptions of safety.