The place attachment construct has been used by leisure researchers and practitioners to refine our understanding of certain leisure behaviors for over two decades. Despite the construct's importance to natural resource-based leisure, little empirical work has appeared in the leisure literature examining the construct's antecedent processes; that is, the processes that lead to recreationists'; attachment to settings. This study examined one antecedent, activity involvement, using covariance structure analysis. The authors examined a model suggesting that place attachment (i.e., place identity and place dependence) would be predicted by activity involvement (i.e., attraction, centrality, and self expression) among four groups of hikers along the Appalachian Trail (i.e., day hikers, overnight hikers, section hikers and through hikers). Using LISREL's multigroup procedure, results indicated that the place identity dimension of place attachment was best predicted by the self expression and attraction dimensions of activity involvement, whereas the only predictor of place dependence was self expression. These relations were consistent for all hikers. Type of use (i.e., day hiker, overnight hiker, section hiker and through hiker), however, was shown only to moderate the correlation between activity involvement dimensions, attraction and self expression, and place attachment's place identity and place dependence. As hikers' activity involvement and attachment grew, the correlations between these constructs declined. This result indicates that these scales' discriminant validity improves as respondents more readily identify with the attitude object used in the item wording.