Hierarchical linear modeling in park, recreation, and tourism research

Myriad research contexts in parks, recreation, and tourism are characterized by the existence of effects "nested" within other effects, but only very rarely are these effects acknowledged and incorporated into designs. Failure to account for these effects not only prevents researchers from assessing effects of nested variables, but it also creates a violation of the assumption of independence of observations that is fundamental to most such commonly used sampling distributions as t and F. Hierarchical linear modeling (HLM) is a statistical technique that provides a solution to this problem. HLM allows researchers to account for nested effects in studies that use unbalanced designs (unequal sample sizes per group), studies that use repeated measures, or other designs that create linear dependency among observations. In this paper, we review the nested effects problem and illustrate applications of HLM using a set of experience sampling data and a set of evaluation data in which intact groups are nested within a treatment variable.