Using data from a questionnaire administered to the same population (Alberta, Canada) in two surveys conducted four years apart (1988 and 1992), this paper investigates three issues related to the replication of research on leisure constraints and the interpretation of temporal changes in the experience of constraints: (1) the magnitude of changes in the reporting of constraints over time; (2) the extent to which such changes are a function of differences in sample or population characteristics; and (3) stability or variability of relationships between aspects of constrained leisure and socio-economic sources of variation. The comparison showed very little temporal change in aspects of leisure constraints, measured with respect to the unfulfilled desire to start a new activity and responses to 15 constraints items. Moreover, the majority of significant differences were accounted for by differences in the age and income structures of the two samples. Also, socioeconomic variations in constraints remained stable over time.KEYWORDS: Leisure constraints, sum research, leisure trends, replication.INTRODUCTIONLack of replication is one of the most serious problems of leisure research (Godbey, 1989), since there is little assurance that patterns of leisure behavior identified in studies conducted at particular times and in specific places hold true elsewhere. More importantly, failure to replicate previous research may mean that regularities in patterns of association between leisure behavior and explanatory or dependent variables cannot be generalized with any reasonable degree of certainty, thus attenuating the development of leisure-related theory. Until recently, the growing sub-field of leisure constraints research has been no exception to this lack of replication. It is true that some efforts to re-test earlier findings have begun to appear, for example McGuire et al.'s (1989) re-examination of research conducted by Jackson and Dunn (1988), and papers by Searle and Brayley (1992), Hultsman (1992) and Jackson and Rucks (1993) on an issue first examined by Jackson and Dunn (1991). However, none of these studies has represented an exact replication of a previous investigation.Roughly every four years since 1979, the Alberta government has conducted a large-scale questionnaire survey among the adult population of the Province of Alberta, Canada. These surveys have collected information about the recreation and leisure behavior, perceptions, motivations and preferences of Albertans. One of the most notable aspects of the surveys has been the inclusion of a series of questions about the kinds of constraints that people may encounter in fulfilling their leisure aspirations. While this constraints information, together with that from the other sections of the questionnaires, has been collected primarily for the purposes of recreation forecasting, planning and management, it has also proven useful from an academic standpoint. Results of the surveys have been published in the leisure research journals (e.g., Jackson, 1983, 1990a; Jackson & Dunn, 1988; Searle & Jackson, 1985) and are freely available to leisure scholars and practitioners for secondary analysis (Ross, 1990).Various modifications have been made to the constraints-related components of the surveys as knowledge about constraints has increased and as the requirements of the government have evolved. From one perspective, these changes are beneficial, in the sense that the questions have been based on the most recent concepts and have provided information relevant to immediate practical and applied needs. On the other hand, incommensurate data from the various surveys have limited the potential for charting and explaining changes over time (see, for example, Jackson, 1990b; Jackson & Dunn, 1991).In this context, the key feature of the 1992 survey was its exact replication of the content, questions and wording of its 1988 counterpart. …