SOME SPATIAL ASPECTS OF PUBLIC URBAN RECREATION IN COLUMBIA, SOUTH CAROLINA Lisle S. Mitchell and Paul E. Lovingood, ]r.a Recreational opportunity is valued highly in a modern urban society and is no longer considered a luxurv. In particular, community subsidized recreational facilities and services are important in an increasingly complex urban environment. Nevertheless, social science research on the subject of urban recreation has been remarkably meager. (1) The contributions of geographers to the spatial aspects of this research is on the increase, but they have yet to develop and test a comprehensive set of principles and models from which a viable theory of urban recreation can be constructed. The authors address this problem with the objective of empirically testing several widely endorsed but apparently contradictory generalizations about the spatial variation of urban recreation consumption. As a case study, this research aspires to neither definitive conclusions nor profound theoretical implications. Rather, it examines the explanatory utility of these spatial constructs within the context of a single urbanized area with the ultimate goal to develop them into hypotheses worthy of rigorous testing in similar contexts. The generalizations to be investigated are: 1) the density of public parks is negatively associated with central cities and with those areas of the city which are on the lower end of the socioeconomic scale; and 2) suburbs tend to be void of public recreational opportunities. The first generalization is derived from statements such as "the residents of the central city generally do not have easy access to outdoor recreation," and "low income city residents, particularly in the slum areas, recent in-migrants, minority groups and elderly people have little or no access to outdoor recreation. . . ." (2) The second generalization follows from ". . . many of the fastest growing suburban areas have thus far made extremely inadequate provision for parks and with the available land taken up in large-scale developments, there are neither vacant lots nor vacant sites upon which parks may be constructed later." (3) These generalizations are in obvious conflict because they allude to the absence of public parks and recreational opportunity in different parts of the city. Both of these statements cannot be true. The procedures outlined below are designed to clarify this contradiction and to develop the generalizations into acceptable hypotheses. *Dr. Mitchell is associate professor and Dr. Lovingood is professor of geography at the University of South Carolina. This paper was accepted for publication in February 1975. 94 Southeastern Geographer PROCEDURE. Three steps were followed in the investigation of the previously stated generalizations. In the first step, the urbanized area of Columbia, South Carolina, was divided into high and low park density regions. Median park density was employed to differentiate the regions. Median density, purely by chance, was exactly one park per square mile. There were, therefore, 30 census tracts with less than one park per square mile and 30 with more than one park per square mile. The high park density region consists of the central cities of the Columbia urbanized area (Columbia, West Columbia, and Cayce) minus Columbia's Central Business District and its functionally related industrial areas plus one residential census tract. The low park density region includes census tracts on the periphery of the study area and those in or near Columbia's CBD (Figures 1 and 2). The second step was an analysis of selected population, family, housing, and income variables for the two regions. The variables were subject to the following analytical techniques: map analysis, simple correlation analysis, linkage analysis, and factor analysis. Finally, the generalizations were examined in terms of the results of the analytical techniques, and the findings were considered in a comprehensive research framework. UIlANlZED AIEA COLUMIIA. SOUTH CAIOUNA MT] ------------- COUNTY tOUNOMT ------------- CITT LIMITI¦¦INSTITUTION«! AItAI \ Figure 1 Vol. XV, No. 2 95 Figure 2 GENERALIZATION ONE: ANALYSES AND RESULTS. The findings of this study cast doubt on the veracity of the first generalization. Map analysis reveals that parks are concentrated beyond Columbia's CBD and functionally related census tracts and in the residential neighborhoods of the three central cities (Figures 1 and 2). A spatial gradient of park locations shows that 75 percent of the parks are located within four miles of...