Political science and public administration literature suggest that consolidated metropolitan government is preferable to fragmented local government. The literature argues that consolidation provides economies of scale that reduces administrative overhead and improves efficiency, while allowing public administrators to equitably distribute regional resources. The literature of special districts suggests that decentralization of metropolitan government is preferable to consolidation. The literature argues that smaller local governments provide public administrators greater autonomy, allowing special districts to be more responsive to local needs. This study tests the competing hypotheses of political science and public administration, and special districts. To test the competing hypotheses, this study compares the service levels and efficiency levels of Illinois park districts and Illinois parks and recreation departments. In this study, three hypotheses are identified. The first, originating from the literature of parks and recreation administration, states that park districts provide higher service levels than municipal parks and recreation departments. The second, also originating from the literature of parks and recreation administration, states that park districts provide higher levels of autonomy than parks and recreation departments, which in turn provide hgher levels of services. The third, originating from the literature of political science and public administration, states that parks and recreation departments provide higher levels of efficiency in delivering services than park districts. Illiois was selected for study because it has the largest park district system in the United States, with 305 park districts, and because Illinois park districts receive 38 percent of the national Gold Medal Awards for Excellence in parks and recreation administration. Illinois was also selected because it has one of the largest municipal parks and recreation department systems in the United States, with 125 agencies, permitting quantitative analysis of service level and efficiency level outcomes between the two structures of government. Service level, autonomy, demographic, and administrator profile data were collected by surveying an equal number of randomly selected park district administrators and randomly selected parks and recreation department administrators. Data was analyzed using comparative means testing, bivariate and multivariate linear regression to test the relationships between service levels, administrative autonomy and efficiency levels as identified in the literature. In addition, these quantitative techniques were used to test competing hypotheses identified in other studies, suggesting that staffing and professional training levels are explanations for service and budgetary levels. The aggregate results of this study show that structure of government is an important determinant of total recreation program levels, with park districts providing two more programs on average than parks and recreation departments. This study also demonstrates that agency service area population, rather than structure of government, is the primary determinant of the number of facilities that an agency operates. This study demonstrates that a strong relationship exists between structure of government and administrative autonomy. But since structure of government is not a determinant of per capita total service levels, no meaningful relationship between autonomy an per capita service levels appears to exist. Testing the relationship between structure of government and efficiency, this study finds that structure of government is not an important determinant of agency efficiency. Considering other potential explanations of agency service levels and efficiency, including the hypothesis that professional training, job satisfaction, length of service on the job or administrative salary are determinants of service levels or efficiency levels, this study finds that only longevity of administrators in their current posisition has an impact on service levels or efficiency. The implications of these quantitative findings are that neither park districts nor parks and recreation departments are preferable in terms of providing higher levels of services or efficiency. Neither parks and recreation theory that park districts are preferable to parks and recreation departments, nor political science and public administration theory that parks and recreation departments are preferable to park districts are affirmed. [ABSTRACT FROM AUTHOR]