Multisite Qualitative Policy Research: Optimizing Description and Generalizability

Abstract

The classical qualitative educational research design is the case study. Studies of school life (Cusick, 1973; Wolcott, 1973), of the larger social forces affecting schooling (Ogbu, 1974), and of efforts to promote planned educational change (Smith & Keith, 1971) have used qualitative data in describing a single social setting. Typically, such studies emphasize in-depth description but provide a weak basis for generalization to other settings. The last decade, however, has seen the emergence of a new form of qualitative research, one intended to strengthen its ability to generalize while preserving indepth description. These multisite qualitative studies address the same research question in a number of settings using similar data collection and analysis procedures in each setting. They consciously seek to permit cross-site comparison without necessarily sacrificing within-site understanding. Although having some roots in academic social sciences (e.g., see Clark, 1970; Whiting, 1963; Whiting & Whiting, 1975), multisite qualitative research arose primarily in response to pressures from the federal government in the 1970s for studies that could overcome some of the weaknesses of large quantitative evaluations without being limited by the particularism of the single-site case study. Like many hybrids, it is today quite robust. However, these multisite qualitative studies were typically expensive endeavors and were done for specific policy purposes, which the current federal administration seems neither to value nor to feel it can afford. There are two important reasons for reflecting on the historical development and potential utility of multisite qualitative policy research at this time. Although it is unwelcomed by most social scientists, the current hiatus in commissioning policy research at the federal level provides researchers and policymakers with an opportunity to consider these issues in some detail. Further, the field of policy research has matured to the point where such considerations can be very fruitful. In recent years qualitative researchers have moved beyond the need to defend the legitimacy of their craft in the policy arena (Rist, 1977; Smith, 1978; Stake, 1978). Moreover, quantitative researchers are beginning to acknowledge a role for qualitative research in policy and evaluation studies (Cronbach, 1982; Hoaglin, Light, McPeek, Mosteller, & Stoto, 1982) and to consider the proper balance of qualitative and quantitative techniques (Cook & Reichardt, 1979; Smith & Louis, 1982). In addition, practitioners of multisite qualitative policy research now exhibit sufficient confidence in their craftsmanship to begin a process of public self-criticism with an eye to improving their methods (Firestone & Herriott, 1982; Miles, 1979; Smith & Louis, 1982; Yin, 1981). Efforts to examine multisite