A Preferred Method for Obtaining Rankings: Reactions to Physical Handicaps

Abstract

their original form rather than combining them into composite indices or collapsing categories prematurely. Not only would this allow a clearer picture of just what is going on in the particular system under analysis, but it would also increase our ability to compare one piece of research with another. For example, data on occupation, education, and income are collected in relatively standard ways, but they get combined into indices of socioeconomic status in a wide variety of ways, which makes comparison between studies unnecessarily difficult. The disadvantage may lie in over-enthusiastic use of the second advantage combined with under-enthusiastic use of the first advantage. The ability to control for large numbers of variables is meaningful only if there is some theoretical rationale behind it. Hopefully, however, attention to the theoretical foundations of this technique will discourage fishing expeditions as more than temporary expediencies. Multiple regression analysis of dummy variables is a specialized tool. The explicit theoretical foundation which Coleman's work provides may be especially important in assuring proper use of this tool.