ABSTRACT
Scaling is a procedure in which we combine a number of relatively narrow indicators into a single, summary measure that we take to represent the broader, underlying concept of which each is a part. The advantages inherent in the ability to develop a single number to represent a complex concept provide strong incentives to employ scaling techniques. The chapter describes four different approaches to the development of meaningful scales and describes their strengths and weaknesses with respect to validity and reliability. The four different scalings are likert scaling, guttman scaling, and thurstone scaling and semantic differential scales. The scales used most frequently in political science research are Likert scales and semantic differential scales. Guttman and Thurstone scaling techniques are less frequently employed. There are several other scaling techniques that one might employ in survey research. Discussion of these three, however, should suggest the types of options available and the criteria one must consider when developing measures of broad concepts.
