Measurement issues have loomed large in the history of stereotype research, as in many other areas of psychology. In a scientific context, one cannot study something one cannot measure, and new measurement devices allow us to ask new questions and encourage new ways of looking at familiar problems. There is inevitably an intimate relationship between definition and measurement, and between both and theory. For example, those who believe that stereotypes are inaccurate will build this into their measurement. Those who believe that stereotypes are collective will try to discover the extent to which trait ascriptions are shared by a number of people. Therefore, measurement is hardly theoretically neutral and packs conceptual baggage. This chapter discusses the major strategies for measuring stereotypes and stereotyping.
The easiest way to assess the content of stereotypes is simply to ask people what traits or features they associate with a given group, and such "free-response" methods make a certain amount of theoretical sense. If stereotypes consist of features associated with categories, then we can make use of a time-honored method of studying associations; strong associations will be given early and without much thought. So if a person readily mentions "violent" when asked to describe an African American, or "lazy" when asked what he or she thinks about professors, we may fairly assume that these traits are strongly associated with their groups. Free-response methodologies have been extensively used to study stereotypes and prejudice (e.g., Allen, 1996; Deaux, Winton, Crowley, & Lewis, 1985; Devine & Baker, 1991; Eagly & Mladinic, 1989; Esses, Haddock, & Zanna, 1994; Esses & Maio, 2002; Haddock & Zanna, 1998b; Jackson, Lewandowski, Ingram, & Hodge, 1997; Monteith & Spicer, 2000; Niemann, Jennings, Rozelle, Baxter, & Sullivan, 1994). 34