Measuring Social Stereotypes with the Photo Projective Method

Article excerpt

This study aimed to measure social stereotypes with the Photo Projective Method (PPM) developed by Noda (1988). PPM is a new technique based on a projective method to capture perceived environments by photographs. Ten university students were provided with cameras and requested to take three pictures of "something representative of our university" and three pictures of "something not representative of our university." The results showed that stereotypes were measured on a microlevel and macrolevel by PPM. Also PPM allowed participants greater freedom in response production, therefore eliciting far more elaborate responses than language-based methodologies. The possibilities and advantages of PPM for measuring social identity, social representation, and other areas of social cognition are discussed.

Keywords: Photo Projective Method, social stereotypes, shared image.

Research on stereotypes has been an important topic in social psychology for many years. The concept of stereotype was formulated by Lippmann (1922), when he described the stereotype as the picture in our head. Originally, the term stereotype had a neutral meaning without negative or positive connotations. However, in many cases negative connotations are given to it, because stereotypes are associated with prejudice.

Stereotypes and prejudice have been measured by using language responses. However, on traditional racism scales, subjects could control their responses intentionally, responding as if they did not have a prejudice. To minimize this problem, the Modern Racism Scale (McConahay, 1983) included items that are not related to prejudice, to minimize the intentional hiding of prejudice. Also, the Ambivalent Sexism Inventory (Glick & Fiske, 1996) for measuring gender stereotypes used a similar approach. Compared with traditional scales, these scales more accurately measure prejudice even if subjects are motivated to hide it. However, because subjects can still intentionally control their explicit language responses, paper-and-pencil types of measuring methods have not completely solved the problems of intentional dissimulation.

In order to solve this problem, other researchers have tried to measure implicit stereotypes by using an automatic activation of a stereotype. For example, Gaertner and McLaughlin (1983) measured stereotypes by an implicit method based upon the quickness of the response to words. They measured the response times for a participant to perceive an association of words for a positive adjective and "black", and for a negative adjective and "black", computing the difference between the response times as an index of the participant's stereotype.

In recent years, the Implicit Association Test (IAT: Greenwald, McGhee, & Schwartz, 1998; Shiomura, Murakami, & Kobayashi, 2003) has been widely used as a measure for assessing social attitudes and stereotypes. In this test, two target concepts (e.g., flower vs. insect) appear in a 2-choice task on a computer display, and the evaluation of an attribute (e.g., good vs. bad) is displayed on the PC screen. The reaction time difference between highly associated categories (e.g., flower and good) and less closely associated categories (e.g., insect and good) implicitly measures the differential association of the two concepts with the attribute. This technique allows for the measurement of a stereotype without intentional bias because it is difficult to control the response intentionally.

But Bosson, Swann and Pennebaker (2000) pointed out two problems. The first problem was that the correlation between the explicit measurement and implicit measurement was low. The second problem was that the reliability measured by the test-retest method was also low compared with traditional measures - whereas the low correlation between them had merit, because that allowed prediction of different aspects of social behaviors, controlled verbal responses and spontaneous nonverbal behaviors (Cunningham, Preacher, & Banaji, 2001; Dovidio, Kawakami, & Gaertner, 2002; McConnell & Leibold, 2001). …


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.