Two possible mediators of attitude change, internalization and identification, were investigated in a laboratory setting. Internalization was assumed to underlie change when respondents are attracted to others who hold the same opinion and can argue in favor of their attitude position; identification was assumed to mediate change when respondents are attracted to similar others but cannot necessarily support their position. A causal analysis based on these assumptions confirmed the independent

more »

... xistence of internalization and identification as mediators of attitude change. The analysis suggested that internalization involves more valid change than identification does and that attraction toward similar others is affected by both internalization and identification. These conclusions are supported in terms of both individual and treatment variation. The results suggest that attraction toward similar others does not necessarily reflect true attitude change but that valid change can be detected even in laboratory settings.