Meta-expectations: A theory of expectancy effects in social interaction

Date of Completion

January 2005


Psychology, Social




Social influence is a defining concept in social science: People are who they are because of the influence of others. However, the concept of influence is particularly vague regarding what is meant by "the other." The other has been conceptualized in two ways: the specific interaction partner and the generalized other. A general model of social influence is presented that proposes three expectations that influence behavior simultaneously: self-expectations, generalized other expectations, and partner expectations. Meta-expectations ---an individual's perceptions of others' expectations for his or her behavior---mediate the effects of others', both specific and generalized, expectations on behavior. Moreover, meta-expectations influence one another. ^ Four studies are presented that test various parts of the meta-expectation model. Study 1 manipulated both specific and generalized other expectations and communicated them to participants. Analysis of variance reveals that both generalized and specific others' expectations influence behavior. Path analysis suggests that meta-expectations mediate the effect of others' expectations on behavior. ^ Study 2 again manipulates others' expectations but partner expectations were not communicated to participants. Latent variable structural equation modeling suggests generalized other expectations affect behavior and were mediated by generalized other and partner meta-expectations. There was no effect of partner expectations on behavior. However, participants' meta-expectations about their partner's expectations did affect behavior. Studies 1 and 2 suggest accuracy of partner meta-expectations is relatively weak and demonstrate individuals infer specific others' expectations from generalized other meta-expectations. ^ In Study 3, a Social Relations Model analysis demonstrates no accuracy for meta-expectations among well-acquainted groups. Partner meta-expectations correlated with self-expectations suggesting that inaccurate partner meta-expectations may be the result of projection. Additionally, there was evidence for assumed reciprocity. ^ In Study 4, a repeated measures multivariate analysis of variance suggests that partner meta-expectations are not unique but indicate two types of others: peers and authority figures. Latent variable confirmatory factor analysis confirms the existence of two factors and suggests individuals base their specific other meta-expectations on roles. ^ The studies presented provide consistent evidence of the influence of both specific and generalized others on behavior and demonstrates that these effects are mediated by individual's perceptions of them or meta-expectations. ^