Over the past 50 years, the expectation states research program has generated a set of interrelated theories to explain the relation between performance expectations and social influence. Yet while the relationship between performance expectations and social influence is now axiomatic, the reported effects do differ in magnitude, sometimes widely. The authors present results from the first formal meta-analysis of expectation states research on social influence. Their findings indicate that theoretically unimportant study-level differences alter expectation states and the baseline propensity to accept or reject social influence. Data from 26 separate experiments reveal that protocol variations, including the use of video and computer technology, sample size, and the number of trials, have important but previously unrecognized effects. The authors close by discussing the more general implications of their research for future investigators.
