We used both quantitative and descriptive procedures to summarize the results of studies examining personality moderators of interpersonal expectancy effects. Five hypotheses associated with this research were identified. Three of these hypotheses specify that expecters who have a greater need to influence others, who have a greater ability to encode nonverbal messages, and who are better liked by their targets should produce target behavior more congruent with their expectancies. The remaining two hypotheses specify that targets who are more susceptible to social influence attempts and who are better decoders of nonverbal communication should be more likely to conform to their expecter's wishes. On average, results were in the predicted direction, but effect sizes were small. Expecter personality may be more influential than target personality. The relation between personality and bias may be influenced by incentives, the length of the task, and the social status of the expecter.In the early 1960s, on the basis of an unanticipated finding in his dissertation research, Robert Rosenthal began a vigorous line of inquiry meant to document and explain the existence of unconscious experimenter bias (Rosenthal, 1985). In one of the first studies of its kind, Rosenthal and Fode (1963) asked 10 graduate and undergraduate students, all of whom had research experience, to help with the development of a test of empathy.Each student experimenter showed subjects 10 photographs of people's faces and obtained the subjects' ratings concerning the degree of success or failure shown in the faces. All experimenters were given identical instructions and were told to administer the photo-rating task in identical ways. However, half of the experimenters were told that it was a well-established finding that the photos depicted successful people, and half were led to expect unsuccessful ratings. In fact, both groups used the same pictures. Results revealed that experimenters expecting more successful ratings obtained more successful ratings from their subjects.The years following this demonstration of experimenter expectancy effects witnessed an enormous amount of related research. Rosenthal and Rubin (1978) summarized over 200 expectancy studies conducted in research settings. The majority of these involved person perception tasks, but studies involving reaction time, inkblot tests, animal and human learning, laboratory interviews, and psychophysical judgments were also represented. The preponderance of evidence in each research area supported the contention that experimenter expectancies can influence research outcomes. Life Cycle of an Artifact McGuire (1969) observed that artifacts in behavioral research have a three-stage life cycle.

Researchers examining personality moderators of experimenter expectancy effects have focused on five hypotheses. Experimenters with stronger interpersonal control orientations, more positively evaluated interpersonal interaction styles, and greater ability to encode nonverbal messages are believed to be more likely to produce expectancy bias. Subjects with greater need for social approval and greater nonverbal decoding ability are believed to be more susceptible to bias. In this study each experimenter administered a photo-rating task under positive or negative expectancies to four subjects, each of whom also interacted with three other experimenters. All five personality moderator hypotheses were tested. Support was found only for the experimenter control orientation and subject need for social approval hypotheses. There was also evidence for a boomerang effect—subjects low in need for social approval gave ratings opposite to the experimenter's outcome expectancy. Finally, effects appeared stronger when positive expectancies were communicated than when expectancies were negative.

No abstract