Avoiding artifact in the search for bias: The importance of assessing subjects' perceptions of the experiment.

Ss' awareness of the experimenter's hypothesis has been recognized as a threat to the validity of laboratory experiments. After finding that the spontaneous behavior of uninstructed Ss was not similar to the behavior of Ss instructed to play a role, D. E. Carlston and J. L. Cohen (see record 1981-12761-001) suggested that concerns with this problem are exaggerated. In the present paper, methodological problems with the Carlston and Cohen study are identified, and a study designed to examine them is reported. 40 male and 40 female undergraduates participated in a conceptually similar experiment. Ss were or were not explicitly told an experimental hypothesis and were or were not instructed to play the role of a good or cooperative S. An extensive postexperimental questioning procedure was used to determine the effect of the manipulations on Ss' perceptions. Overall data from the explicit-hypothesis conditions replicated the Carlston and Cohen finding. However, only Ss who indicated on the postexperimental questionnaire that they did not believe that the explicitly stated hypothesis was the true purpose of the experiment replicated their results. When Ss were not blatantly told the hypothesis (implicit-hypothesis conditions), uninstructed control Ss and good-role-play Ss behaved similarly. Results are interpreted and discussed in terms of Ss' phenomenological views of the experiment rather than traditional S roles. (16 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)