Gary D. Barko

Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)




Previous research and discussion has suggested that task ambiguity and experimenter-subject familiarity with the task are important factors affecting the generality and strength of the experimenter bias effect. These factors were conceptualized as inter-task ambiguity (inherent in the task) and two types of intra-task ambiguity (experimenter's familiarity with the task and subject's familiarity with the task). It was hypothesized that experimenter bias would: (1) be a significant factor in a more ambiguous task and not a significant factor in a less ambiguous task, (2) more likely be communicated to the subjects who were less familiar with the task, and (3) more likely be communicated to the subjects as the experimenters became more familiar with the task.

Nine upperclassmen experimenters were randomly assigned to one of three bias level, high, medium, and low (expect +5, 0, or -5 mean ratings on the person perception task). Each experimenter tested 10 subjects (female volunteers from introductory psychology) in a random order. Five subjects rated high ambiguity photos (mean ratings near 0 on the success-failure dimension under a no-expectancy condition) and five viewed low ambiguity photos (mean ratings high in the success direction under the no-expectancy condition).

Five analyses of variance were computed; a significant experimenter bias effect was found with the high ambiguity stimilus photos but not with the low ambiguity photos. However, the bias levels were not communicated in the predicted order. The high and low expectancy experimenters were both significantly higher than the medium bias experimenters. The hypotheses pertaining to intra-task ambiguity were not supported by the data.

From this study one can conclude that the experimenter bias effect is a function of ambiguity inherent in the task. The role of intra-task ambiguity is less clear and no conclusion can be stated based on the present data.