Document Type


Degree Name

Master of Arts (MA)




Faculty of Science

First Advisor

Alan Auerbach

Advisor Role

Thesis Committee Member

Second Advisor

Sid Hellyer

Advisor Role

Thesis Committee Member

Third Advisor

Josephine Naidoo

Advisor Role

Thesis Committee Member


Scientific researchers are expected to control the “Experimenter” or “Rosenthal” effect, in which the experimenter’s (E’s) expectations somehow induce the experimental results. Some seventy studies dealing with experimenter bias effect (EBE), many of them recent, were critically reviewed. Contradictory results and failures to replicate were reported. The most plausible explanation involved the personality characteristics of, and the dominance relation between, E and S. Accordingly, the present study examined one personality trait which seems to underlie the EBE, namely, submission to authority.

To elicit EBE, students serving as Es administered Ss Rosenthal’s Photo Rating Test (RPRT), a series of photographed faces which Ss rate on an ordinal “success-failure” scale. Some Es were led to expect that their Ss would perceive “success” in the faces; other Es expected their S to perceive “failure.”

In a preliminary study RPRT was examined by having students rate the pictures on two occasions under neutral, i.e., “no-expectancy” conditions. Only 11 out of the 20 pictures received near-zero ratings, and a large rating dispersion was found. Consequently, results were calculated separately for all the pictures and for the most “neutral” ones.

Focussing on S personality variables, Experiment One tested the hypothesis that Ss who were more submissive to authority were also more susceptible to EBE. Ss selected for either high or low dominance were given RPRT by medium dominant Es were led to expect high ratings and others to expect low ratings from their Ss.

Simultaneously, focussing on E variables, Experiment Two tested the hypothesis that dominance and submission in E affected S’s susceptibility to EBE. The task and the expectancies given to Es were the same as in Experiment One.

In both experiments E and Ss were classified as dominant, medium dominant or submissive if they scored within the upper, middle, or lower range of their sample’s distribution on Gold’s Dominant-Submission Scale. The interaction between dominance and experimental situation responses was examined bya hidden observer, recording smiles, glances and test duration, and by post-test questionnaires.

No EBE was found in either experiment, yet submissive Ss reported a significantly higher pressure to rate the pictures in a certain way than dominant Ss.

The results of both studies were interpreted in terms of problems surrounding this area of research. Finally, a proposed model of personality and situational variables which can be expected to elicit EBE was presented.

Convocation Year


Included in

Psychology Commons