Abstract
Biases of +5 (positive bias) or −5 (negative bias) were induced in 4 male and 4 female Es high on n Avoid Success and 4 male and 4 female Es low on n Avoid Success. Instructions tape-recorded by these Es for Rosenthal's photo-rating task were then played for 64 male and 64 female Ss who rated 8 photographs on degree of perceived success or failure by means of a rating scale which ranged from +10 (extreme success) to −10 (extreme failure). Failure of either the chi-square test or the analysis of variance to indicate a significant experimenter bias effect questions Barber's acceptance of previous research which indicated that “the experimenter bias effect is produced by subtle (paralinguistic) cues” but supports his contention that the effect is not the pervasive and easily demonstrable phenomenon Rosenthal has claimed. The only analysis of variance component to reach statistical significance, Ss' sex, appeared due to a response set in which female Ss tended to view individuals pictured in the photo-rating task as more successful, regardless of E's induced bias.