Abstract
12 9- to 10-yr.-old boys and girls at each of two levels of test anxiety were Ss in a marble-dropping task under social reinforcement. Ss were tested by female Es expecting either that boys would perform at a higher rate than girls or the reverse. The performance measures were the base rate of response (the number of marbles dropped during the first minute of the task) and a series of six difference scores reflecting changes in rate of response from the base-rate level. The general prediction was that Es would obtain data consistent with the expectations (bias) given them. The results were: (a) no bias effects in the data collected by Es biased to expect boys to perform at a higher rate than girls, (b) for Es biased toward girls, low test-anxious children were influenced by Es' expectations (E-bias effect), but the high test-anxious children were influenced by the social reinforcement in the situation (social reinforcement effect). The implications of experimental investigations of adult expectancy effects on children's performance for a complete understanding of teacher-expectancy effects were discussed.