This inquiry employs meta-analysis (1977) to account for variability in the outcomes of experiments testing the effects of teacher expectancy on pupil IQ.
The tenuous process of expectancy induction, wherein researchers supply teachers with information designed to elevate their expectancies for children actually selected at random, is viewed as the Achilles’ heel of Pygmalion (1968) experiments. It was hypothesized that the better teachers know their pupils at the time of expectancy induction, the smaller the treatment effect would be. The data strongly supported this hypothesis.
Hypotheses that the type of IQ test (group vs. individual) and type of test administrator (aware vs. blind to expectancy-inducing information) influence experimental results were not supported.
The hypothesis that expectancy effects are larger for children in Grades 1 and 2 than for children in Grades 3–6 was supported. However, surprisingly, statistically-significant effects reappeared at Grade 7.
Theoretical implications and questions for future meta-analytic research are discussed.
[I am highly skeptical of this attempt to rescue the Pygmalion effect by post hoc moderators. The claimed pattern makes zero sense, and finding such subtle interaction patterns in so few studies with such a small d = 0.11 average effect is almost certainly very statistically-underpowered.]
[APA PsycNET summary] Meta-analysis was used to examine the variability in the outcomes of experiments testing the effects of teacher expectancy on pupil IQ. The tenuous process of expectancy induction, wherein researchers supply teachers with information designed to elevate their expectancies for children actually selected at random, is viewed as problematic in “Pygmalion” experiments, as developed by R. Rosenthal and L. 1968. It was hypothesized that the better teachers know their pupils at the time of expectancy induction, the smaller the treatment effect would be. Data strongly support this hypothesis. Hypotheses that the type of IQ test (groups vs individual) and type of test administrator (aware vs blind to expectancy-inducing information) influence experimental results were not supported. The hypothesis that expectancy effects are larger for children in Grades 1 and 2 than for children in Grades 3–6 was supported. However, statistically-significant effects reappeared at Grade 7. Theoretical implications and questions for future meta-analytic research are discussed.