Examined a natural quasi experiment as a test of the teacher expectancy hypothesis. 1st-grade achievement scores were obtained for 79 younger siblings who had been preceded in school by bright or dull older siblings. It was hypothesized that pupils taught by the same teacher as their older siblings (expectancy condition) would perform better than those taught by a different teacher (control) if their older siblings had been good students and worse than the controls if their older siblings had performed poorly. Analyses of variance indicated significant (p < .05) Expectancy Condition * Older Sibling Performance interactions on 3 achievement subtests and a 3-way interaction of these with the sex of the younger sibling on a 4th subtest. Results provide support for the teacher expectancy hypothesis and an indication of 1 means by which expectancies can be induced naturally in the classroom. (23 ref) (PsycINFO Database Record (c) 2016 APA, all rights reserved)