Judging students' academic self-concepts accurately is assumed to be a necessary condition for adaptive classroom instruction. Previous studies found moderate correlations between teachers' judgements and the selfreported self-concepts. Reference values for interpreting these correlations are missing. In the present study, self–other agreements with a zero-acquaintance premise (see Ambady & Rosenthal, 1992) were proposed as such a reference value.
A total of 88 teachers judged the self-concepts of their own students. Additionally, 97 persons judged students' self-concepts based on 30-second videos of the students. Intra-individual correlations were calculated between judgements and actual self-concepts.
The analyses showed that agreements in the natural classroom setting and in a zero-acquaintance setting both displayed generally moderate correlations and did not differ significantly from one another.
Based on the results, the accuracy of teachers in judging students' self-concepts is considered to be low. Possible reasons for the low accuracy of teachers' judgement accuracy are discussed.