Does immigration background matter?: How teachers' predictions of students' performance relate to student background

Investigated whether there is an unintentional ethnic bias in teachers' evaluations of immigrant students. Data from a large-scale field assessment, the Cognitively Activating Instruction and the Development of Students' Mathematical Literacy Study (COACTIV) conducted in 2003/2004 through the Program for International Student Assessment (PISA) were used. The final sample consisted of 2,088 students (between 13 and 18 years of age) and 305 teachers. Students were given two mathematical problems, one considered more linguistically complex than the other. The teachers were asked to predict whether the students would be able to solve the problems, and the predictions were compared to the students' actual performance. Multilevel analyses were used to determine whether performance was over- or underestimated or accurate and whether the students' immigration and language background played a role in the evaluations. Results revealed that teachers had a tendency to overestimate the performance of all students but that they were significantly more likely to overestimate the performance of bilingual immigrant students, specifically on mathematical word problems. It is concluded that teachers may be unaware of the language comprehension difficulties that immigrant students face and should be trained to evaluate the teaching material for potential problems.