Teacher expectations of multilingual students’ achievement : can teachers accurately predict science performance on a linguistically accommodated test?

Assessment is a powerful tool for both teachers and students to get more insight into students’ competences. However, multilingual learners may be underper- forming on tests because linguistic barriers hinder them in demonstrating their content knowledge. In this study, we investigate whether multilingual assess- ment holds the potential to adjust erroneous teacher expectations of multilin- gual students. We discuss a randomised experiment in which 1,154 students in 5th grade (age 9 to 12) were randomly assigned to one of four test conditions: (1) a bilingual test, (2) a bilingual test with additional read-aloud accommo- dation in both the language of schooling and home language, (3) read-aloud accommodation in the language of schooling and (4) a control condition with- out accommodations. Flemish teachers were asked how well they expected their students to perform on a science test. By using chi-square tests, we examine the accuracy of teacher expectations and investigate whether this differs according to students’ linguistic backgrounds and test accommodations. The results sug- gest that when multilingual students use read-aloud accommodations in their L1, they more often tend to do better than expected by their teachers. Implications for testing culture, test accommodations and teacher expectations are discussed.Assessment is a powerful tool for both teachers and students to get more insight into students’ competences. However, multilingual learners may be underper- forming on tests because linguistic barriers hinder them in demonstrating their content knowledge. In this study, we investigate whether multilingual assess- ment holds the potential to adjust erroneous teacher expectations of multilin- gual students. We discuss a randomised experiment in which 1,154 students in 5th grade (age 9 to 12) were randomly assigned to one of four test conditions: (1) a bilingual test, (2) a bilingual test with additional read-aloud accommo- dation in both the language of schooling and home language, (3) read-aloud accommodation in the language of schooling and (4) a control condition with- out accommodations. Flemish teachers were asked how well they expected their students to perform on a science test. By using chi-square tests, we examine the accuracy of teacher expectations and investigate whether this differs according to students’ linguistic backgrounds and test accommodations. The results sug- gest that when multilingual students use read-aloud accommodations in their L1, they more often tend to do better than expected by their teachers. Implications for testing culture, test accommodations and teacher expectations are discussed.A