How is social inequality reproduced by teachers’ recommendations and expectations? Examining implicit bias among teachers.
Underperforming students with an ethnic minority status or a socio-economically disadvantaged background are more likely to repeat a grade or switch programs compared to similarly performing peers. As gatekeepers, teachers play a crucial role in altering students’ trajectories by recommending interventions such as grade retention and tracking. This study examines how teacher expectations and recommendations reproduce social inequalities in secondary education. Drawing on literature linking student sociodemographic characteristics to biased teacher judgements, we investigate whether these biases shape expectations for higher education enrollment and how these expectations impact recommendations. Using a factorial video vignette experiment among 956 teachers in 46 schools, teachers evaluated fictive students varying in SES, ethnicity, name, and language proficiency, while holding academic performance constant. Multilevel regression showed higher expectations for more privileged students. SES and ethnicity emerged as predictors, with high-SES and ethnic majority students receiving higher expectations. Name cues and language proficiency also influenced expectations. Remarkably, the most privileged students were most likely to be recommended a program change, with no differences between profiles for grade retention. It may be that teachers assume underprivileged students to be in a lower track, as expectations about past performance were similarly biased. Implications are discussed.
Authors
- Jannick Demanet (ORCID: https://orcid.org/0000-0003-2770-8917)
- Joana Pipa (ORCID: https://orcid.org/0000-0001-7150-145X)
- Timo Van Canegem (ORCID: https://orcid.org/0000-0002-3817-6146)
- Mieke Van Houtte
- Francisco Peixoto
Publication Details
- Journal
- Ghent University Academic Bibliography (Ghent University)
- Published
- 2026-01-01
- Primary Topic
- School Choice and Performance
- Type
- article
- Field-Weighted Citation Impact
- 0.00