The briefs explore whether (in)accurate teacher perceptions are systematically linked to students’ socio-economic and ethnic background, or gender.

• Preliminary Results of the BIPE Project I – Ethnicity (PDF, 6KB)
• Preliminary Results of the BIPE Project I – Gender (PDF, 6KB)
• Preliminary Results of the BIPE Project I – Income Quintile (PDF, 6KB)
• Preliminary Results of the BIPE Project I – Social Class (PDF, 6KB)

These preliminary results investigate the discrepancy between students’ own rating of their academic ability (i.e., academic self-concept) and teachers’ rating of students’ academic abilities.

In both cases, such discrepancies (i.e., the disagreement between student ratings and teacher ratings of, for example, mathematical skills) are captured by the idea of inaccurate perceptions, where an inaccurate perception is defined as the teacher over- or under-estimating a student academic ability, in comparison to the students’ own rating of it.

The briefs explore whether (in)accurate teacher perceptions are systematically linked to students’ socio-economic and ethnic background, or gender.

• Preliminary Results of the BIPE Project II – Ethnicity (PDF, 5KB)
• Preliminary Results of the BIPE Project II – Gender (PDF, 5KB)
• Preliminary Results of the BIPE Project II – Income Quintile (PDF, 5KB)
• Preliminary Results of the BIPE Project II – Social Class (PDF, 5KB)

Publications

We are currently working on three papers. Here are the abstracts:

Ethnic biases in primary school teachers’ perceptions of their students’ academic attitudes: evidence from the British Millennium Cohort Study

In this paper, we use again data from the Millennium Cohort Study (MCS), a large-scale, rich survey on children and young people in the UK. We analyse whether there is a relationship between student ethnic background and teachers perceptions of students’ attitudes towards school. We focus on students aged 10 to 11 years in England. We assess the discrepancy between teacher’s perceptions and student’s own reports of the student’s report on whether they enjoy school, whether they give their best in school, misbehave in class and find school boring. The discrepancies are calculated again using the residual method outlined above: a linear regression model predicts teacher reports with student report as main independent variable and, subsequently, the ‘distance’ between a teacher report and the regression line (residual) represents the discrepancy. We argue that larger discrepancies or ‘biases’ reflect misunderstandings and conflict in the teacher-student relationship.

Preliminary analyses suggest that teachers tend to, on average and ‘compared’ to the students’ own reports, under-estimate how often students from families who identify as Pakistani, Black Caribbean, and Black African background try their best at school and enjoy school.

Student SES and teacher-student perceptual discrepancies of school effort and enjoyment

Teacher perceptions of student academic attitudes influence teacher grading behaviour and other important teacher judgements. Moreover, congruence between teacher and student perceptions reflects positive teacher-student relationships and is critical for the creation of effective learning environments. This study investigates discrepancies between teacher and student perceptions of student’s school enjoyment and effort, and whether these discrepancies are associated with student socioeconomic status (SES). It draws on representative survey data on children at the end of primary school in England (UK Millennium Cohort Study) and Scotland (Growing Up in Scotland study) and uses a residual method to compute perceptual discrepancies. We find that, compared to students’ own perceptions of their enjoyment and effort, teachers significantly underestimate the effort and enjoyment of low SES students and overestimate the same attitudes for high SES students. The association between SES and discrepancies between teacher and student perceptions remains significant even when student prior ability and socioemotional and behavioural difficulties are held constant. Results for England and Scotland are generally similar indicating that education systems have little impact on social inequalities in perceptual discrepancies between teachers and students.

Socioeconomic bias in teacher assessments of primary school students: the mediating role of student attitudes, student behaviour and parental involvement

This paper aims to investigate socioeconomic biases in teacher assessments of school performance of primary school students. A social bias in teacher assessments is typically measured as a discrepancy between teacher assessment (e.g., ratings of student academic performance) and student scores on standardised ability tests that is systematically linked to the student’s social class or socioeconomic status. A core limitation of existing research is that experimental studies – which are common in the field – although powerful in uncovering causal relationships, lack external validity as they usually involve teachers assessing fictional student cases. Moreover, they are limited in their ability to uncover the processes that underlie these biases. These socioeconomic biases in teacher assessments might arise because teachers unconsciously factor in some student characteristics that relate to their socioeconomic background. Examples of these characteristics could be students’ behaviour and attitudes towards learning or parental involvement in children’s education. Other factors that have received relatively little attention so far are “macro-level factors” such as education policy and teacher training.

In this paper we use three large-scale secondary data sets – the British Millennium Cohort Study (MCS – England), Growing Up in Scotland (GUS – Scotland), and the National Education Panel Study (NEPS – Germany) – to (1) identify socioeconomic biases in teachers’ assessments of students’ academic abilities, and (2) analyse the extent to which these biases are mediated by student behaviour in the classroom, academic attitudes, and parental involvement in children’s education. We use linear regression analysis and causal mediation analysis. All three data sets provide comparable measures of subjective teacher assessments of students’ abilities in different subjects and various cognitive ability measures gathered through standardised ability tests. The longitudinal structure of the data sets helps to minimise methodological problems that could arise through issues such as reverse causality. The multi-country approach enables us to explore how “macro-level factors” such as education policy and teacher training influence teacher bias.