Inferences about teacher expectations from national assessment at key stage one

Abstract

Background. Concern has been expressed about whether the relatively low achievements of, for example, boys and African Caribbean pupils, might be due, in part, to teachers having low expectations for pupils from these groups. Developments in national assessment provide a way of looking further at this question.Aims. To show how the two measures of attainment used in national assessment—teacher assessment and Standard Assessment Tasks (SATs)—relate, and thus to draw inferences about whether and how teacher expectations vary systematically across socio‐demographic groups. Also, to show that there are different statistical approaches to this question.Samples. Data from pilot studies of national assessment at Key Stage One (ages 6 to 7); between 6000 and 7000 pupils in England; background data on gender, ethnic group and social class.Methods. Differences between the ideas of individual mismatch between teacher assessments and SATs, group mismatch and a measure of bias are elaborated, and these different ideas are linked to different statistical models for categorical data.Results. No evidence of systematic individual mismatch but some evidence of group mismatch by gender and ethnic group. The results for bias suggest that teachers might have expectations for boys, ethnic minority pupils and pupils from less advantaged backgrounds which are too low.Conclusions. Care is needed in defining aspects of mismatch and bias when studying teacher expectations. Possible teacher biases against some groups of pupils merit further investigation.