Towards a typology of gender‐related school effects: some new perspectives on a familiar problem

Abstract

The superior performance of girls, especially in terms of GCSE examinations, continues to attract attention. Much of the debate to date, however, has focused on absolute differences in attainment rather than differences in pupil progress. Furthermore, the role of the school in contributing to the differential progress of the two sexes has been little explored in previous research on school effectiveness which has been largely gender‐blind. The research reported here tracks the progress of over 450,000 pupils in almost 2,700 English mixed secondary schools from Key Stage 3 to GCSE.First, a typology for understanding gender‐related school effects, which relates schools' overall ‘effectiveness’ to the relative progress of the two sexes in individual schools, is developed. Then the extent to which schools are distributed across the different cells of the typology is considered. The analysis confirms expectations that there are a very considerable number of schools where girls have been making greater progress than boys between KS3 and GCSE and also that there are hardly any where boys' progress has been superior to that of girls. An unexpected finding, however, was the discovery that in almost half the schools in the country the progress of the two sexes was broadly equal. Furthermore, in a substantial minority of schools raising the overall levels of performance of both sexes may well be a more pressing priority than reducing the gender gap. The implications of the framework for the development of strategies designed to reduce the ‘gender gap’ are then considered.

Figure 1b: GCSE point score v. KS3 average levels for two schools with different levels of effectiveness

Figure 1a: GCSE point score v. KS3 average level—national lines for boys and girls

Notes

Faculty of Education, Cambridge University, Trumpington House, Hills Road Site, Cambridge CB2 2PH, UK. Email: [email protected]

Most of the research to date using large‐scale data‐sets has been conducted on secondary schools but see, for example, Sammons et al. (1993 and 1997) and Strand (1999). Recently, some more extensive work on primary schools, based on national data sets, has begun (see, for example, Schagen & Benton, 2003 and Kyriakides, 2003).

The analysis might be extended, in due course, to include pupils not entered for any examinations or those statemented for SEN purposes.

It would have been useful to have had information about individual pupils in terms of whether they were in receipt of free school meals as well as their ethnic background. Unfortunately, these variables only became available for the DfES's 2002 cohort, not the 2001 one.

Two basic Models (referred to here as 1 and 2) were run. As part of the exploratory work, however, a number of variations on each model were also considered (referred to here as a, b, etc.).

The relevant figures are shown in Appendix 1. Figure 6a provides the basis for Table 3b. It should be noted that Figure 4a in Appendix 1 has been supplied for completeness but has not been employed in constructing Table 3 because the assumptions about ‘overall effectiveness’ have not been varied.

The relevant cells are 8 and 9. Added together these produce figures of 32.7% (Table 3a) or 32.3% (Table 3b).

Cell 3 in Tables 3a and 3b and, possibly, Cell 6.