Socioeconomic background, gender and subject choice in secondary schooling

Abstract

Background: The desirable extent of curriculum choice to be offered to students remains a central policy question in England. Previous studies of the impact of the introduction of a common curriculum for 14–16 year olds in 1988 have suggested that some gender differences were narrowed as a result. These studies examined subject choice either in terms of students' ex ante preferences in advance of enrolling for subjects or in terms of ex post aggregate data on examination entries. There is some conflict between the evidence provided by these two sources. One possible reason for this conflict could be that existing ex post evidence does not examine the interaction between student characteristics or the effect of school-level variables.

Purpose: This study aims to identify effects of social class and gender on subject choice for 14–16 year olds in England over and above effects that are attributable to students' ability. Effects that operate at school level are separated from those that act at the level of the individual.

Sample: The sample is drawn from the schools that participated in the Yellis system for providing analysis of the examination results achieved by 16-year-old students. The sample of 664 schools and 112,412 16-year-old students was selected by including all schools who had participated in the Yellis project for at least five years during the period 1994–2002. All of the sample schools were participating in the scheme in 1998, the year for which a cross-sectional analysis is presented in this paper.

Design and methods: Statistical (probit) models are used to investigate effects of student and school characteristics on the probability of a student entering for examination in each of six option subjects: Business Studies, French, Geography, German, History and Home Economics. The models take account of levels in the data and identify interactions between the student characteristics.

Results: After taking prior ability into account socioeconomic background effects, taken together, exert a stronger effect than gender on the likelihood of entering for examination in history. Socioeconomic background effects are also stronger in the case of business studies. The effect of socioeconomic background is stronger for females than for males. There are also strong social effects operating through the characteristics of the cohort of students at the school. The proportion of students eligible for free school meals has a significant effect on the probability of entering for examination in geography, German or History.

Conclusions: In so far as the results from this study can be compared with previous research they support the conclusions of previous ex post studies rather than ex ante studies in terms of gender preferences in subject choice. This might suggest some difficulty in generalising from the kind of ex ante data gathered previously. The evidence of the effect of socioeconomic background at individual and school level suggests that current policy aiming to increase subject choice within and between schools will deepen differences between the subjects studied by students from different socioeconomic backgrounds.

Keywords:

• subject choice
• examination entries
• social class
• gender

Acknowledgements

This paper is an outcome from ESRC Project RES-000–22–0090 ‘Within-school competition and pupil achievement’. We are grateful for helpful comments received from colleagues attending the British Educational Research Association annual conference, UMIST, September 2004, and three anonymous referees.

Notes

1. A separate probit regression was used to check the estimation of the effects of the variables that can be related to socio-economic background . This comparison used one subject, history, and included just the variables student ability, mother's occupation and books at home. There was very little variation in the pseudo-R 2 (from 0.0387 to 0.0365), thus justifying the specification presented in the paper.

2. For that reason, the model was carefully run again and the result confirmed.