Religion, religiosity and educational attainment: evidence from the compulsory education system in England

ABSTRACT

This paper investigates the effect of religion on the educational attainment of pupils in their final year of compulsory education in England. The results show that pupils that identify with any religion have better academic performance than other pupils, after controlling for various family, parental and neighbourhood characteristics. The outperformance is reinforced by previous attendance at religious classes but there is no similar effect from considering religion to be very important to their life. Allowing for religion-specific effects shows that Muslim pupils outperform Christian pupils although the performance of the latter group is boosted by attendance at religious classes.

KEYWORDS:

• Educational attainment
• religion
• religiosity
• England

JEL CLASSIFICATION:

• I21
• J15

Acknowledgments

This work contains statistical data from ONS which is Crown copyright and reproduced with the permission of the controller of HMSO and Queen’s Printer for Scotland. The use of the ONS statistical data in this work does not imply the endorsement of the ONS in relation to the interpretation of analysis of the statistical data. This work uses research datasets which may not exactly reproduce National Statistics aggregates.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Among individuals that do not progress beyond KS4, those with five or more GCSEs (including English and Maths) graded A*-C earn around £100,000 (2013 prices) more, on average, over their lifetime than individuals that fail to achieve this benchmark (Hayward, Hunt, and Lord 2014).

² The question is phrased as follows: ‘In the last 12 months have you ever gone to classes or courses connected with any religious establishment you might go to (such as a church, chapel, synagogue or mosque)?’ This question was not asked in LSYPE in wave 3.

³ The ethnic groups follow the classification defined for oversampling (Bangladeshi, Indian, Pakistani, Black African, Black Caribbean, Other and Mixed Heritage).

⁴ An alternative strategy that is not reliant on the conditional independence assumption would be to identify a variable that determines whether a pupil belongs to a religion, is uncorrelated with the error term and has no direct effect on educational attainment and use this as an instrumental variable. We were unable to find a convincing candidate for such a variable in the dataset and are not aware of any policy change or other major event that induced quasi-experimental variation in the sample cohort’s religious affiliations.

⁵ Strand (2011) analysed an intermediate stage of compulsory schooling, KS3, whereas the present analysis focuses on the final outcome at KS4. Our specification therefore omits control variables observed at KS3 but not at KS4, namely, measures of parental involvement in school, parental supervision, absence from school, pupil’s future planning attitude, homework and academic self-concept. We also exclude variables described by Strand (2011) as ‘parental attitudes and behaviours’ and ‘student risk and protective factors’ on the grounds that these may be functions of religion. The former are measures of whether the pupils lives in a household with a computer, received private tuition, whether the parents want the pupil to continue in full-time education post-16 and whether the parents quarrel with the pupil. The latter are measures of whether the pupil has special educational needs, played truant in the last 12 months, has been excluded in the last three years or has had contact with the social services or the police as well as measures of pupil aspirations and their attitude to school.