Parental social class and school GCSE outcomes: two decades of evidence from UK household panel surveys
Parental social class and school GCSE outcomes: two decades of evidence from UK household panel surveys
Key takeaways
Bibliography: Stopforth, S., Gayle, V., Boeren, E., 2021. Parental social class and school GCSE outcomes: two decades of evidence from UK household panel surveys. Contemporary Social Science 16, 309–324. https://doi.org/10.1080/21582041.2020.1792967
Authors:: Sarah Stopforth, Vernon Gayle, Ellen Boeren
Tags: #Methods, #Literature-Review, #Statistical-methods, #Synthetic-Cohorts
Collections:: Methods
First-page: 309
This paper investigates social class inequalities in English school qualifications. The analytical focus is pupils’ outcomes in General Certificates of Secondary Education (GCSEs). The original aspect of this paper is the operationalisation of data from the British Household Panel Survey (BHPS) and the UK Household Longitudinal Study (UKHLS), which facilitates analyses from 1991 to 2013. We observe a general trend of improved educational outcomes in more recent cohorts of school pupils, which is consistent with national results. The central empirical finding is that there is a persistent social class gradient. Pupils growing up in families in less advantaged social classes have less favourable school GCSE outcomes. This is especially concerning, because having fewer good GCSEs is likely to limit children’s participation in more advanced education and restrict their options in the labour market. Changes in the structure and content of GCSEs lead us to conjecture that sociological analyses of social class inequalities in school qualifications will continue to be important. We highlight the limitations of using administrative educational data, and we outline the data resources that would better facilitate the study of social class inequalities.
content: "@stopforthParentalSocialClass2021" -file:@stopforthParentalSocialClass2021
Reading notes
Imported on 2024-06-26 11:24
⭐ Important
- & The central empirical finding is that there is a persistent social class gradient. Pupils growing up in families in less advantaged social classes have less favourable school GCSE outcomes. (p. 309)
- & The overall message has been that children living in families in less advantaged social classes generally have less favourable educational outcomes (see Corrigan, 1979; Douglas, 1964; Halsey, Heath, & Ridge, 1980; Hargreaves, 1967; Lacey, 1971; Wedge & Prosser, 1973; Willis, 1977) (p. 309)
- & qualifications that English children gain at school are important because they are strongly related to participation in post-compulsory education (Payne, 2000), youth unemployment (Rice, 1999) and future labour market experiences (Babb, 2005; Jones, Joyce, & Thomas, 2003; Murray, 2011). (p. 309)
- & Pupils from families in more advantaged social classes, for example those characterised by managerial, administrative and professional occupations, have more favourable GCSE outcomes than pupils from families in less advantaged social classes (Connelly, Murray, & Gayle, 2013; Connolly, 2006; Demack, Drew, & Grimsley, 2000; Gayle, Murray, & Connelly, 2016; Playford & Gayle, 2016; Strand, 2014; Sullivan, 2001). (p. 310)
- & The first is that there is a lack of suitable large scale, nationally representative youth data. (p. 310)
- & The second obstacle is that despite the UK having an impressive set of birth cohort studies, the 1946, 1958, and 1970 cohorts pre-date the introduction of GCSEs (p. 310)
- & The third obstacle, however, is that administrative educational data resources do not ordinarily include detailed measures of parental social class. (p. 310)
- & Taylor (2018) commented that FSM eligibility does not capture the broader multi-dimensionality of social advantage and disadvantage. (p. 310)
- & We matched the young person’s GCSE information with information on parental social class, parental education, and other measures, which were collected from interviews with both the young person and their parents. (p. 311)
- & The timing of the data collection, in both studies, does not neatly map onto the English school year. We have carefully organised the data into ‘synthetic’ school year cohorts. The mechanics of the data wrangling processes are elaborated in Stopforth (2020). (p. 311)
- & There is no single overall or ‘agglomerate’ measure of school GCSE outcomes. Following Schmitt and Wadsworth (2006), Connelly et al. (2013), Shakeshaft et al. (2013), and Gayle et al. (2016), we use the number of GCSEs gained at grades A*–C as the outcome measure throughout this paper (p. 312)
- & here is also evidence of an association between household tenure and children’s school outcomes (Connelly et al., 2013; Gayle et al., 2016). (p. 312)
- Nice table - emulate this in thesis:
- & Table 1. Descriptive statistics and mean school GCSE outcomes (grades A*–C). (p. 314)
- & Models for count data are most appropriate for modelling the number of GCSEs gained at grades A*–C (see Cameron & Trivedi, 1998). A sensitivity analysis comparing alternative statistical models suitable for count data was undertaken separately on the two samples.6 A significant likelihood ratio test provided evidence of over-dispersion and therefore the negative binomial regression model was preferred to a Poisson model (see Long & Freese, 2014). There were high percentages of young people with zero counts in both samples. A significant Vuong test provided evidence that a zero-inflated model was most suitable for these data (see Vuong, 1989). (p. 315)
- Look into this - important:
- & The models are adjusted for the complex designs of the surveys (see ‘svy’ in StataCorp, 2017) (p. 315)
- & The negative binomial regression model does not have an equivalent to R2 in OLS regression. Cox-Snell Pseudo R2 measures were calculated, but this measure is illustrative rather than confirmatory (see Long & Freese, 2014). (p. 320)