Department of Education

Viewing archives for Quantitative Methods Hub

Young people who have had experience of being in the care of the state, usually due to neglect or maltreatment, are disproportionately likely to be not in education, employment or training (or NEET) in early adulthood.  This session will outline some of the findings from the quantitative strand of a study funded by the Nuffield Foundation and due to report in January 2022: http://www.education.ox.ac.uk/research/care-leavers-employability-in-england.  In particular, it will use binary logistic regression to explore employment outcomes in young people’s 21st year using linked administrative data from the National Pupil Database, Individualised Learner Records and the Longitudinal Educational Outcomes dataset.  The analysis compares the general population with young people with different levels of engagement with the social care system, with a focus on identifying particular supporting or risk factors.

Research at the intersection of social science and genomics, “sociogenomics”, is transforming our understanding of the interplay between genomics, individual outcomes and society. It has interesting and maybe unexpected implications for education research and policy. Here we review the growing sociogenomics literature and discuss its implications for educational researchers and policy makers. We cover key concepts and methods in genomic research into educational outcomes, how genomic data can be used to investigate social or environmental effects, the methodological strengths and limitations of genomic data relative to other observational social data, the role of intergenerational transmission, and potential policy implications. The increasing availability of genomic data in studies can produce a wealth of new evidence for education research. This may provide opportunities for disentangling the environmental and genomic factors that influence educational outcomes and identifying potential mechanisms for intervention.

A significant proportion of pupils move school during their school career for reasons other than standard structural moves between educational stages. Little is known about the underlying causes of these moves and the characteristics and experiences of mobile pupils are challenging to research. This seminar presents analyses of the English National Pupil Database (NPD), tracking a cohort from age 5 to 16, to better understand when school moves occur, the characteristics of mobile pupils and the impact of mobility on attainment. Findings reveal a sizable underlying rate of moves in England of about 1.5–2% per term and identify differences in mobility related to disadvantage, school phase, ethnic group and SEND status. The predictive power of the data for identifying mobile groups, however, is low, highlighting the need for more research, policy and practice in this area to better understand individual mobility circumstances. With regards to attainment, analyses identify moves during the secondary phase – but not the primary phase – as being significantly associated with lower pupil academic progress. This seminar presents results from across these analyses and discusses their implications for policy, practice and further research.

We present an analysis of A-level subject choices at around age 16 for a cohort of students in English schools who completed their studies in 2014. We examined both the National Pupil Database and a unique rich dataset on the subject preferences and subsequent choices between the ages of 16 and 18 (i.e. GCSE and A-level). We found substantive differences between students’ preferences and actual choices of ‘hard’ and ‘soft’ post-16 subjects (i.e. A-level). These differences were strongly associated with falsification of students’ expectations of examination grades taken at age 16 (i.e. GCSE) in the core subjects of English and mathematics. The sizes of these falsification effects were much larger than other significant associations such as gender, ethnicity, and social class. This suggests that subject choices are not rigidly framed by stable individual preferences and they are therefore open to influence from new information, persuasion, and opportunities.

This paper examines inequalities in the match between student and degree quality using linked administrative data from schools, universities and tax authorities. We analyse two measures of match at the university-subject level: undergraduate enrolment qualifications, and graduate earnings. We find for both that disadvantaged students match to lower quality degrees across the entire distribution of achievement, in a setting with uniform fees and a generous financial aid system. While there are negligible gender gaps in academic match, high-attaining women systematically undermatch in terms of expected earnings, driven by subject choice. These inequalities in match are largest among the most undermatched.

Alternative Provision (AP) schools provide education for pupils who can’t go to mainstream schools. Although they are commonly associated with pupils who have been permanently excluded, they also cater for pupils with a range of other needs including school refusers, those who are ill and those without a school place. Despite increased attention being paid to the sector in recent years, very little information is published about pupils who spend time in AP or their outcomes. In this session I will present work I have been undertaking to fill in some of the evidence gaps on AP using administrative data. Around 3% of pupils spend some time at a state-funded AP school during their school career. However, their outcomes (attendance, attainment and post-16 destinations) tend to be poor. More broadly, the session will cover the administrative data resources available to researchers studying the education system in England. These allow detailed investigation of narrow segments of the population for which data may be sparse in other sources (e.g. survey datasets).

In a paper recently published on majority-Muslim hate crimes, as well as the present paper on segregation in schools, I rely on descriptive quantitative mappings as a backdrop for qualitative research. In my presentation, I will focus on cross-over methodology but also present some key results. This paper examines ethnic segregation in schools as a case of political inaction, or non-decision making. Since non-decision making is about the paths not taken, its study requires a combination of careful factual observation with an ethnographic approach to the counterfactual policy analysis. To enable such analysis, the focus of this study is on Halifax, an English town presenting archetypal scenarios of ethnic segregation and mixing in schools and neighbourhoods. After presenting a descriptive “landscape” of ethnic segregation and mixing, the subsequent qualitative analysis offers insight into the inner workings of value-based and procedure-based non-decision making that have wider application, transcending any given locality and sphere of public policy.

Please register ahead of the webinar at this link.

Dr Gabriella Conti

Early intervention programmes can play an important role in improving children’s health and their cognitive and socio-emotional development. While existing evidence demonstrates the potential benefits that these programmes can have, particularly for the most disadvantaged, much less is known about the factors that drive effectiveness in scaled-up programmes.

In this paper, we investigate the important but under-researched question of workforce quality within the context of the Family Nurse Partnership (FNP). This is a large-scale home-visiting programme in England targeting first-time teenage mothers, which has previously shown benefits for children’s cognitive development (Robling et al., 2015). For identification, we exploit a unique feature of the assignment process of the family nurses to the clients within the FNP teams.

Conditional on a small set of variables governing the assignment process, nurses were assigned to clients to equalise caseloads within teams. We present evidence that, for a wide range of client and nurse characteristics, there is no systematic relationship between clients and nurses conditional on these assignment variables. We then present results on the effects of family nurse quality on the cognitive, socio-emotional and health outcomes of the child.

First, we find evidence of substantial heterogeneity. A one-standard deviation (SD) increase in family nurse quality leads to a 0.22 SD increase in birthweight, to a 0.25 SD increase in child’s cognition at age 2, and to a 0.29 SD increase in child’s socio-emotional development at age 2. We also show a strong correlation between nurse effectiveness for different outcomes; of the nurses who are in the top quintile of effectiveness in boosting cognitive development, almost half are in the top quintile for socio-emotional development. We also find evidence that nurses can improve maternal mental health and reduce unhealthy behaviours, potentially mediating their effect on children.

However, despite a very rich set of characteristics such as demographics, training, experience and qualifications, we can only explain between 10 and 15% of this variation in family nurse effectiveness. These results are reminiscent of the literature on teacher quality, where observable characteristics have little power in explaining variation in teacher’s value-added. Our results show that the quality of the workforce matters, and that we are just starting to understand its determinants.

Please register ahead of the webinar at this link.

Dr Matt Dickson

We compare estimates of the effects of education on health and health behaviour using two distinct natural experiments in the UK Biobank data. One is based on a widely used policy reform while the other, known as Mendelian randomization (MR), uses genetic variation. The policy reform is the raising of the minimum school leaving age (RoSLA) from 15 to 16 which took place in the UK in 1972.

MR exploits germline genetic variation that associates with educational attainment and is a strategy widely used in epidemiology and clinical sciences. Under the assumption of monotonicity, these approaches identify distinct local average treatment effects (LATEs), with potentially different sets of compliers. The RoSLA affected the amount of education for those at the lower end of the education distribution whereas MR affects individuals across the entire distribution.

We find that estimates using each approach are remarkably congruent for a wide range of health outcomes. Effect sizes of additional years of education thus seem to be similar across the distribution. Our study highlights the usefulness of MR as a source of instrumental variation in education.