Give Them Time? - The Effects of Deferred School Start

Lead Research Organisation: University of Strathclyde
Department Name: Economics

Abstract

Parents with children who were born around school year cut-off dates face a dilemma: Should they send their children to school as soon as possible, thus entering them into primary school as the youngest members of their cohort? Or should they delay school entry by one year, thus entering their child as one of the oldest cohort member? Proponents of the latter practice argue that deferral equips children with more maturity which improves school readiness and long-term outcomes. Opponents argue that the costs of deferral - ranging from an additional year of nursery fees to a 1-year loss in life-time earnings - do not justify the benefits, or may even be counter-productive as underchallenged deferrers become demotivated. Policy makers who decide on funding for both (pre-)schools also have an obvious interest in the issue.

We provide new and - given recent reforms to deferral rules - timely answers that will help guide parents, practitioners, and policy makers. For this we turn to Scotland which is uniquely suited to investigate both short-term and long-term effects of school starting deferral. We use individual-level pupil census data to track the universe of Scottish pupils from 2007 until the present. Our skilled and experienced research team has a successful track record of near-perfectly linking these high-quality data with attainment data and will now extend data linkage to information related to pupils' behaviour, health, qualifications, and post-secondary school destinations. We are also among the first to analyse the effect of deferral on the experience and the act of bullying.

To ensure an innovative and robust research design, we exploit two natural experiments. First, long-standing cut-off dates determine whether pupils are eligible for deferral. Our regression discontinuity design compares pupils who will typically start school in the year they turn five (e.g. those born on December 31st) with pupils who are likely to defer school entry by 1 year (e.g. born on January 1st). As long as the only difference between pupils on either side of the threshold are differences in deferral rates, we can isolate the causal effect of deferral. Second, in 2021 parts of Scotland were chosen to pilot a scheme that provides automatic nursery funding entitlement to all children who defer school entry. We will analyse the short-run effects of this reform by way of a differences-in-differences approach that compares changes in outcomes among pupils in local authorities that are part of the pilot with pupils from local authorities without the entitlement.

In order to ensure impact, we partner with several stakeholders. The Scottish Government (SG) has agreed to provide access to the various datasets required for this project. SG has also expressed an explicit interest in our research ensuring that our findings will indeed inform and shape policy. We will take advantage of the excellent networks that the researchers' institutions - including the Fraser of Allander Institute at Strathclyde and Robert Owen Centre for Educational Change at Glasgow - have with key local and national stakeholders. These institutions also have outstanding platforms to disseminate research outputs to a wide audience. These links will ensure input from a range of practitioners and academics and facilitate dissemination across disciplines and local policy makers.

An important secondary goal and legacy of this project is to build a pupil database for education researchers in Scotland and beyond. Researchers can use this resource for future projects or build on our work by adding more information as it becomes available. Combined with free-of-charge data provision by way of our partnership with SG, this aspect ensures that our project offers excellent value for money. To support wide adaption and to further capacity building, we have assembled a team stretching across 3 Scottish universities that includes both early career and experienced researchers.

Publications

10 25 50