Nuffield Early Language Intervention (re-grant)

This page covers the second (effectiveness) trial of Nuffield Early Language Intervention, testing a scalable model under everyday conditions in a large number of schools. To read about the first (efficacy) trial - testing whether it could work in schools under best possible conditions - click here.

The Nuffield Early Language Intervention is designed to improve the spoken language skills of children in the first year of primary school. Targeted at children with relatively poor spoken language skills, it is delivered to groups of two to four children, three times a week, alongside some individual sessions. Trained Teaching Assistants (TAs) run the programme, which lasts for 20 weeks during their first year of school (Reception). Sessions focus on listening, narrative and vocabulary skills.

The intervention was developed by academics at University of Oxford (led by Charles Hulme and Maggie Snowling) together with team members from Sheffield (Silke Fricke) and York (Claudine Bowyer-Crane) with funding from the Nuffield Foundation. A previous EEF evaluation found a positive impact on language skills. This trial will test the Reception programme in 200 schools. The resources are being published by Oxford University Press, and training for TAs will be provided by Elklan – a specialist training agency focused on supporting speech and language development.

Why are we funding it?

This trial follows on from several previous studies of the intervention. A 30 week intervention (starting in Nursery then continuing into Reception) was initially developed and evaluated with funding from Nuffield Foundation, and training delivered by the developers. This suggested very large impacts on language skills (effect sizes of 0.8 standard deviations). The EEF funded an efficacy study in 34 schools, with the developers stepping back from training and with analysis done independently by IFS, found an impact of 0.16 (2 months) for the 20 week Reception intervention. It also followed the children 6 months later, and found the effect grew. Both studies were randomised trials, but with individual level randomisation and in a relatively small numbers of schools. This new effectiveness trial will randomise at the school level, and will test whether a scalable version of the training and resources can produce similar results.

How are we evaluating it?

A team from RAND Europe, led by Julie Bélanger and Alex Sutherland,  will independently evaluate the programme. At least 200 schools will be recruited and randomised to either be in the intervention group or the control group.

Before randomisation, teachers will be asked to assess the language skills of the pupils using a simple app being developed by the delivery team. Alongside standardised measures of language, this will identify the target pupils (5 per class). Delivery of the training will be in late 2018 and intervention will begin in January 2019.

Outcomes of the target pupils will be assessed at the end of the academic year (summer 2019), using a range of language assessments that will be integrated into a composite measure. An implementation and process evaluation will assess whether the intervention is being delivered as intended.

This trial follows an initial phase of recruitment, pre-testing and intervention delivery in 2017/18. Post-test data will not be collected from this full sample due to concerns about the data quality, but the work has informed this full evaluation and shown that the delivery partners can work with the required number of schools. 

When will the evaluation report be due?

The evaluation report will be published in Spring 2020.