A review of the evidence on early language development

A new review of the evidence on early language development, commissioned by the Education Endowment Foundation in partnership with Public Health England, has examined the most effective ways to support young children with delays in their early language development between birth and five years old.

James Law and colleagues looked at the existing evidence to find out which interventions have the greatest potential for boosting young children’s language skills and reducing inequalities in outcomes. They identified 44 intervention studies which focused on language and related skills in pre-school. All the studies were randomised controlled trials or quasi-experimental, matched study designs. Positive effect sizes were found in relation to receptive language in 29 studies. They found one of the best ways to improve early language development for this group is by training teachers in early years settings so that they can deliver cost-effective and evidence-based interventions to those children who have fallen behind.

In addition to high-quality early years provision, the researchers identified interactions with parents as key, highlighting the need to promote positive interaction between parents and their children before they start pre-school.

The report also stresses the need for better monitoring of children’s progress at different stages of their development, to catch those children falling behind and to identify those who need targeted, specialised support.

Source: Early Language Development: Needs, provision and intervention for preschool children from socio-economically disadvantage backgrounds (October 2017), Public Health England and Education Endowment Foundation

What difference does it make?

We regularly quote effect sizes in Best Evidence in Brief as a measure of the impact of an intervention or approach. But what is the impact of a normal school year on children, and how much of that impact is due to the school? A study by Hans Lutyen and colleagues, published in School Effectiveness and School Improvement, attempts to find out.

The study analysed 3,500 pupils from 20 mostly independent (private) English primary schools on four different learning outcomes. These measures, part of the Interactive Computer Adaptive System (InCAS), were reading, general maths, mental arithmetic and developed ability, the last of which measures items such as vocabulary and non-verbal pattern recognition.

Children were measured on these outcomes from Years 1 to 6. Using a regression-discontinuity approach that exploited the discontinuity between the youngest pupils in one year and the oldest pupils in the year below, the researchers were able to identify the overall progress of the children, and the extent to which this was a result of the impact of the school.

The results showed a declining impact of a school year as children got older. The effect size of Year 1 ranged from +1.18 for mental arithmetic to +0.8 for general maths. By Year 6, effect sizes varied from +0.88 for general maths to +0.49 for reading and developed ability.

The effect of schooling itself accounted for an average of between 23.5% and 43.4% of this impact across the four measures. Put another way, the effect size of schooling in Year 1 ranged from +0.55 for reading to +0.31 for developed ability. By Year 6, effect sizes had fallen to between +0.27 for general maths and +0.08 for reading and developed ability.

The researchers suggest that, when setting benchmarks for educational interventions, it is not only important to consider the phase of the educational career, but also the specific measure.

Source: The contribution of schooling to learning gains of pupils in Years 1 to 6 (February 2017), School Effectiveness and School Improvement

Children with reading difficulties may have undiagnosed hearing problems

A report published by the Nuffield Foundation, which compared a group of children with dyslexia to children with a history of repeated ear infections (otitis media with effusion, OME) to see if there were any similarities in their phonological and literacy difficulties, has found that their difficulties may be due in part to undiagnosed hearing problems.

Julia Carroll and Helen Breadmore compared a group of 36 children with dyslexia to 29 children with OME, and also to control groups of typically developing children of the same age and groups of younger children at the same reading level. This made a total sample size of 195 children, ages 8 to 10, from 20 schools in the UK. All of the children completed a series of tests to establish their reading and writing skills and also their phonological skills (ability to manipulate speech sounds) and morphological skills (knowledge of grammatical word structure). Eighteen months later, the children’s reading, spelling and phonological awareness was re-tested, and a hearing screening conducted.

The results showed that the children with dyslexia had different patterns of literacy difficulties than children with OME, although there were some overlaps. The children with dyslexia showed difficulties with both phonological and morphological skills, whereas children with OME had difficulties only on phonological tasks. Both dyslexic children and children with OME had lower levels of reading than the age-matched control children.

The results from the hearing screening eighteen months later found that 9 of the 36 children with dyslexia had mild or moderate hearing impairments, of which their parents and teachers were unaware. The researchers suggest, therefore, that children with reading difficulties should be screened for hearing problems so that they are able to receive more structured support that could help them improve their literacy skills.

Source: Morphological processing in children with phonological difficulties (October 2017), Coventry University and University of Warwick Briefing Paper for The Nuffield Foundation

Evaluation of a parent-delivered early language enrichment programme

A study, published in The Journal of Child Psychology and Psychiatry, evaluates the effectiveness of a parent-delivered language programme on pre-school children’s language and emerging reading skills.

Kelly Burgoyne and colleagues conducted a randomised controlled trial with 208 pre-school children (mean age 3 years, 1 month) and their parents living in socially diverse areas of the UK. Children and parents received either an oral language programme or an active control programme targeting motor skills. Parents delivered the 20-minute sessions to their child at home every day over 30 weeks. Children were assessed at pre-test, post-test, and 6 months after post-test on measures of language and motor skills. Early literacy skills (letter-sound knowledge, phoneme awareness and regular and irregular word reading) were assessed at 6 months after post-test only, as children were non-readers at pre- and post-test.

Children who received the language programme made larger gains in language skills (effect size = +0.21) and narrative skills (effect size = +0.36) at post-test than those children who received the active control programme, and these results were maintained six months later. Improvements were also seen in letter-sound knowledge (effect size = +0.42) and regular word reading (effect size = +0.35). No evidence was found that the control programme improved motor skills.

Source: Evaluation of a parent-delivered early language enrichment programme: evidence from a randomised controlled trial (September 2017), Journal of Child Psychology and Psychiatry. doi:10.1111/jcpp.12819

Evaluation of an early language intervention

A randomised controlled trial, conducted by Silke Fricke and colleagues, looked at the effect of an oral language intervention and compared the extent to which a 30-week programme beginning in nursery and continuing for 20 weeks in Reception was more effective than delivering a 20-week programme starting in Reception.

Children from 34 nurseries in the UK were randomly allocated to a 30-week intervention (n= 132), a 20-week intervention (n=133), or an untreated waiting control group (n=129). Allocation was minimized for gender, age and verbal skills. The children in the 30-week intervention group received the Nuffield Early Language Intervention programme for 10 weeks in nursery and continued for 20 weeks in Reception. The 20-week intervention group received only the final 20 weeks of the intervention, beginning when they entered primary school. The control group received their usual schooling.

Children in both the 20- and 30-week programme intervention groups showed greater improvement in oral language skills on measures including the CELF Expressive Vocabulary and CELF Sentence Structure subtests, and the Information Score from the Renfrew Action Picture Test, compared to children in the control group (effect size for the 20-week programme = +0.21; effect size for the 30-week programme = +0.30). However, there was no evidence to suggest that either programme improved early literacy or reading comprehension skills.

Source: The efficacy of early language intervention in mainstream school settings: a randomized controlled trial (October 2017), Journal of Child Psychology and Psychiatry, 58. doi:10.1111/jcpp.12737

Implication of subject choice on university entry

The introduction of the English Baccalaureate (EBacc) performance measure for schools in England means that schools are incentivised to encourage pupils to study this set of subjects (to count towards the EBacc, a pupil must achieve GCSE grade C or above in a range of subjects including English, maths, history or geography, two sciences, and a language). But are some subjects “better” to study than others for getting to university?

A recent Centre for Longitudinal Studies Working Paper seeks to understand the implications of subject choices at age 14 (when pupils pick their GCSE options), and if these choices then play a part in whether pupils go to university and where they end up studying.

Using data collected from the longitudinal survey, Next Steps, Jake Anders and colleagues looked at the different probabilities of applying to university, entering university, and attending a high-status university for pupils who study the full set of subjects required for EBacc, compared to those who study other combinations of subjects. Using both regression modelling and propensity score matching to test the robustness of the results, they found that pupils who study the full set of EBacc subjects are slightly more likely to apply for and to attend university (a positive effect of 4 and 3 percentage points, respectively). However, the results from the regression model imply that pupils with a full set of EBacc subjects are less likely to get into a high-status university.

The researchers emphasise that the differences are not large, and ultimately it’s far more important to perform well in whatever subject is studied, so the likely implications of more pupils studying EBacc subjects should not be exaggerated.

Source: Incentivising specific combinations of subjects: does it make any difference to university access? (August 2017), CLS working paper 2017/11. Centre for Longitudinal Studies