If school meals are free, how do we know who needs help?

Do I get a pupil premium with that? Chris Radburn/PA Archive

As September approaches, primary schools up and down the country are preparing to roll out universal free school lunches for all children in reception, year 1 and year 2. But if all meals are free, what does this mean for the government’s flagship measure of deciding how to give extra money to schools to help children from disadvantaged backgrounds?

The school system uses a child’s eligibility for free school meals as an indicator of their socio-economic disadvantage. Families with parents in receipt of some benefits, including income support and income-based Job Seekers’ Allowance, receive free lunches. At the moment, about one in five pupils is eligible. The Department for Education also publishes data on the achievement of such children to gauge how poor children progress in the education system.

But what does eligibility for free school meals signify? Children living in lone-parent families or with parents on benefits constitute most of those eligible and it is a reasonably good indicator of low household income. Students eligible for free meals have, on average, much lower educational achievement. While 65% of pupils not eligible for free school meals achieve 5A*-C GCSEs including mathematics and English, only 38% of pupils who are eligible, actually do so.

Far from perfect

The government uses the free schools meals measure for pragmatic reasons: schools can check parent eligibility through a centralised computer system and most parents are keen to apply to get free meals for their child. However, the measure is far from perfect.

Some children in “working poor” households are not identified, nor are children whose families choose not to claim a free meal for dietary, cultural or other reasons. Eligibility is also cyclical. In an economic downturn the proportion of students eligible rises.

Ideally we want a measure that indicates a child’s persistent educational disadvantage, such as low parental education. Temporary unemployment of a parent can trigger eligibility for free school meals, but may not signify that a child is educationally disadvantaged. Conversely, in an upturn, the proportion of children eligible falls. These children may still suffer long-term educational disadvantage associated with low parental education and insecure household income, despite a temporary upturn in their economic circumstances as their parent secures a job.

Despite these limitations, eligibility for free school meals is used to identify pupils who attract the pupil premium, an additional payment paid to schools for disadvantaged students. Since 2011, all pupils who have ever been eligible for free school meals during the previous two years are entitled to this premium of around £1,000.

This provides an incentive for schools to identify children who are eligible, although schools have found it difficult to identify all the eligible students. Around one in ten students eligible for free school meals are still not claiming them. On the other hand, children who are only temporarily eligible for free school meals will attract the pupil premium for at least two years – and these pupils may not be the most needy if they don’t suffer from long-term disadvantage.

Even with this relatively simple system in place, identifying poor students is challenging and about to get more so. Pupils in infant schools will shortly receive free meals at school regardless of their home circumstances.

Although schools will still record pupils who meet the existing criteria for free meals – and will want to in order to get the pupil premium – the incentive for parents to register is lessened given that these pupils will receive food at school regardless.

There will be a new indicator in the schools census database to separately identify pupils who are eligible for free meals in the conventional sense. We will therefore continue to be able to monitor the progress made by free school meal students and to allocate the pupil premium on that basis. However, there is a risk that in the early years, a critical period in children’s development, many parents will no longer bother to register their child as eligible.

An added complication will also be the introduction of universal credit, set to be implemented nationwide by 2016. Government data indicates that if all school-aged children’s eligibility for free meals was extended to everybody on universal credit, the numbers of children receiving free meals would triple and the costs would be prohibitive. Therefore within the system of universal credit, eligibility for free school meals will still need to be determined separately.

Other options on the menu

Are there alternatives? Some researchers have recommended using census data to identify schools with socio-economically deprived pupils. Schools could also use such data to allocate the pupil premium. Census data would provide information about the level of deprivation of the neighbourhoods in which children live. Such measures would include employment rates and the proportion of people in a neighbourhood who have low education. In our work, we have found such neighbourhood-based measures do indeed predict pupil attainment.

Census measures are in some ways less crude than the free school meals indicator. For example, census data could distinguish between two schools that have the same rate of children on free school meals, but that are in different situations.

In one school the non-eligible pupils might all live in very deprived neighbourhoods and, in the other, most might live in wealthier neighbourhoods. The downside is that the census data can only provide information on pupils’ neighbourhoods rather than on individual circumstances. We will still need some kind of indicator to identify individual pupils.

One solution would be to use administrative data on parents’ receipt of benefit. To do this we would need their National Insurance number and to overcome any legal restrictions on the use of these data for this purpose. The advantage of this approach would be that it would provide an accurate record but it may be politically difficult.

An alternative would be to use students’ low prior achievement as an indicator that they are at risk of educational disadvantage. This would work well in secondary schools but is far more problematic in primary schools. Very early measures, such as the assessments of children undertaken by teachers in reception, are often fraught with error.

A final alternative is for schools to collect additional data from parents, such as their education level. This might be more difficult to extract from parents. A trial to determine whether these approaches are feasible is arguably now pressing, given the forthcoming changes to the free school meal system.