We all love Sure Start.
All the party leaders have praised it. Wonks and academics love the idea. It is a rare example of "early intervention" – an attempt to move from curing problems to preventing them emerging in the first place.
Rolled out since 1998, it is seen as one of the most enlightened programmes of the last government. Sure Start offers parenting advice, childcare, and health advisers. It started operating in the most deprived areas first.
There's evidence from academics that you get the biggest returns to investment in education are when children are very young. A programme like Sure Start wouldn't have to improve educational achievement very much to make it a worthwhile investment either: Research based on cohort studies finds that the lifetime earnings impact for an individual of gaining five GCSEs at grades A to C compared with not is between £80,000 and £100,000. And graduates earn far more than non-graduates.
Furthermore, there are plenty of anecdotes about the amazing long term effects of similar-sounding programmes – like the Perry preschool project in the US. In fact Sure Start is heavily based on a previous US initiative called Head Start.
It sounds like it must work, and must be a good investment. But is it?
The programme has changed quite a bit over the years. At first it was run as a series of local programmes, which varied a lot. The first 524 local programmes were established between 1999 and 2003. After 2005 the model was standardised: local authorities would provide children's centres in which all these different services were brought together seamlessly under one roof.
On average, Sure Start local programmes (SSLPs) cost around £1,300 per eligible child per year at 2009-10 prices (or £4,860 per eligible child over the period from birth to the age of four).
That's a lot of money. But part of that cost is offset, because by providing childcare, the programmes helped some parents move into work. By the time children reached the age of five, SSLPs had already delivered economic benefits of between £279 and £557 per eligible child.
But to be cost effective, the local progammes would have to be delivering some significant educational improvement.
The government didn't set up a proper randomised control trial, so evaluation is quite difficult. Fortunately there are some measures taken on a uniform basis for children in the programme, and those who are not. The Foundation Stage Profile measures children's cognitive development at the start of school.
The Department for Education evaluation published at the end of last year says that:
"no differences emerged between the [Surestart and non Surestart] groups on 7 measures of cognitive and social development from the Foundation Stage Profile completed by teachers".
It notes that:
"It is disappointing that no effects were discerned for “school readiness” as measured by the Foundation Stage Profile."
On a more positive note, the study suggests that there were improvements in the "home learning environment" and that parents involved in the scheme used less harsh discipline on their children. But these effects may be to do with the measurement differences between the Sure Start group and the control group.
Some people might hope that the programme became more effective after 2005. It may be too early to tell. A commons committee on the subject criticised the lack of data collected on the performance of the scheme.
However, one study from academics at Durham University published at the end of last year covers the period up to 2008. It suggests that children's school readiness hasn't improved: in fact it has got slightly worse in reading and slightly better in maths – but neither are really significant. That's not conclusive proof that the scheme isn't working. But it isn't a good sign.
The US Head Start progamme on which Sure Start is based has been running since the sixties. When it was reauthorised in 1998, Congress mandated the government to carry out a proper randomised control trial. The first results were published at the start of last year.
While children seemed to be doing a little better during their time on the programme, the effects were so small that they had faded away by the time the children reach the end of their first grade in school:
"No significant impacts were found for math skills, pre-writing, children’s promotion, or teacher report of children’s school accomplishments or abilities in any year."
"the benefits of access to Head Start at age four are largely absent by 1st grade for the program population as a whole. For 3-year-olds, there are few sustained benefits, although access to the program may lead to improved parent-child relationships through 1st grade."
What does this mean?
There are lots of different responses.
So far the political response is to try and improve the programme. A report by the Children, Schools and Family Committee of the House of Commons published before the last election suggested that Sure Start centres should: focus on their core business rather than trying to do everything; try to improve the quality of their staff; and get better at reaching the most troubled families who don't use the service.
The new Government is also trying to refocus the programme on its original goal of serving deprived communities – rather than trying to spread the service, and the budget, across the whole country.
But aren't there wider implications?
I think the central importance of the early years is overdone. John T Bruer's book "The Myth of the First Three Years" reviews the neuroscience on which this now-common idea is based, and shows how it has been whipped up into a kind of early years determinism. Fortunately, he suggests, our course is not actually set by age three – our brains are much more plastic than that.
If early intervention programmes are not going to be the "wonder drugs" that they have sometimes been hyped up to be, where does that leave us?
Both left and right have responses. The left traditionally argues that centrist technocratic fixes like Sure Start won't work, because they don't fix the underlying inequalities that cause social problems. In mirror image, the right argues that they won't work because they don't address the underlying cultural issues that cause social problems. These arguments are strengthened if programmes like Sure Start don't work.
I do think culture and the wider context are important. Recent academic work on children from poor backgrounds who go on to perform well - confirms the importance of culture and parenting. These things are difficult for the government to change with micro-interventions like Sure Start and Head Start. And they are profoundly affected by wider welfare policy.
But I don't think we should totally give up on early education interventions. Part of the problem is that the effect of such schemes fades, unless it is followed up better education in later years. That's why the recent expansion of school reforms to include primary schools is so important.
In a way, Sure Start and Head Start are not that ambitious. Perhaps we shouldn't be surprised that better childcare and health advice have not had the revolutionary consequences that some had imagined. But as we learn more about how small children develop, it is possible that people will come up with more radical and more effective programmes in the future.