Sure Start was one of a number of social policy initiatives brought forward by the New Labour government in the late 1990s, which incorporated ideas about the importance of area-based initiatives (in health as well as education),Footnote ¹ of strengthening communities, of tackling social exclusion and child poverty in particular, of making interventions ‘evidence-based’ and of ‘modernising’ public services by encouraging joined-up government and an ‘integrated’ approach that cut across departmental and professional boundaries (Blair, Reference Blair1997). Labour made family policy an explicit part of its legislative programme and devoted particularly large expenditure to subsidising part-time early years education for all three and four-year olds and childcare on both the demand and supply sides. The amount spent on Sure Start was less, but it was a ‘flagship’ policy for the new administration that was also perceived to be popular (Glass, Reference Glass2006; Eisenstadt, Reference Eisenstadt2002).Footnote ²

Sure Start was an early intervention programme, intended to bring together a range of services, including family support, health services and support for special needs as well as childcare and education, in disadvantaged areas. The aim was to ‘invest’Footnote ³ in early childhood and to ‘help to ensure that children, particularly those at risk of social exclusion are ready to learn when they arrive at school’ (HMT, 1998: para. 1.14). Beginning in 1998 with 60 ‘trailblazer’ local programmes (SSLPs) and with a further 250 SSLPs planned, expenditure was more than doubled in 2000 in order to fund 530 programmes by 2004.Footnote ⁴ Labour made a ten-year funding commitment to Sure Start and also funded a major national evaluation project.

It took three years for an SSLP to reach maturity, and the final wave of local programmes was not approved until 2002. Yet in 2003 it was decided that SSLPs would be replaced by Children's Centres, which would cover the whole country, albeit with a limited set of services for better-off areas. By 2006, the majority of Sure Start local programmes were functioning as Children's Centres. Naomi Eisenstadt, the Director of the Sure Start Unit set up in 1998 under the auspices of both the DfES and DH, wrote in 2003 to assure Sure Start local programmes that ‘by embedding SSLPs in the local authority's strategic vision for the delivery of children's services in your area, we will ensure that the additional Children's Centres’ funding will build on what you have already started and improve mainstream services’ (Eisenstadt, Reference Eisenstadt2003). In this interpretation, the move to create a network of Children's Centres represented the ‘mainstreaming’ of Sure Start into a universal service. Indeed, the new Children's Centres were called ‘Sure Start Children's Centres’, with advice offered by the Sure Start Unit (DfES, 2006) on logos and ‘rebranding’, reflecting the importance of ‘market recognition’ in what has always been a mixed economy of early years childcare and education provision. However, some local authorities preferred to refer to the new centres simply as ‘Children's Centres’ from the first,Footnote ⁵ and by the late 2000s this usage was also common in many of the policy documents emanating from central government. Indeed, because Children's Centres were to be a universal service, only a minority originated in SSLPs. Many more were added onto schools and to existing early years provision such as neighbourhood nurseries and early excellence centres, and some to primary health care trust buildings and community centres.

This paper examines the extent to which the transition to Children's Centres represented continuity or change, and suggests that while the policy was framedFootnote ⁶ in terms of continuity, in practice the extent to which changes were made was just as striking. In particular, governance underwent radical change and, while many elements of the Children's Centres’ offer looked similar to those of SSLPs, the emphasis was rather different, particularly in the importance accorded childcare and early years education.

The paper uses historical methods to explore, first, the nature of the shift in policy, drawing on the full range of policy documents from the Department of Education and Skills (from 2007 the Department of Children, Schools and Families), the Treasury, the Sure Start Unit, Parliamentary Papers and the official guidance. The findings of the National Evaluation of Sure Start are used both for the information they provide about the workings of the SSLPs and for the part they played in influencing policy. Second, the paper explores what these sources reveal as to reasons for the policy shift, in terms of changes in the government's own agenda and external pressures. This part of the paper also draws on key informant interviews, which provided very different perspectives on the key reasons for the promotion of Children's Centres,Footnote ⁷ and documentary evidence from two London boroughs. Finally, the paper reflects on the implications for a broader understanding of policy change, drawing attention to the importance of the role of strategic framing and of the politics of policy-making in particular.

Sure Start local programmes

Provision for young children under four was made the subject of one of New Labour's cross-departmental reviews when it came to power in 1997. Officials from 11 government departments considered academic papers suggesting that evidence from the UK and the US showed that early intervention programmes were important for the development and life experience of children (HMT, 1998; see also Waldfogel, Reference Waldfogel1999; Melhuish, Reference Melhuish2004). The aim of the programme that emerged from the Review process – Sure Start – was intended to improve children's health and their social, emotional and cognitive development, and to strengthen families and communities in disadvantaged areas. School readiness was thus defined more widely than cognitive development, as was also the case with the American Head Start programme, which was begun in 1965 as both a child and community development programme in the context of the wider War on Poverty (Zigler and Styfco, Reference Zigler and Styfco2004). The policy problem was identified by British policy-makers mainly in terms of family functioning and child poverty among the socially excluded, and the thinking behind the creation of Sure Start focused on finding a more integrated approach to tackling social exclusion among young children and their families (HMT, 1998; Glass, Reference Glass1999). There were large hopes for Sure Start, which was seen as something of a ‘magic bullet’ that would, in the long term, help to reduce youth crime, teenage pregnancy, family breakdown and poverty. The introductory booklet accompanying the setting up of the fifth wave of SSLPs in 2002 talked of ‘breaking the cycle of disadvantage’ (DfES, 2001, Foreword). However, Professor Sir Michael Rutter, adviser to the national evaluation project, stated that, in his view, the ‘Government statements on what it wanted SSLPs to achieve have lacked both consistency and precision’ (Rutter, Reference Rutter, Belsky, Barnes and Melhuish2007: 199).

The Sure Start local programmes launched in 1998 offered outreach and home visits; support for families and parents (including support groups, drop-in and parenting sessions); support for good-quality play, learning and childcare; primary and community healthcare and advice about child and family health; and support for children with special needs. These constituted the core offer, but SSLPs could also provide additional services: for example, advice on housing or welfare benefits. The local programmes could offer a particular service themselves, or give money to a provider to fund more of a particular service (for example, an extra health visitor), or provide information to parents as to where to seek help. As the National Evaluation showed, SSLPs were able to call on different amounts of existing provision, and while most ‘inherited’ reasonable health provision, there was usually little by way of childcare (Allnock et al., Reference Allnock, Tunstill, Meadows, Akhurst, Chrysanthou, Garbers and Morley2005). Each SSLP enjoyed autonomy and, while it was obliged to provide the core offer and to work towards nationally determined targets, how it did so was up to each local programme to decide.

A set of key principles was laid down for SSLPs in the first official Guidance offered by government: to coordinate, streamline and add value to existing services (which required staff to work in cooperative and new ways); to involve parents; to avoid stigma; to ensure lasting support for children and families; to act in culturally appropriate and sensitive ways; to achieve specific objectives; and to promote accessibility for all local families (DfEE, 1999a). Naomi Eisenstadt (Reference Eisenstadt2002: 3, 4) commented that ‘probably’ the most important of these was the injunction to involve parents and carers: ‘The fundamental premise is that better outcomes can only be achieved with the active participation of parents.’ The National Evaluation found that parents were represented on the partnership boards of all the 12 SSLPs studied in depth (Ball, Reference Ball2002). Parents were also to be the subjects of many dimensions of the intervention. Indeed, only provision for special needs and for early years educationFootnote ⁸ focused on the child alone; other parts of the SSLPs’ core offer were usually focused on the parent and child, or the parent alone (Tunstill et al., Reference Tunstill, Allnock, Akhurst and Garbers2005a). Many SSLPs expected parents to be present while their children were attending a childcare session at the SSLP in order to learn how to interact with their child (Anning et al., Reference Anning, Chesworth and Spurling2005).Footnote ⁹ Engaging local communities and local parents was central to the ethos of Sure Start (although, turning this on its head, Clark (Reference Clark2006) has argued that Sure Start was as much about the promotion of good parenting and the effective surveillance of mothers in particular (see also Hey and Bradford, Reference Hey and Bradford2006)).

Certainly, the national targets for Sure Start reflected the focus on parents (invariably mothers) as much as children. In common with all public services, Public Service Agreements (PSAs) were set by the Treasury for Sure Start. Thus, under the health objective, targets were set for reducing the proportion of mothers who smoked, for increasing the proportion who breastfed and for reducing the proportion of children admitted to hospital. The PSAs for children's cognitive, social and emotional development proved both hard to measure and difficult to achieve. In practice, as Ellison et al. (nd, circa Reference Ellison, Hicks and Latham2004–5: 23, 15) commented, there was an ‘inevitable artificiality about the way activities [in SSLPs] are allocated to PSA targets’ not least because ‘the targets themselves were not necessarily what Sure Start has actually been trying to achieve’. Many SSLPs made substantial efforts to increase parental self-esteem and confidence, which were also hard to measure and did not in any case figure in the PSA targets. Sure Start delivered a range of services, depending on local demand and assessment of need. But some targets, for instance achieving a reduction in mothers who smoked, were not necessarily supported by the local community (Fox, Reference Fox, Weinberger, Pickstone and Hannon2005). As one London voluntary agency contracted to provide services put it: ‘Sure Start is a very target driven process, but parents are told that it is their wishes that are paramount. Often they cannot reconcile these two concepts’ (sic) (Westminster Children's Society, 2003: 2).

The PSAs were frequently reviewed and the emphasis shifted substantially over time. The 2002 Spending Review specifically changed the aims of Sure Start to include an increase in the availability of childcare as a separate aim and, by 2004, the provision of childcare was tied firmly to the target of getting parents into employment. But SSLPs did not develop into major childcare providers and in fact the Neighbourhood Nursery Initiative in disadvantaged areas complemented Sure Start in this respect.Footnote ¹⁰ Norman Glass, who as a civil servant in the Treasury had been involved in the setting up of Sure Start, maintained that supporting parents into work was ‘an ambition for which it arguably was not funded, let alone intended’ (Glass, Reference Glass2006: 55). Because of its wide aims, it was possible for Sure Start to mean all things to all people, but the ethos of Sure Start local programmes was centred on local responsiveness, parent participation and child development. There was no specific focus on early years education and care linked to the provision of opportunities for parental employment.

Children's Centres

The core offer for the Children's Centres launched in 2003 was: integrated early education and care, parental outreach, family support, health services, information about other childcare providers and community services, effective links with Jobcentre Plus and workforce training. Children's Centres were to be a universal service, but in the 70 per cent of areas that were better off, the intention was to focus mainly on the provision of an information service. The greatest emphasis was put on the need to provide integrated childcare and early years education, and to reach the most disadvantaged, hard-to-reach parents – both aims being strongly associated with the concern to promote adult employment and to address the issue of workless families, which are disproportionately numerous in the UK compared to other EU Member States (NAO, 2007). The first major report of quantitative findings from the National Evaluation had identified this group as being particularly hard to reach by SSLPs (NESS, 2005). Increasing concern about the evidence-base for programmes delivered by SSLPs, also raised by the National Evaluation, resulted in an effort to promote more standardised family support programmes from the list issued by the National Academy of Parenting Practitioners, established in 2007.

The principles underpinning Children's Centres were said to be the same as SSLPs; Margaret Hodge, the Minister responsible, told the Select Committee on Education and Skills that Children's Centres would be ‘community driven, parent driven and will retain that essence of Sure Start which has made it so successful’ (House of Commons Education and Skills Committee, 2005, Q. 479). In 2006, Beverley Hughes, Minister for Children and Families, described the Sure Start local programmes to this Committee in terms of an ‘experimental stage’, which had let a ‘thousand flowers bloom’ (House of Commons Education and Skills Committee, 2006, Q. 7), laying the groundwork for the new universal, mainstream service. These statements were part of a series of ex post facto constructions of the shift from SSLPs to Children's Centres in terms of continuity rather than change. Yet in the new emphasis on childcare and education and on the nature of governance, change was more apparent than continuity.

Children's Centres were first mooted in the 2002 Report of the Inter-Departmental Committee on Childcare as a particularly effective way of providing good-quality, integrated childcare and early years education (DfES, DWP, HMT and Women and Equality Unit, 2002, chapter 4). A major focus of this report was on ways of meeting the government's objective of getting 70 per cent of lone mothers into work, which in turn was seen as part of what was needed to halve child poverty by 2010. The much greater place given to childcare provision in the core offer of Children's Centres from their launch in 2003 was reflected in the changes in governance at central government level. In 2002, cross-departmental responsibility for Sure Start moved from the DfES and the DH to the DfES and the DWP. The Sure Start Unit was merged with the Early Years Division and Childcare Unit, creating a single group responsible for SSLPS, neighbourhood nurseries and the National Childcare Strategy, underlining the stronger emphasis on childcare and its connection to supporting parents into work. Children's Centres were set targets for the creation of new childcare places and, from September 2008, centres were obliged to provide fully integrated early years education and full daycare, which the Effective Provision of Pre-School Education (EPPE) project had found to be particularly beneficial for child outcomes (Sylva et al., Reference Sylva, Melhuish, Sammons, Siraj-Blatchford and Taggart2004).

In respect of local governance, Children's Centres were made the responsibility of local authorities, which managed the allocation of the government grant (SSLPs had been funded directly by the DfES, with the lead partner on the partnership board managing the grant). In part, this change also reflected the focus on childcare and education. The 2002 Inter-Departmental Review of childcare pointed out that the Early Years Development and Childcare Partnerships, which had been given the responsibility of managing the childcare market at the local level in 1998 (Lewis, Reference Lewis2003; Penn and Randall, Reference Penn and Randall2005), were more successful when the local authority was actively involved (DfES, DWP, HMT, Women and Equality Unit, 2002), and the 2006 Childcare Act gave local authorities a statutory responsibility to ensure that there were enough childcare places for parents in work or training. In part, the change was also linked to the profound restructuring of children's services undertaken under the auspices of the Every Child Matters agenda (HMT, 2003), launched in 2003, for it would have proved difficult to leave SSLPs outside the reconfigured services for children at the local level.

The first, brief, official guidance for Children's Centres issued in 2003 reiterated the need to involve parents, but by consultation above all, and to develop an understanding of community development rather than to encourage active participation, as had been the intention with SSLPs.Footnote ¹¹ While SSLPs were never ‘bottom-up’, in so far as parents were very much a minority voice on the partnership boards responsible for their governance, responsiveness to local needs and preferences, community development and empowerment were explicit aims of the programme (Williams and Churchill, Reference Williams and Churchill2006) in a way that they were not for Children's Centres.

Children's Centres were made part of ‘mainstream’ provision for young children in all local authorities, but it is less certain as to whether in so doing a Sure Start approach was ‘mainstreamed’. The local responsiveness that had let a thousand flowers bloom in terms of the nature and balance of the SSLPs’ service offer gave way to a more specified service offer with an emphasis on integrated childcare and education for children in conjunction with links to Jobcentre Plus for their parents, and greater control by the local authority to make sure it happened. A similar trajectory from a community-oriented to a more centrally mandated programme marked the experience of Head Start in the US (Levitan, Reference Levitan1969; Cohen, Reference Cohen1996). Newly created Children's Centres were often attached to schools (as had been the EECs),Footnote ¹² which reflected the importance attached to early years education and care in the Children's Centres’ offer and to the new requirements under the 2006 Childcare Act for local authorities to ensure an adequate number of places.

However, ‘mainstreaming’ Sure Start into a universal service had significant cost implications. On average, SSLPs in their third year of operation spent £900 per child (at 1999–2000 prices), with a minimum spend of £350 and a maximum of almost £2,500 (Meadows, Reference Meadows2006). The funding for Children's Centres was less than that for SSLPs. The Minister responsible, Beverly Hughes, told the 2006 Select Committee on Education and Skills that it was close to 66–70 per cent of the level reached by SSLPs (House of Commons Education and Skills Committee, 2006, Q.29). The Public Accounts Committee expressed fears about the dilution of support for disadvantaged areas (House of Commons Committee on Public Accounts, 2007), and the National Audit Office (2006) expressed concern about the sustainability of Children's Centres. At the local level there was also concern about ‘who will pick up the tab on the new services once central government money dwindles’ (City of Westminster, 2002: 2).

The reasons for the policy shift

The reasons for the shift to Children's Centres are related in part to changes in government's policy agenda in respect of the service offer made by SSLPs and of services for young children more generally, and in part to evidence of programme failure offered by the National Evaluation.

First, Sure Start's ethos and practice became increasingly out of step with government policy in regard to childcare. The emphasis on making provision for integrated childcare and education, linked to supporting parents into work, was very different from the relatively minor part accorded childcare in Sure Start local programmes, and reflected the way in which government's thinking on childcare in relation to social exclusion had developed. SSLPs had often provided some childcare – in the form of informal drop-in and ‘stay and play’ provision as well as crèches and sessional day care – mainly in order to encourage parents to take part in other Sure Start activities, but their provision of formal childcare for working parents remained low (Tunstill et al., Reference Tunstill, Allnock, Akhurst and Garbers2005a). The aim was thus to support parents seeking to attend courses that might improve their employability or to enhance the parent/child relationship. Much less attention was paid to the cognitive development of children, especially in the crèches (Anning et al., Reference Anning, Chesworth and Spurling2005; see also Mathers et al., Reference Mathers, Sylva and Joshi2007). In their report for the National Evaluation on the employability of parents in 25 SSLPs, Meadows and Garbers (Reference Meadows and Garbers2004) concluded that SSLPs acted mainly as a bridge for parents into the education, training and employment services of other organisations, not least because of the common local perception that the mothers of young children should stay at home. No SSLP in their sample told parents that they should be working.

But in 2003, the Work and Pensions Select Committee warned that the number of childcare places was not likely to be enough to allow the 2010 targets on child poverty and employment for lone mothers to be met (House of Commons Work and Pensions Committee, 2003: 15). Government plans for the development of childcare over the period 2004–2014 (HMT, DfES, DWP and DTI, 2004) made specific reference to the increased contribution that would be made by Children's Centres. At the same time, government was pushing ahead with more out-of-school care for older children: by 2010, all schools would be obliged to be open from 8 a.m. to 6 p.m. – the better to serve the needs of working parents. Some local authorities made an immediate connection between the need to provide ‘extended schools’ and new Children's Centres (for example, City of Westminster, 2005: 4–5), the majority of which have been located in schools rather than other community-based institutions. This in turn tended to give more emphasis to educational attainment than to the community-based initiatives and community-led services promoted by SSLPs, which were more likely to have been based in community institutions other than schools.

Second, the government's agenda for services for young children more generally underwent substantial change. Children's services underwent fundamental review as a result of the Every Child Matters agenda, implemented in the wake of the inquiry into child protection. Stress was placed on the importance of prevention, together with a strong emphasis on children's educational achievement. Integrated servicesFootnote ¹³ for children were to be provided under Directors of Children's Services appointed by local authorities, and it would have been impractical to leave out the SSLPS, which were accountable directly to central government in the form of the Sure Start Unit.

The new vision of a universal network of integrated children's services at the local level also meant the end of Sure Start as an area-based programme. Both the National Evaluation and the Daycare Trust (a long-standing campaigning group for formal childcare provision) had made the point that not all disadvantaged children lived in deprived areas (Tunstill et al., Reference Tunstill, Meadows, Allnock, Akhurst, Chrysanthou, Garbers, Morley and van de Velde2005b; Land, Reference Land2002). The provision of Mini-Sure Start local programmes in rural areas and in small pockets of urban deprivation from 2002 was designed to address this, but pressure for more universal coverage remained. In 2003, the Work and Pensions Select Committee argued in favour of a universal Children's Centres model (House of Commons Work and Pensions Committee, 2003: 22). By the mid-2000s, Labour politicians were speaking of the importance of ‘progressive universalism’ (for example, Balls, Reference Balls2005), which in the case of services meant universal provision, with the greatest help for those in the most need. This principle was articulated clearly in the 2005 pre-budget report (HMT and DfES, 2005), and in the 2007 Children's Plan, which promoted the idea of universal services operating in a preventative system (DCSF, 2007).

Third, strong evidence of programme failure was offered by the National Evaluation, which was all the more serious given the premium the government attached to evidence-based policy as a means to its pragmatic commitment to identifying and funding ‘what works’.

The emphasis on area-based programmes in the policy fields of housing, health and education, as well as early years, faded out of the policy picture, largely because of the tension between central direction and local empowerment (see Coote et al., Reference Coote, Allen and Woodhead2004). As Naomi Eistenstadt commented: ‘The key debate, which still rages about Sure Start, was the commitment to local flexibility; [and] funding based on plans to deliver specific child outcomes rather than specific services’ (Eisenstadt, Reference Eisenstadt, Belsky, Barnes and Melhuish2007: ix). The idea of local flexibility incorporated a commitment to community ‘empowerment’ and to local responsiveness, which sometimes made the pursuit of PSA targets (which, pace Eisenstadt, covered mothers’ behaviour as well as children's welfare and development) difficult. Despite central control, SSLPs had enjoyed autonomy in the way in which they went about their work, which resulted in tensions regarding the effectiveness of the programmes they chose to use and how far their work was oriented towards achieving better outcomes as opposed to community engagement. It was difficult both to work towards PSAs and to put responsiveness to parent-determined needs first.

The 1999 Guide to Evidence-based Practice said that Sure Start had been ‘planned on the basis of sound evidence from international research’ (DfEE, 1999b: 3), but, in their volume summarising the findings of the National Evaluation, Melhuish and Hall (Reference Melhuish, Hall, Belsky, Barnes and Melhuish2007) claimed that while research evidence was critical in winning the argument for Sure Start at the beginning, it was overlooked in the detailed planning and operation of the local programmes. Sure Start had no in-built plan for evaluation; each local programme was different in terms of the nature, volume and mix of services it offered. Nor did the initiative require the use of standardised programmes, for example to encourage good parenting. All this made it very difficult to assess the outcomes (Kane, Reference Kane2008; see also Hansen and Hawkes, Reference Hansen and Hawkes2009), which is what the quantitative part of the National Evaluation hoped to do and which was particularly important for a government focused on outcomes rather than process.

It was additionally difficult to come to any quick conclusions about outcomes associated with SSLPs. Not only did the local programmes usually take three years to become fully operational, but the hope was that the effects of the programmes would be revealed in terms of school readiness, and also later in the child's life: for example, in regard to increasing rates of educational achievement and decreasing rates of juvenile crime. However, it was expected by some civil servants and politicians that some of the PSA targets set for Sure Start would be achieved in two or three years. But as Kane (Reference Kane2008)Footnote ¹⁴ argued, it was too early and too difficult to assess the impact of Sure Start.

The National Evaluation, which started 18 months into the second wave of SSLPs (in 2000), published its interim findings in 2005 and they were not encouraging. In particular, the National Evaluation found only modest benefits for the moderately disadvantaged and small adverse effects for those most at risk – the very people policy-makers were most concerned to reach (NESS, 2005). A similar conclusion regarding the effects on the most disadvantaged was reached in the US in respect of Early Head Start (Love et al., Reference Love, Kisker, Ross, Schochet, Brooks-Gunn, Paulsell, Boller, Constantine, Vogel, Fuligni and Brady-Smith2002), which served the under threes.

Glass (Reference Glass2006) has suggested that government moved to change Sure Start in anticipation of these findings. Press reactions to the national evaluation's interim findings, which were leaked prior to publication, were strong, with a majority claiming that Sure Start had ‘failed’ (Phillips, Reference Phillips2005; Marrin, Reference Marrin2005; Toynbee, Reference Toynbee2005; Ward, Reference Ward2005; Wheatcroft, Reference Wheatcroft2005). When the Minister responsible, Margaret Hodge, appeared before the Education and Skills Committee in 2005 and said that Sure Start Children's Centres would ‘retain that essence of Sure Start which has made it so successful’ (see above), the Chairman replied that two thirds of SSLPs had been shown to be unsuccessful (House of Commons Education and Skills Committee, 2005, Q.480). A supplementary memorandum submitted by Hodge after her appearance before the Committee suggested that no evidence of impact should not be interpreted as evidence of no impact (House of Commons Education and Skills Committee, 2005: 191; see also Glass, Reference Glass2006). In 2006, as the first substantial guidance was being issued to Children's Centres, Tony Blair, then Prime Minister, talked about the importance of reaching the hard-to-reach families, who had remained excluded (Blair, Reference Blair2006). The primary policy goal in setting up Sure Start had to do with tackling social exclusion, but the interim findings from the National Evaluation seemed to indicate that it had not succeeded in respect of the most disadvantaged families.

Professor Sir Michael Rutter, writing in a personal capacity rather than as an adviser to the National Evaluation, criticised policy-makers for rejecting both pilot projects and a random controlled trial (RCT) when setting up Sure Start, which he viewed as crucial to the provision of a robust evidence base: ‘Cynics might also argue that an RCT had to be ruled out because, if it did not show that it was effective, it carried the danger of showing that a key Government policy was a mistake’ (Rutter, Reference Rutter2006: 136). He suggested that, in the end, evidence-based policy was trumped by political pressure (Rutter, Reference Rutter, Belsky, Barnes and Melhuish2007), and that the mainstreaming of Sure Start was happening without demonstration of its effectiveness. As Tunstill et al. (Reference Tunstill, Allnock, Akhurst and Garbers2005a: 168), who were involved in the qualitative research on process for the National Evaluation, put it: ‘At face value it may be argued that in this case national policy has evolved in advance of conclusive evaluation findings.’

However, it was doubtful that findings from the evaluation could ever be conclusive, given the problems the evaluators faced in establishing causal relationships for a programme that varied so much from place to place. Indeed, in 2008, new quantitative findings from the National Evaluation reversed the findings in respect of the most disadvantaged families: now no adverse effects were observed (Melhuish et al., Reference Melhuish, Belsky, Leyland and Barnes2008; NESS, 2008). The evaluators concluded, albeit cautiously, that this could be attributed to the ‘maturing’ of SSLPs as they became Children's Centres with more ‘clearly focussed services with better guidance’ and claimed that the improvement was ‘partly influenced by the early NESS findings’ (NESS, 2008: 30). They gave the example of the successful use of the same standardised parenting programme in 11 SSLPs in north and mid-Wales (Hutchings et al., Reference Hutchings, Bywater, Daley, Gardner, Whitaker, Jones, Eames and Edwards2007).Footnote ¹⁵

But the idea of politics trumping the best efforts of social scientists was only part of the story. The evaluators may well have done the best they could in the absence of rigorous implementation (Rutter, Reference Rutter, Belsky, Barnes and Melhuish2007); however, the findings they published in 2005 in particular made them policy actors. These findings may well have influenced the guidance issued in respect of Children's Centres, but in the first instance they made it difficult to defend SSLPs. The policy shift to Children's Centres was in part a political reaction to the 2005 findings.Footnote ¹⁶ Notwithstanding the shortcomings of the Evaluation's methodology, the findings made it possible to charge Sure Start with ‘failure’.

By the time more positive findings were issued in 2008, SSLPs had been ‘mainstreamed’ according to policy-makers, or, more plausibly in the light of the extent of the changes to the programme, ‘absorbed’ into Children's Centres, which in turn had been brought into the mainstream policy development of children's services and childcare and education. The guidance for Children's Centres highlighted the importance of increasing childcare places and of reaching the most disadvantaged families, the first of these being in line with government's own rapidly expanding policy agenda for childcare and early years education, and the second acknowledging the question mark the 2005 National Evaluation posed over the SSLPs' success in tackling social exclusion.Footnote ¹⁷

Changes in central government's policy agenda for Sure Start and for policies for young children more generally were important in explaining the shift to Children's Centres, but it is also the case that government was ‘hoist on its own petard’ of promoting evidence-based policies when the National Evaluation cast doubt over Sure Start's achievements. The government's commitment to rapid and transformative policy change was difficult to maintain in parallel with a commitment to evidence-based policy development, which is likely to be much slower, requiring the use of pilot projects and/or RCTs, which are in turn likely to be difficult to defend politically in a highly centralised state where differences in the provision of services (for example, in respect of NHS drug policies) are met with public disapproval.

Conclusion: understanding the policy change

The transition from SSLPs to Children's Centres constituted a substantial change in policy, which was nevertheless presented by politicians as continuity rather than change. The overarching policy goal of tackling social exclusion remained, but government's ideas as to how best to achieve it shifted, resulting in changes in the programme's content, governance and reach.

In respect of programme content, a focus on support for children and their parents gave way to an emphasis on children's cognitive development on the one hand, and parents’ employment on the other. In addition, the mainstreaming of Children's Centres as a new, universal local service under the control of local government marked a major change from the centralised control of SSLPs, which had nevertheless (somewhat paradoxically) enjoyed substantial autonomy in regard to the service offer they made. While statements as to government's commitment to ‘empowering’ families continued to be made (for example, Cabinet Office and DCSF, 2008), parental participation, together with responsiveness to local demand and local need, gave way to a more centrally specified service offer. The PSAs for Children's Centres were more closely aligned with the core offer that emphasised early years education and childcare, and links with Jobcentre Plus.

However, representing Children's Centres as a continuation of the ‘essence’ of Sure Start allowed government to claim the success of a flagship programme, while also changing it significantly to meet both its changing ideas as to how best to meet its core goal of tackling social exclusion, as well as academic and press criticism about the failure to reach the most disadvantaged. Taken together these points explain the policy shift, with the negative findings of National Evaluation providing the catalyst for a substantial policy change that was presented as the development of Sure Start.

The effort strategically to frame the policy (Campbell, Reference Campbell1998) in terms of continuity while making important changes enabled government to hang on to a flagship policy in which considerable investment of public money had been made and to avoid both blame for policy failure in the wake of the National Evaluation's interim findings published in 2005 and a possible backlash by voters angry at the abandonment of a popular programme. Paul Pierson's (Reference Pierson1994) work on policy change in a period of welfare state restructuring has shown the political importance of ‘blame avoidance’. Thus, a government committed to transformative policy change nevertheless sought to claim policy continuity. Strategic framing to justify reform and as an aid to the pragmatic politics of policy-making is the most important factor in understanding how politicians changed course on Sure Start.

The idea of framing the policy shift as the ‘mainstreaming’ of Sure Start, while nevertheless making significant changes, was a way of satisfying both critics of Sure Start and its many committed defenders. In 2008, the government announced its wish to put Sure Start Children's Centres ‘on a legislative basis’ to ensure that they become ‘an established part of the universal services available for young children and their families’ (DCSF, 2008a: 1). This will make governance of Children's Centres much clearer. In addition, having committed large amounts of public expenditure first to SSLPs and then to Children's Centres, this move may be interpreted as a desire institutionally to embed this early years’ initiative and to protect it from possible future political attack.