ChildForum Office of Pre-Primary EducationThe lead public advisor on early childcare and education

National organisation for the ECE sector

See: About Us


Assessing the Evidence on Early Childhood Education Effects on Children: A Systematic Review

lunch-time at a Kita in GermanyBy Dr Sarah Alexander
© 2008



The methodologically soundest (best) research over the past 20 years on the effects of participation in early childhood education and care was selected and assessed.

Studies were selected using a set of strict criteria, including that they:

  • followed children from a young age (as opposed to sampling children when they have left or are due to leave to go to school), and
  • were not reports of high cost intervention overseas programmes for different ethnic groups impossible to duplicate here in NZ for the same outcomes.

As the review was a systematic and critical one, and independently conducted, it did not suffer from biases often inherent in literature reviews and publicly funded reviews.  

The findings of this critical and systematic review may be controversial because the findings differ from political claims about what research says. 

This paper presents the major pieces of research and the strongest evidence available on what the positive and negatives effects of childcare are, and discusses the evidence in the light of current policy directions in New Zealand.


Key Findings

1.  The most robust research points to parents/family having greater impact than the childcare/ECE experience on children’s developmental outcomes. This is no surprise and confirms what early childhood professionals generally know and acknowledge. Therefore, effort put towards supporting families to provide great home learning environments for children and allowing parents to make child-rearing arrangements that are best for their child and for family well-being would pay off.

2.  Children attending full-time ECE/childcare as compared to part-time (around 12.5 hours per week or 2.5 hour sessions) do not have significantly better developmental outcomes. In other words it is the experience of attending a group early childhood programme that matters, and more time in the programme does not equal greater benefits for children

3.  The evidence also points to ECE/childcare having both developmental risks and benefits. There can be cognitive gains (at least in the short-term and dependent upon the effectiveness of the primary school children go on to) but there can also be negative outcomes for children’s health, mothers’ sensitivity in interaction with their children, problem behaviours and aggression in children.

4.  The combined effects of home factors, ECE factors, and primary school factors are important to consider. The British EEPE 3 – 11 project notes that attending a less academically effective primary school can wash out any benefits to be gained from attending a more effective or a higher quality preschool. Children who did not attend preschool or attended a lower quality preschool can “catch up”. Results from the Early Childhood Longitudinal Study-Kindergarten Cohort in the US similarly point out that the longer-term effects of preschool experience can depend on classroom experiences during at least the first years of school (Magnuson, Ruhm & Waldfogel, 2007). Public policy should therefore also be concerned about transition to primary school and ensuring that any benefits children gain from attending child care/ECE continue to be supported by effective academic instruction.

5.  Children attending childcare/ECE programmes can experience worse health outcomes than those who do not. They are more likely to need antibiotics for preventable illnesses. The evidence also indicates that children who may be more at risk may be protected by breastfeeding and being in a familial setting, especially before 2.5 years. 

On the basis of the best evidence, it seems unwise for public policy to promote a view that more intensive participation in good-quality (non-parental) ECE is best for every child and to encourage parents to use ECE if this is not their first choice or the right choice.  The research evidence suggests it would be wise to look for ways to support parents, and recognise that the primary purpose of early childcare and education is to support parents and families and not to replace them. The potential benefits of attending an early childhood programme will not be as strong for children otherwise. 

Furthermore, policy that seeks to make early childhood centres more like schools, risks institutionalising children's learning and makes it more difficult for centres and home-based ECE services to support parents.

As parents and family have the greatest impact on child outcomes it seems essential that information such as that provided in this paper is given to parents and not kept from them. 

Below is a copy of the report


Assessing the Evidence on ECE Effects on Children


Early childhood education/childcare policy in New Zealand is based on ideology and not on evidence of what is best for children. In other words, our ECE policy is not evidence-based. In September 2002 the government released a 10-year plan for ECE and the New Zealand Council for Educational Research Competent Children, Competent Learners study was drawn on to justify the values underpinning the plan and ECE policy. But the study had limited findings relating to ECE effects and quite major methodological problems. Policy development and implementation proceeded without a clear understanding and knowledge of positive and negative childcare/ECE effects.

A systematic review (in difference to a general review of the literature) of the best evidence, independently conducted, using a set of criteria for selection was carried out. This first working paper outlines the major pieces of evidence and discusses these in the light of current ECE policy considerations and directions in New Zealand.

The best evidence points to parents/family having a far greater impact than the childcare/ECE experience on children’s developmental outcomes. This suggests that effort put towards supporting families to provide great home learning environments for children and allowing parents to make child-rearing arrangements that are best for their child and for family well-being would pay off. Such efforts would more likely make a greater difference to children’s immediate and long-term developmental outcomes than simply encouraging parents to make more use of nonparental childcare as present ECE policy does.

Children attending full-time ECE/childcare as compared to part-time (around 12.5 hours per week or 2.5-hour sessions) do not have significantly better developmental outcomes. There is no advantage to be gained for children attending non-parental ECE for longer hours than parents require for childcare. This suggests that the way ECE policy financially incentivises early childhood services to increase the length of sessions and to promote 20 hours or 6-hour days as an optimal minimum, is unnecessary in terms of any added benefits for children. (It may also be at the cost of children not accessing formal ECE still not being able to access it).

The evidence also points to ECE/childcare having both developmental risks and benefits. While there may be cognitive gains (at least in the short-term) there can also be negative outcomes for children’s health, mothers’ sensitivity in interaction with their children, problem behaviours and aggression in children. Public policy that emphasises telling parents about the benefits of formal ECE and not about risks and other childcare options may increase the risks for children further. Enabling parents to be informed and to make informed choices would put parents in the position of being able to better manage risks and understand the size of potential benefits of ECE in interaction with home factors and later schooling choices and experiences.


A Story

Once upon a time there were 307 mostly well-to-do mostly Pakeha/NZ European children going to kindergarten, childcare, playcentre and one provider of family day care (Barnardos) in the Wellington region.

When they were nearly 5, 15 one-minute observations of them were done by strangers and the children were interviewed and given specific tasks to measure the competencies these grown-ups were interested in. Their centre and family day-care caregivers or teachers were asked to provide ratings of their competencies. Their parents were questioned about their use of early childhood education/childcare (ECE) and family. The quality of the preschool or service they were currently in was rated on 21 criteria taken mostly from instruments to measure quality in United States preschools, and organisational data was recorded like the number of qualified staff and ratios.

Every couple of years after that the children were checked on. When they were 8, well past their time in early childhood education, these children were joined by another 242 selected via a separate telephone survey study. And so it was said that this was a study of some 500 children.

The research, called Competent Children: Competent Learners, concluded that quality ECE plays an important part in children’s performance and continues to do so throughout their schooling. This was described as providing a “powerful legacy” (Wylie, Thompson, & Lythe, 2001, p. xviii). From the outset, the research had a political purpose as Helen May (2001, p. 237) wrote:

“The longitudinal Competent Children Project, initiated by Anne Meade and Cathy Wylie in the early 1990s, had been a politically astute move to document the long-term impact of early childhood programmes on the competencies of children in the education system.”

Cathy Wylie (2001, p. 1) explained:

“Policy interest in the value of early childhood education played a significant role in the original funding of the Competent Children project.”

And, Trevor Mallard (2001) as Minister for Education, said:

“From the beginning in 1992, this long-term study related particularly to early childhood resourcing and targeting, and this continues to be of important strategic value.”

In his role as Minister of Education Trevor Mallard cited the Competent Children study frequently. The Minister used the study to explain, or provide backing for, the Government taking a more hands-on approach to:

  • raising the intensity (hours and total time/years) of children’s participation in ECE;
  • getting children not yet in ECE into ECE, especially in poorer areas;
  • convincing parents of the benefits for children of ECE (usually meaning non-parental ECE or childcare);
  • making the Diploma of Teaching the only qualification recognisable for working in ECE and bringing in teacher registration requirements, and
  • increasing government subsidy payments to non-parental ECE services to provide publicly funded 20 hours of Free ECE for every 3- and 4-year-old.

The Minister actively promoted the research saying that “this New Zealand research is nationally and internationally regarded as a flagship study of early childhood education in New Zealand.” (Mallard, 2004). Assessing the evidence on ECE/Childcare Working Paper 3 On the release of the latest Age-16 Competent Children research, Steve Maharey (2007) as Minister of Education said:

“The research shows making early progress in literacy and numeracy is an important factor in determining whether a student remains at school after they turn 16 … This is more evidence to support participation in quality ECE, which will be boosted by the government’s 20 Hours Free ECE scheme. From 1 July, 92 thousand three- and four-year olds will be eligible to sign up for this popular policy.”

The annual ‘Social Reports’ published by the Ministry of Social Development state that “participation in ECE” can be used as, and is, an “indicator of social wellbeing” and cite Competent Children as providing the evidence for why participation is a predicator of future educational outcomes.

The success of the study in achieving its political purpose can be summed up in Anne Meade’s question to a recent Ministry of Education symposium on progress on the 10-year ECE Strategic Plan: “When will we start another longitudinal study of Competent Children?” (Meade & Royal Tangaere 2007, power point slide No. 26)

That in a nutshell is the story of the Competent Children: Competent Learners study. But the story cannot end here.

Within the climate of political praise and promotion for the Competent Children study we have not been able to question how the study stacks up against other research and how well it meets basic criteria for good research. We also have not really begun to look at what other New Zealand and international research might offer too.


Assessing the Competent Children Study Evidence

1. Validity and Generalisability

The research reports correctly note that the sample of children was not nationally representative. The sample was limited to the Wellington region, which may not have been a problem except that it was heavily biased toward children from families with high-incomes, who were Pakeha/European and whose mothers held a trade or tertiary qualification.

Furthermore, the children were selected from a narrow range of early childhood service types. Too many (75 out of 151) of the centres who were approached in the Wellington region either refused to participate or were considered by the researchers to be ineligible to participate. Due to the very small number of A’oga Amata in the study (n = 3) and the absence of other Pacific Island language nests and Kohanga Reo no conclusions should be drawn about these service types or about ECE effects on Maori and Pacific children.

The sampling methods used in the study were problematic for a third reason as well. Seventy percent of the children had attended or were concurrently attending other early childhood services. Given that children were not sampled until close on school starting age and that much of their other ECE experience was excluded from the study this raises a serious problem with the sample methods.

An important aspect of validity is whether the evidence looks real and seems plausible. Here there are further problems with the Competent Children research. Did it really capture children’s competencies adequately? Would parents agree that the ECE service their child was in at 4.6 to 5 years was responsible for their child’s continuing achievement in numeracy and literacy and that this ECE experience affected whether their child stayed at school till 16 or left? Remember the characteristics of the children sampled (mostly from high income families with well-educated mothers). Also consider that the ECE experience is a relatively short one in the lives of children, and the roles that family and other educational experiences play.

The research was designed to view ECE as something quite separate from the family - as seen in the study’s research question of “Can early childhood education experiences temper the influences of family backgrounds on children’s competence (positively or negatively)?” The view taken does not make sense. The interactions between the settings of home and ECE are important in the development of young children, and traditionally ECE has not been provided as an intervention but rather as an experience that compliments and builds on what the family provides. Then there is also the issue of whose competencies and whose values underpin the selection of competencies, for example there is a heavy emphasis placed on academic competencies in the study but does this mean that children’s gross-motor skills and talents in areas such as art and music are less or not important?

2. Causality

It is not possible to say whether any differences relating to children’s competencies at age-5 or later were due to their participation in the ECE service the researchers sampled. They may have been due to children’s earlier or other concurrent ECE experiences, to family experiences or to other child characteristics. The research was not good at identifying and controlling for a range of possible confounders. The researchers correctly note that their research reports only associations between variables and does not show causality. As the ECE experience was not measured over time (there was only a snapshot taken at close on 5 years of age) the paramount conditions for causality, being a sequence of events, is missing (Abramson, 1988). A Ministry of Education commissioned Strategic Literature Review on ECE noted in regards to the Competent Children study “… it would also be useful to follow children from birth and carry out prospective analyses of the influence of the quality of the three Microsystems of home, early childhood centre and school.” (Smith et al., 2000, p.43).

3. Coherence and Credibility

Any piece of research at least in part reflects the interests of the researcher and very likely that of the funding agency, some research more than other research. What all researchers should be able to do is to suspend their personal views and let the data speak for itself. This may cause some internal conflicts for the researcher but this is what an investigation is – the seeking of evidence and letting the evidence speak for itself.

When reading, and before using, research it is wise to check for possible bias. Often the research questions are the first signal of any potential bias or alignment to a particular view. Then a second clue is whether conclusions are reached and recommendations are made that do not match the data. A third clue is whether the researchers or their organisation has a personal stake in the outcomes of the study.

Any research to do with non-parental ECE/childcare effects is very political. Positive or negative effects can be interpreted in different ways by those who argue that exclusive mothers’ care is best and those who argue that childcare/ECE can be good, if not better for children, as long as its quality. It is very hard for researchers not to get caught up in ‘the day care wars’ (Karen, 1994).

When a lead researcher in the U.K.’s Families, Children and Child Care study gave a presentation to the National Childminders Association, the finding that young children do less well in groups was widely reported in the international and local media (Roberts, 2005). The research team had little choice because of the politically charged public debate and backlash that resulted but to issue a statement suggesting that the media had misinterpreted the contents of the presentation.

The American National Institute for Child Health and Human Development Study was initiated in 1991 in part to bring together “on a single team warring camps of researchers” (Shellenbarger, 2007, p. B1). This makes the findings from the research arrived at by the NICHD researchers from different universities and research organisation, more credible and offers the research some protection from political criticism.

The NZ Competent Children project has a known history of political associations. Trevor Mallard as Minister of Education did not want such associations to cloud people’s perception of the study. He made a point of explaining that the NZ Council for Educational Research who put out the research “is an independent, educational research organisation which provides educators, students, parents, policy makers and the public with innovative and independent research, analysis and advice” (Mallard, 2004). What he did not tell the media in his press release was that the Minister of Education nominates a representative for the NZCER Board and the Government provides a substantial financial operating grant to NZCER. Anne Meade was an adviser to Prime Minister David Lange as Minister of Education. In 1988 she was appointed by Lange to chair and author the first government report for the reform of ECE ‘Education to be More’. She knew the value to policy-makers of having research to back ideology. In 1992 as a senior lecturer at Victoria University she initiated the Competent Children project in conjunction with NZCER, and in early 1993 she joined NZCER as its Director. While Anne Meade’s contribution to the project diminished after the initial reports, this personal/political stake in the research can be seen to carry over into helping to ensure its uptake and its use to support the Government’s 10-year strategic plan ‘Pathways to the Future’.

In the Competent Children study factors other than ECE experience come up and show more strongly in each report. The reports from the study however still conclude and make recommendations for the importance of good-quality ECE for all children and better resourcing of ECE services by Government. The larger body of research expresses more caution about the impacts of ECE experience on children who are not from economically disadvantaged backgrounds. The larger body of research points to some negative effects of non-parental ECE on children. Effect sizes reported within the larger body of research point to the much greater impact of family factors compared to ECE factors.

4. Measuring Policy Impacts

There are three main ways of measuring policy impacts. The first is the conventional test of statistical significance. The problem with this test is that most of the time the test reflects the size of the study sample. And, the more you manipulate variables and look for statistically significant associations the more likely you are to get results. Sometimes the results may not always make much sense, for example, see discussion of first-time associations with aspects of ECE quality in Wylie, et al. (2004). This kind of test can be quite meaningless in a policy context. For policy purposes it is the size and strength of any test result, plus the robustness of the research and relevancy of the results to the general population that is important.

A better approach is to look at the magnitude of the difference or the ‘effect size’. Effect sizes are absent in the Competent Children early reports of ECE data. The researchers could not compare children participating in ECE to those who were not, because only children in ECE were sampled. Also, the research was not concerned with comparing children’s home learning environment with their ECE environment, so the research cannot tell us for example, if attending ECE has a large, moderate, small, or negative impact on children who experience a high-quality home learning environment.

A third approach to measuring the impact of ECE is cost-benefit analysis. Policy researchers use this to compare the overall costs of a policy intervention – with the economic or social benefits able to be achieved. This approach to measuring policy impacts is often used in measuring the impact of programmes for disadvantaged children, and often very small effects may be shown to be worthwhile in terms of their savings in the long term. The Competent Children study was not designed to measure the impact of ECE using cost-benefit analysis and so in this respect also it is not useful for insights into policy impacts.

5. Implications for Policy from the Competent Children Project

Putting to one side the range of limitations with the study in respect of researching and reporting on the impacts of ECE experience on children, some consistencies in reported findings between the Competent Children study and other research are apparent and these may have implications for policy.

The study results at age-16 tell us that 5 features of the ECE environment are related to a limited range of children’s competencies: staff responsiveness, staff guiding children in activities, staff asking children open-ended questions, staff joining children in their play, and a print-saturated environment (Hodgen, 2007). With the exception perhaps of a print-saturated environment, this finding lends support to an already large body of mostly US literature on ECE features statistically related to children’s social and academic outcomes. The contribution of these aspects is low – about 4% of the variability in students’ numeracy and logical problem-solving scores at age 16 (Wylie & Hodgen, 2007, p.2). These associations, however, may be artefacts of parents’ choice of ECE service – saying more about the parents’ selection of ECE environment and the people working in it than about the effect on children of being in that environment.

If, however, the ECE features identified in the Competent Children study cannot be explained by parent selection or other uncontrolled factors, then the findings give reason to be cautious in promoting more extensive ECE participation for children. Other New Zealand research reports that toddlers in childcare centres can experience little interaction with their teachers (Smith 1996a; Smith, 1996b). Around a third of 200 toddlers in 100 childcare centres were observed over 20-minute periods to have no any adult attention/interaction. When toddlers initiated interaction with staff (which they did frequently) a large proportion (a third) of their initiations were ignored. Joint involvement/play of teachers and children in activities was observed in only about 7% of the total observation time across centres. Smith (1996b, p.12) comments “In a busy centre, it may be difficult to find the time to focus on shared attention with children”.

The Competent Children research team explained how they tried with various analyses to show the importance of the qualification level for ECE staff, but they could not (Wylie, Thompson, & Kerslake-Hendricks, 1996). They had to “push [their] exploration of the effects of ECE training on ECS quality to the limits of [their] sample by focusing only on playcentres which had sufficient cell sizes of different children to staff ratios” (p. 116). Indicative associations only were found between the length of training and playcentre quality ratings and only on two subscales. Also, at odds with current policy to define the ECE qualification as a diploma of teaching and to require teacher registration were findings showing that children in playcentres and family daycare services performed better on some cognitive competencies than children in kindergartens and childcare centres (Wylie et al., 1996). The latest report at age-16 nevertheless makes the policy recommendation that “we need to ensure that services keep their qualified staff (turnover rates remain high in ECE services) and continue to enlarge the number who are qualified” (Wylie & Hodgen, 2007, p.3).

The time that children spend in ECE (starting age and hours) does not make a difference to any of the competencies when family factors are added to the analysis according to the Competent Children study (Hodgen, 2007). In other words, more intensive participation in ECE is not necessarily better or worse for children in the long-term. Smith et al. (2000) note that earlier and longer participation can benefit disadvantaged children in overseas early intervention programmes. The evidence for children attending ‘normal’ ECE programmes does not show this and is more mixed. Let us now look at other evidence.


What other NZ and International Research Offers in the Way of Information and Insights

Search and Evaluation Criteria

An extensive search of data bases for international and local literature childcare effects was conducted. Studies reported over the past 20 years that met the following criteria were examined:

  1. Included a control group of children not participating in ECE for comparison.
  2. Were longitudinal (i.e., following children while they were in ECE and following them up for some years after).
  3. Involved a random selection of children from the population or all the children in a birth cohort or region.
  4. Concerned typical ‘regular style’ ECE programmes in communities (as opposed to early intervention or atypical high-cost programmes e.g., High Scope/Perry; Headstart).

 Much of the research on the effects of non-parental ECE is plagued by methodology problems. Literature reviews can ignore or overlook this and report whatever evidence happens to fit the frame of reference for the review. In contrast, a systematic review process enabled the identification of the best evidence which is summarised in this paper.

The critical work of Penn and Lloyd (2007) on the long-term cost benefits of early childhood interventions highlights the value of the systematic review process as its involves mining the research evidence in great detail. Spelling out why we should be cautious about what we read in academic papers (or don’t read but are told to believe!) is more help in getting policies and practices right for children than letting established assumptions about early child care and education outcomes lie.

Most studies have focussed on only the very economically poor ethnic minority children and on high-cost special programmes, like Headstart and the Perry Project in the US. Such programmes are probably impossible to duplicate exactly in NZ and the cost of doing so would be out of the reach of most communities. Any benefits for children are unlikely to be replicated in other settings unless the conditions of the programme can be duplicated. Reviews of the literature on non-parental ECE effects often over-generalise and regard early intervention programmes and typical community ECE programmes as one and the same thing (Belsky, 2006). But they are not. To include what is known about the outcomes for economically disadvantaged children in exceptional experimental programmes often for research purposes as meaningful for other children in other contexts is a daft thing to do.

Studies of the childcare experience over-time (i.e., longitudinal research) provide a more comprehensive view of effects than studies that take just a snapshot of children’s ECE/childcare experience. Employing longitudinal methods more successfully isolates the effect of childcare on children as Hickman (2006) reports after carrying out both cross-sectional and longitudinal analysis on data from the Early Childhood Longitudinal Study – Kindergarten Cohort of 1998-9. While cross-sectional analysis indicates that children who have centre-based ECE experience exhibit advanced maths and reading skills over home children but poorer peer-related social skills this did not hold true when employing longitudinal analysis. Longitudinal analysis showed that the cognitive benefits of centre-based ECE did not persist and that social skills deteriorated, suggesting that children benefit less from the centre-based experience than previously thought.

A small selection of among the very best most rigorous longitudinal studies on normally occurring/regular non-parental early childhood care and education will now be presented and key findings of interest to policy makers, researchers, and providers of formal ECE services will be highlighted.

The Christchurch Health and Development Study

The Longitudinal Study of Early Childhood Education and Subsequent Academic Achievement looked at the relationships between duration of attendance in ECE and measures of intelligence, reading comprehension, mathematical reasoning, and school ability up to the age of 13 years for an original birth cohort of 1265 children (Fergusson, Horwood, & Lynsky, 1994). It has rarely been cited by ECE policy-makers and researchers. The study was probably ignored because the findings were not politically palatable. Yet the research was quite rigorous and much more so than the Competent Children study. It involved a large cohort of children, both attending and not attending ECE services, and controlled for a variety of possible founding factors. As well as mothers education, socio-economic status or family income, ethnicity and gender the researchers controlled for position in the family, paternal education, birth weight, breastfeeding during the first year of life, maternal emotional responsiveness and a measure of the frequency with which the mother made punitive responses to the child’s behaviour.

Fergusson et al concluded from their findings that increasing exposure to ECE is associated with small detectable increases in ability and achievement scores. But,

“… at the same time, the relatively small effect sizes found and the uncertainties of the evidence suggest it would be unwise to aggressively promote the view that early education of the type provided to this cohort makes an important contribution to subsequent academic achievement. At best any benefits found in this study are small and it is possible that even these benefits may be due to uncontrolled factors rather than the benefits of early education.”

For these reasons the results of the present study are viewed as providing upper limit estimates of the likely contribution of New Zealand early education services to subsequent academic achievement. (p. 115) Unfortunately, this much more rigorous study still missed a small but potentially confounding factor. And that is, that there are fundamental differences in parent involvement across different types of early childhood services. In kindergartens and childcare centres, the child carers or teachers are not the parents. In contrast, playcentres require parents to engage in training and this may lead to an enhancement of parenting skills which in turn may have been the dominant factor in children’s academic performance. The NICHD study which is next outlined defined childcare as any care provided on a regular basis by someone other than a child’s mother. In other words, care by a father was defined as in-home childcare/ECE along with care/education by paid caregivers/teachers and other relatives.

The U.S.’s National Institute of Child Health and Human Development Study of Early Child Care (NICHD)

The NICHD study followed from one month old some 1,300 children living in 10 different locations in the United States. Whereas the Competent Children provided only a snapshot of children’s ECE experience at the point of starting school, an important strength of the NICHD study is that it viewed “reliance on a single assessment at one point during the first five years to be an [in]adequate representation of a child’s experience in childcare” (NICHD, 2002, p.137). Information about amounts and types of care was collected by the researchers every 3 to 4 months. Associations between family features and children’s development were discovered to be two to three times stronger than links between childcare/ECE and development. Key findings from the report published by NICHD (2006) and from other NICHD papers are outlined briefly below.

Family Features: Family features held greater and more consistent predictive power than ECE features. The association between family features and children’s developmental outcomes was two to three times stronger than links found between ECE features and children’s development. Most children in exclusive maternal care at around 1 year and at 2 and 3 years of age had cognitive, language and achievement scores like those of children in childcare. Therefore, exclusive maternal care is not a bad thing for children’s development and ECE/childcare does not necessarily put children at an advantage over those who do not experience it. Family features that predicted children’s cognitive/language and social development included parents’ education, family income, two-parent compared to single-parent father, mothers’ psychological adjustment and sensitivity – and the social and cognitive quality of the home environment.

Intensity of ECE participation: The quantity, or average amount of time that children spent in childcare/ECE each week, was not found to be related to children’s cognitive skills, language skills, or to their school readiness prior to school entry. Children did not gain any greater benefit from spending more time in childcare, even in childcare rated as being of high-quality. The more time children spent in childcare during their first 3 years the less sensitivity mothers showed in their interactions with them. This was also found at 4.6 years and in first grade, but interestingly only for white children. Even with controls for several possible mediators of the quantity of childcare effects (e.g., quality of centre, type of care, observed parenting etc), more time in childcare/ECE consistently predicted higher levels of behaviour problems – defiance/disobedience and aggression (NICHD, 2003b).

Type of Childcare/ECE: Centre-based ECE had both positive and negative effects. While centre-based ECE contributed positively to children’s cognitive development up to age 4.6 years and to more positive social behaviours to age 3 years, the group nature of centre-based ECE also meant more problem behaviours just before and after children started school.

Health: Centre-based ECE attendance was associated with increased ear infections, upper respiratory and stomach illnesses (NICHD, 2003c). Children in large group care (more than 6 children) were 2.2 times more likely to have upper respiratory tract illness, 1.6 times to have an ear infection, and 1.4 times to have a gastrointestinal tract illness as children not in ECE or in small group care. Children in ECE between 2 – 3 yrs were less likely to contract gastrointestinal respiratory tract illnesses between 3 – 4 ½ yrs than children who started after they were 3, suggesting increased immunity as a result of exposure. The hours that children attended centre/group ECE made no difference to health risk levels.

Quality: The quality of childcare was associated with children’s cognitive and language development, but slightly (NICHD, 2003a). The links between childcare quality and social development outcomes were weaker still. Effect sizes showed that family features and parenting were more important for children’s cognitive, language and social development than the quality of childcare alone. Children who experienced higher quality care showed slightly more positive outcomes than those in lower quality childcare. A possible limitation is that childcare quality was determined using criteria to measure the quality of the childcare arrangement and not based on the quality of experience children were coming from or would have otherwise received had they not been in childcare.

The U.K.’s Effective Provision of Preschool Education (EPPE) Project and EEPE 3 – 11 years Project

This study investigated the impact of preschool experience on some 3,000 children from age 3 attending one of the main types of services in England: playgroups, local authority nurseries, care, and education (integrated) centres, private day nurseries, nursery schools and nursery classes. From the outset the project was concerned primarily with documenting children’s preschool experience and looking at its quality and effectiveness. However, a small comparison sample of around 300 children who had no group preschool experience was recruited when they started primary school. The latest reports on children’s cognitive and social development to age 10 are provided by Sammons et al. (2007a & 2007b). Other papers drawn on here in summarising the results include Sylva et al. (2004) and Siraj-Blatchford et al. (2006).

The EPPE research goes further than other childcare studies in that the researchers formulated a way of distinguishing between ‘quality’ and ‘effective’ preschools. While ‘quality’ was measured in the usual way of pre-defined criteria on which observers provided ratings, ‘effectiveness’ was derived from a ‘value-added’ model of the children’s progress across the 141 preschool settings. Preschools in which the children made more progress in reading or early number concepts were classified as more effective. In addition, the project also examined the impact of the home learning environment on children’s developmental outcomes. Unfortunately, as the comparison ‘home children’ did not enter the study until later less is known about how these children may have differed from the preschool sample in their early home experiences.

Family Features: The early results showed that the home learning environment had a particularly strong effect on children’s cognitive development. After child age, it was the variable with the strongest effect, stronger than both social class and parental education. At age-10 the quality of the home-learning environment during the early years and parents’ (especially mothers’) qualification levels continued to be the most important factors relating to a child’s attainment in reading and mathematics. The data suggested to the researchers that the home-learning environment can be viewed as a ‘protective’ factor in reducing the incidence of special educational needs. The home learning environment was only moderately associated with mothers’ educational level, suggesting that it is perhaps more what the mother does for her child than who she is that makes the difference. Given the significance of family features it was not surprising that the researchers found that preschools which directly promoted activities for parents and children to engage in together had the greatest positive impacts on children.

Intensity of ECE/childcare Participation: Longer days were not found to lead to significant benefits for children’s outcomes over part-day (2.5 hour session or 12.5 hours weekly) attendance. A small effect for longer duration of ECE on academic and social outcomes was found. However, when children started before they were 2 no additional effect of the time before 2 years of age was found. In other words, short exposure to preschool from about 2 years of age has the same benefits for children’s development as more intensive participation. However, more intensive participation can also have some negative effects, for example, higher levels of preschool participation - more than 2000 hours - was associated with higher levels of antisocial/ worried behaviour in children at the start of primary school.

Type of Childcare/ECE: Being cared for by a relative, such as a grandmother, before entering the study, showed moderate effects on children’s cognitive outcomes, higher co-operation, and less anti-social behaviour compared to children in different care arrangements. On entry to school the effect size for reading and mathematics for children who attended preschool compared with the ‘home children’ was moderately significant, suggesting that preschool experience over none does benefit children, but this weakened/faded with each year of schooling. Preschool experience compared to none had a moderate effect on children’s independence and concentration, self-regulation and peer sociability on starting school as would be expected from prior exposure to being in a group setting. On the other hand, home children showed less anxious behaviour on starting school than preschool children.

Amongst the different preschools, significant variation in effectiveness was found, showing that differences within different preschools are likely to be more important than differences between type. Education and care preschools (integrated centres) showed the biggest significant impact on children’s language (0.28), pre-reading (0.28) and early number concepts (0.40), followed by nursery schools (language 0.17; pre-reading 0.19; and number 0.24) and private day nurseries (language 0.21; pre-reading 0.26; and number 0.17). However, most of the effect sizes remain in the mostly small to moderate range. Despite being one of the most expensive providers, children in local authority nurseries tended to make relatively slower progress, especially for pre-reading.

Quality and Effectiveness: Compared to children who did not go to preschool, there was evidence of a small continuing effect on children’s outcomes in mathematics and reading if the preschool they attended was high quality or effective and if they then attended a more academically effective primary school. If children attended a lower quality preschool, they may not have attended one at all because their academic results were not significantly differently from children who had not attended preschool. The significance of positive effects of attending a higher quality or effective preschool also fade for children who go on to a primary school that is academically less effective.

The Canadian Population based Longitudinal Study

Research on childcare/ECE attendance, breastfeeding and the frequency of antibiotic treatments carried out as part of a Canadian population based longitudinal study gives further insights into childcare effects (Duois & Girard, 2004). The study involved 1841 children, first seen at 5 months and then at yearly intervals to 5 years. The study appears to be well-designed and rigorous. It provides further evidence to the NICHD study (summarised above) that childcare attendance may result in worse health outcomes.

Childcare attendance increased the number of antibiotic treatments at each of the studied ages. It raised by 30% the odds of having one antibiotic treatment between 2.5 and 4 to 5 years of age and it doubled the odds of having three antibiotic treatments between 1.5 and 4 to 5 years of age.

When breastfeeding and childcare attendance were combined, children not in childcare were protected against antibiotic treatments by being breastfeed for at least 4 months at 1.5 and 2.5 years of age. For children starting in childcare before 1.5years, the odds of having 6 or more prescription antibiotic treatments was 7.5 times higher if they were never breastfed. The odds were 4 times higher if breastfeed for the first four months.

Dubois and Girard conclude that attending childcare/ECE puts children at risk of worse health outcomes, it “puts children in contact with other children and creates a milieu suitable for the development of infections” (p. 2039). Breastfeeding appears to serve as an important ‘protective’ factor.


A look at International Studies on Student Achievement for any further Insights

The Progress in International Reading Literacy Study (PIRLS)

A paper analysing the 2001 PIRLS results from the International Association for the Evaluation of Educational Achievement (Martin, Mullis, & Gonzalez, 2004) provides some further food for thought regarding childcare/ECE effects on children’s subsequent academic achievement. The PIRLS study examines reading literacy achievement in primary schools across 35 Countries. The findings relating to NZ are interesting and further study of the results would be useful. We see in the results that NZ rated very highly (near the top in countries) in home literacy activities. Most striking was that when children’s participation in 2 or more years of preschool was compared for the lowest achieving third of students and the highest achieving third of students, New Zealand had a high percentage of its lowest achieving students (72%) and a similarly high percentage of its highest achieving students (70%) receiving 2 or more years of preschool. This could suggest one or two things. First, that preschool/ECE attendance in New Zealand may have little subsequent impact on achievement levels in reading literacy at school age or on reducing educational inequities. Second, children’s home learning environment may be a more likely contributor to New Zealand’s overall high ranking in reading literacy than preschool participation.

Programme for International Student Assessment (PISA)

Every three years PISA assesses how far students near the end of compulsory education (at age 15) in OECD and partner economies/industrial countries have acquired some of the knowledge and skills for full participation in society. The latest (2006) country mean scores for mathematics, reading and science place New Zealand high in the country rankings (see Tables 1, 2 & 3 below). However, it is noteworthy that even with New Zealand’s internationally high participation rate in ECE and relatively egalitarian society our students’ achievement is amongst the most widely spread – our students are both amongst the best and the worst academically (see Table 4 for example on science scores). Our education system so far has not been able to address the significant tail end of under (or low) achievement. It is too early to tell yet whether recent government policy to increase the intensity of children’s participation in ECE and to provide 20- Hours Free ECE in only teacher led services will help to reduce educational inequities. However, when you look at funding arrangements and provisions for childcare/non-parental ECE in other high performing OECD and partner countries it is difficult to conclude that our current ECE policy will make much of a further difference to student achievement.

For example, Finland outperforms us and other countries (except for Korea). Yet since 1990 Finish parents have had a right to day care for children under-three years of age either in municipal day-care or by receiving a child home-care allowance in order to care for their own child at home. Since 1997 families could receive a private childcare allowance for providing their children with private care or make use of a daycare centre or family daycare provided by their local municipality. Finland thus has a much more flexible system of parental choice and funding than NZ. As well, children start primary school at age 7 in Finland so an earlier start in formal education may not necessarily be related to higher academic achievement in the long term. Interestingly Canada performs consistently slightly better than New Zealand in its PISA results, yet Canadian provisions for ECE could perhaps be described as less advanced and not as highly regulated and funded than New Zealand’s.


Table 1: Mathematics score in PISA 2006


Mean Score













New Zealand






OECD all countries average




Hong Kong-China







Table 2: Reading score in PISA 2006


Mean Score

Korea 556


Finland 547


Canada 527


New Zealand 521


OECD all countries average




Hong Kong-China



Table 3: Science score in PISA 2006


Mean Score

Finland 563


Canada 534


Japan 531


New Zealand 530


Australia 527


OECD all countries average




Hong Kong-China 542


Chinese Taipei 532


Estonia 531



Table 4: Percentage of students at the 2 lowest (level 1 and below) and 2 highest (levels 5 and 6) proficiency levels on the science scale


Level 1 and Below

Levels 5 & 6







Hong Kong-China












Chinese Taipei















New Zealand



Source: OECD PISA 2006 database. Table 2.1a, PISA 2006: Science Competencies for Tomorrow’s World. Countries are ranked in descending order of percentage of students at Levels 1 and Below Level 1


Conclusions and Discussion

According to the best evidence the impact of the family is much greater than the impact of childcare/ECE. This holds true whatever the researcher-rated quality of the ECE service. The best evidence does not show that good quality ECE is better necessarily than care within the family or has a greater impact on children’s achievement and other outcomes. The best evidence suggests only that non-parental ECE can help support the work of the family. Public policy would be most effective in improving child outcomes if it:

  • Supported parents/families to make child-rearing arrangements they deemed best for their child and family functioning, thereby reducing stress and economic coercion to make decisions they would sooner not;
  • Revised and changed the assumption that non-parental ECE is better for children; and
  • Rewarded/valued and provided effective supports to enhance the informal care and early education provided by parents.

Family characteristics are the strongest predicators of child outcomes. ECE effects are likely therefore to be strengthened if a child’s ECE/childcare experience successfully builds on this. For children at risk of poorer outcomes because of less optimal family backgrounds, ECE/childcare is more likely to make a positive contribution (Hungerford & Cox, 2006).

From a policy perspective, an important goal could be for researchers and officials to do work around the question of how to ensure that childcare/ECE quality (not just standards or proxies for quality such as staff qualification and ratios) is indeed higher quality for the individual child, family, and subgroups of families (Farquhar, 2005). To maximise the benefits of childcare attendance for children, knowledge of children’s home and usual experiences provides an important basis from which to extend.

Children attending full-time as compared to part-time (around 12.5 hours per week or 2.5 hour daily sessions) do not have significantly better developmental outcomes. In other words, there is no advantage to be gained for children attending non-parental ECE for longer hours than parents require for childcare. The evidence suggests that the way funding to ECE teacher-led services is structured to encourage services to provide longer days for children gives no added value for children’s educational achievement. It should be questioned as to whose interests are being served by public policy incentivising services to provide longer hours and parents to use longer hours of childcare (e.g. kindergartens being financially motivated to provide 6-hour days for children).

Participation in ECE brings developmental risks and benefits. The picture is not all as rosy for non-parental ECE/childcare as we would hope or want to believe. Whether public policy is on the right road by telling parents that teacher-led ECE is better for children should be questioned. It may be that if unbiased information on potential risks and the size of possible benefits is given to parents in a timely manner, then parents can make more informed choices and manage risks to better advantage their child’s development.

The combined effects of home factors, ECE factors, and primary school factors are important to consider. As mentioned earlier, the British EEPE 3 – 11 project notes that attending a less academically effective primary school can wash out any benefits to be gained from attending a more effective or a higher quality preschool. Children who did not attend preschool or attended a lower quality preschool can “catch up”. Results from the Early Childhood Longitudinal Study - Kindergarten Cohort in the US similarly point out that the longer-term effects of preschool experience can depend on classroom experiences during at least the first years of school (Magnuson, Ruhm & Waldfogel, 2007). Public policy should therefore also be concerned about transition to primary school and ensuring that any benefits children gain from attending ECE are supported by more effective academic instruction.

Children attending childcare/ECE programmes can experience worse health outcomes than those who do not, or those in small group settings with less than 6 other children. This is something that is important for parents to know and to weigh up in their decision-making on childcare/ECE options. The evidence also indicates that children who may be more at risk may be protected by breastfeeding and being in a familial setting, especially before 2.5 years. There is no data on breastfeeding rates for children in ECE/childcare in New Zealand to be able to estimate risk, and because early childhood centres are not required to have a policy on breastfeeding support it may be assumed that children in centres are at higher risk of worse health outcomes.

Based the best evidence, it seems unwise for public policy to promote a view that more intensive participation in good-quality [non-parental] ECE is better for [all] children. In addition, theory and research on young children’s learning not able to be covered within the scope of this paper points to the value for children of being separated as little as possible from community and family activities and of not closing the adult world off to them through ECE/childcare attendance (for example see Brennan, 2005).

Negative effects on children’s health, anxiety levels, problem and aggressive behaviour, and parent-child relations including maternal sensitivity should not continue to be overlooked and only the evidence of cognitive gains emphasised. A more balanced consideration of the range of childcare/ECE effects is called for. As well, within NZ early childhood policy a more balanced view of the contributions of parental and non-parental ECE is needed. As family/parents have the greatest impact on child outcomes it seems essential that information such as that provided in this paper is not kept from parents. Enabling parents to be informed and to make informed choices would put parents in the position of being able to better manage risks and understand the size of potential benefits of ECE in interaction with home factors, their child’s characteristics, and later schooling choices and experiences.



Abramson, H.J. (1988). Making sense of data. Oxford: Oxford University Press.

Belsky, J. (2006). Early child care and early child development: Major findings of the NICHD Study of Early Child Care. European Journal of Developmental Psychology, 3(1), 95-110.

Brennan, M. (2005). “They just want to be with us.” Young children learning to live the culture. A post-Vygotskian analysis of young children’s enculturation into a childcare setting. Unpublished doctorate thesis, Victoria University of Wellington. Wellington, New Zealand.

Dubois, L. & Girard, M. (2004). Breastfeeding, day-care attendance and the frequency of antibiotic treatments from 1.5 to 5 years: a population-based longitudinal study in Canada. Social Science and Medicine, 60, 2035-2044.

Farquhar, S-E. (2005). How to pick remarkable quality childcare and education: with guidelines for parents, teachers and service providers, course presenters and staff trainers. Wellington: ChildForum Research.

Fergusson, D.M., Horwood, L.J., & Lynsky, M.T. (1994). A longitudinal study of early childhood education and subsequent academic achievement. Australian Psychologist, 29(2), 110-115.

Hickman, L.N. (2006). Who should care for our children? The effects of home versus center care on child cognition and social adjustment. Journal of Family Issues, 27(5), 652-684.

Hodgen, E. (2007). Early childhood education and young adult competencies at age 16. Technical report 2 from the age 16 phase of the longitudinal Competent Children, Competent Learners study. Wellington: Ministry of Education.

Hungerford, A. & Cox, M.J. (2006). Family factors in childcare research. Education Review, 30(5), 631-655. Karen, R. (1994). Becoming attached. New York: Warner Books

Maharey, S. (2007). New research backs government’s Free ECE policy. acks_governments_free_ece_policy_steve_maharey.html Retrieved 28 Apr 2008

Mallard, T. (2001). Early childhood education for a democratic society. Media release 26 October 2001. Mallard, T. (2004). Internationally recognised research to continue. Media statement 9 August 2004.

Magnuson, K.A., Ruhm, C. & Waldfogel, J. (2007). The persistence of preschool effects: Do subsequent classroom experiences matter? Early Childhood Research Quarterly, 22(1), 18- 38.

Martin, M.O, Mullis, I.V.S., & Gonzalez, E. (2004). Home environments fostering children's reading literacy: Results from the PIRLS 2001 study of reading literacy achievement in primary schools in 35 Countries. International Assn for the Evaluation of Educational Achievement. Retrieved 2 February 2008

May, H. (2001). Politics in the playground: The world of early childhood in postwar New Zealand. Wellington: Bridget Williams Books & NZCER. Meade, A. & Royal-Tangaere, A. (2007). Travelling Pathways to 2007. Ministry of Education Early Childhood Education Symposium Proceedings, 2-3 May 2007

NICHD Early Child Care Research Network (2002). Early child care and children’s development prior to school entry: Results from the NICHD Study of Early Child Care. American Educational Research Journal, 39(1), 133-164.

NICHD Early Child Care Research Network (2003a). Does quality of child care affect child outcomes at age 4 ½ ? Developmental Psychology, 39, 451 – 469.

NICHD Early Child Care Research Network (2003b). Does amount of time spent in child care predict socio-emotional adjustment during the transition to kindergarten? Child Development, 74, 976 – 1005.

NICHD Early Child Care Research Network (2003c). Child care and common communicable illnesses in children aged 37 to 54 months. Arch Paediatric Adolescent Medicine, 157, 196- 200.

NICHD Early Child Care Research Network (2006). The NICHD Study of Early Child Care and Youth Development (SECCYD). United States: US Dept of Health and Human Services, NICHD

Penn, H. & Lloyd, E. (2007). Richness or rigour? A discussion of systematic reviews and evidence based policy in early childhood. CIEC online journal, 8(1), 3 -18.

Roberts, Y. (2005). Mum’s care best for kids’ development. NZ Herald, 3/10/05 p. B2. Article syndicated from the London Observer.

Sammons, P. Sylva, K., Melhuish, E., Siraj-Blatchford, I., Taggart, B., Grabbe, Y. & Barreau, S. (2007a). Effective pre-school and primary education 3 – 11 Project (EPPE3-11). Influences on children’s attainment and progress in key stage 2: Cognitive Outcomes in Year 5. Research Brief No: RB828.

Sammons, P. Sylva, K., Melhuish, E., Siraj-Blatchford, I., Taggart, B., Grabbe, Y. & Barreau, S. (2007b). Effective pre-school and primary education 3 – 11 Project (EPPE3-11). Influences on children’s attainment and progress in key stage 2: Social/behavioural outcomes in Year 5. Research Brief No: RB007.

Shellenbarger, S. (2007). Day care is linked to behaviour lasting through 6th Grade. The Wall Street Journal Online, March 26, p. B1

Siraj-Blatchford, I., Sammons, P., Taggart, B., Sylva., K., Melhuish, E. (2006). Educational research and evidence-based policy: The mixed-method approach of the EPPE project. Evaluation and Research in Education, 19(2), 63-82.

Smith, A.B. (1996a). The quality of childcare centres for infants in New Zealand. Monograph Number 4 of the NZARE ‘State of the Art’ series.

Smith, A.B. (1996b). Quality childcare and joint attention. Paper presented at the NZARE Conference, Nelson, 5 – 8 December 1996.

Smith, A.B., Grima, G., Gaffney, M., Powell, K., Masse, L., & Barnett, S. (2000). Strategic research initiative literature review: Early childhood education. Report to the Ministry of Education. Dunedin: Children’s Issues Centre, University of Otago.

Sylva., K., Melhuish, E., Sammons, P., Siraj-Blatchford, I., & Taggart, B. (2004). Effective preschool education. A longitudinal study funded by the DfES 1997 – 2004. London: DfES and the Institute of Education.

Wylie, C. & Hodgen, E. (2007). The continuing contribution of early childhood education to young people’s competency levels. Wellington: NZCER. Wylie, C. (2001). Competent children: Findings and issues from the first 7 years. Paper presented to a Ministry of Social Policy Seminar, Wellington, April 6 – 7.

Wylie, C., Thompson, J. & Lythe, C. (2001). Competent children at 10: Families, early childhood education and schools. Wellington: Ministry of Education and New Zealand Council for Educational Research.

Wylie, C., Thompson, J., Hodgen, E., Ferral, H., Lythe, C. and Fijn, T. (2004). Competent Children at 12. Wellington: Ministry of Education and New Zealand Council for Educational Research.



Are you interested in joining us? 
Become a member and also gain access to our significant online knowledge base 

Membership Options


Who is this for?
Teachers - Student Teachers - Parents
$78.00 one year
$145.00 two years
Your own personal username and password.

ECE Service Provider  

Who is this for?
Service providers with one or more licensed ECE services
Starting from $198.00 for one year
A unique username and password

NZ-International Research in
ECE Journal Subscription

Who is this for?
Libraries, universities, polytechnics and organisations
$150.00 annual renewable in November each year
A username and password or IP address access