What is the benefit-cost ratio of investing in childcare services regarding productivity returns of academic achievement at the end of compulsory schooling?
In this cohort study including 8936 children, childcare attendance was associated with higher academic achievement at the end of compulsory schooling only for children of mothers with lower education. The benefit-cost ratio for each £1 (US $1.40) invested in full-time childcare attendance for children of mothers with low education was £1.71 (US $2.39) for those who achieved a Level 2 General Certificate of Secondary Education qualification.
These results suggest that subsidizing childcare attendance for children of mothers with low education would yield economic benefits in the long run.
Low school preparedness is linked to high school dropout, poor employment, and negative outcomes. Childcare attendance may increase school readiness and foster academic achievement.
To explore whether childcare attendance was associated with academic achievement at the end of compulsory schooling (age 16 years in the UK), whether maternal education level was a moderator, and the benefit-cost ratio of childcare regarding productivity returns of academic achievement.
Design, Setting, and Participants
In this cohort study, data were included from the Avon Longitudinal Study of Parents and Children (ALSPAC) born from April 1991 to December 1992 and the UK National Pupil Database for examination results. Data on academic achievement at age 16 years were available for 11 843 participants. Data were collected from June 2006 to June 2008, and data were analyzed from September 2019 to May 2020.
On average, 3.7%, 5.9%, and 90.4% attended childcare full time, part time, and less than 10 hours per week, respectively. Maternal education was assessed by questionnaire during pregnancy. Analyses included weights for population representativeness and propensity score weights to account for parental selection into childcare.
Main Outcomes and Measures
Academic achievement was defined as no certificate, Level 1 General Certificate of Secondary Education (GCSE; limited training), or Level 2 GCSE (qualification for academic post-16 education; high school diploma equivalent). Lifetime productivity return estimates were withdrawn from previous economic analysis based on pupil’s qualifications.
Of 14 541 children in the ALSPAC study, 8936 children had complete data on childcare attendance, academic achievement, and maternal education levels. Of these, 4499 (50.3%) were male. Attending childcare was associated with higher probabilities of obtaining a Level 1 or 2 GCSE qualification (Level 1: relative risk, 1.41; 95% CI, 1.16-1.73; Level 2: relative risk, 1.62; 95% CI, 1.30-2.01); however, this association was moderated by the child’s maternal education level. When children of mothers with low education attended childcare, their probability of no GCSE qualification went from 28.9% (95% CI, 26.8-31.0) to 20.3% (95% CI, 18.0-22.8), whereas children of mothers with higher education had a probability of no qualification of less than 10% regardless of childcare attendance. The benefit-cost ratio for each £1 (US $1.40) invested in full-time childcare attendance for children of mothers with low education was £1.71 (95% CI, 1.03-2.45; US $2.39; 95% CI, 1.44-3.43) for those who reached a Level 2 GCSE qualification.
Conclusions and Relevance
Promoting universal childcare with facilitated access for children of lower socioeconomic backgrounds deserves to be considered as a way to reduce the intergenerational transmission of low academic achievement.
Academic achievement at the end of compulsory schooling (age 16 years in the UK) is of major importance to individuals, families, and society. Failure to obtain a high school diploma (ie, secondary school dropout) has been linked to substantial societal costs related to poor employment, low income, and negative physical and mental health outcomes.1,2 Whereas academic achievement facilitates upward mobility,1,3 dropping out of high school contributes to low social mobility and intergenerational transmission of economic and health problems.1,3
Socially disadvantaged children may be at higher risk of poor academic achievement than others, partly due to a lack of preparedness for learning and complying with teachers’ instructions at kindergarten entry.4 Experimental or quasiexperimental studies have shown that exposure to higher-quality environments outside the home, such as childcare, can attenuate the impact of socioeconomic status on cognitive and academic trajectories5-7 and that provision of such programs can be cost-effective.8,9 The protective effect of childcare attendance is related to group-based dynamics that promote the development of early social skills (eg, socialization with other children, self-regulation of emotions) and to goal-oriented activities that enable children to start school better equipped to manage classroom demands.10,11
The evidence from observational studies in programs available to the general population (community childcare) is less clear, however. A 2018 meta-analysis of natural experiments12 reported a positive association between community childcare attendance and long-term academic achievement but noted the paucity of longitudinal studies. Further, to adequately investigate the issue, observational studies need to consider social selection of families into childcare services.13 Characteristics known to be associated with academic achievement (socioeconomic status, neighborhood quality, and parental employment) can also influence choices.1
Therefore, the present study examined 3 research questions, with particular emphasis on controlling for social selection. First, was childcare attendance associated with academic achievement by the end of compulsory schooling? Did maternal education moderate the association between childcare attendance and academic achievement? What is the benefit-cost ratio of investing in childcare services regarding productivity returns of academic achievement at the end of compulsory schooling?
Study Design and Participants
We conducted a longitudinal population-based cohort study using the database of the Avon Longitudinal Study of Parents and Children (ALSPAC), which has been linked to the UK National Pupil Database for grades achieved on the Key Stage 4 (KS4) examinations for each individual. ALSPAC is an ongoing epidemiological study of 14 541 pregnant women residing in Avon, United Kingdom. The women had an expected delivery date between April 1991 and December 1992 (85% of the eligible population).14 Written informed consent for use of data collected in questionnaires and at clinic visits was obtained from participants in accordance with the ALSPAC Ethics and Law Committee and local research ethics committees. The original ALSPAC sample is representative of the general population.15,16
Data on academic achievement at age 16 years were available for 11 843 participants. Reasons for nonavailability included enrollment in private school and homeschooling. As we used administrative linked data, some participants had KS4 results available, but they did not complete any questionnaire in the ALSPAC study. They were thus excluded from our final analytical sample. Our final sample thus included 8936 participants with data on maternal education, childcare attendance, and academic achievement (Table 1).
Primary Outcome Measure: Academic Achievement at KS4
The KS4 is the standardized examination at the end of compulsory schooling in the UK, known as General Certificate of Secondary Education (GCSE). This examination is taken at an approximate age of 16 years. KS4 results were categorized as (1) obtaining a GSCE Level 2 qualification (5 or more A* to C grades, necessary for entrance to post-16 education and training courses; rough equivalent of a US high school diploma); (2) GCSE Level 1 qualification (5 or more A* to G grades, which offers more limited educational opportunities than GSCE Level 2); or (3) less than Level 1 qualification, or low academic achievement.
Predictor: Childcare Attendance
Childcare attendance was defined as group-based childcare in settings such as nurseries and playgrounds, with children cared for by someone other than parents or relatives. Mothers reported number of hours per week spent in childcare, if any, at the child’s age 8 months, 15 months, 2 years, and 3 years. To capture persistent differences in childcare attendance, we modeled developmental trajectories of childcare attendance using latent class growth analysis. A 3-group trajectory model was selected, based on fit and parsimony indices (bayesian information criteria) and group classification accuracy (entropy). A lower bayesian value indicated a more parsimonious model, whereas an entropy value closer to 1 indicated greater precision (possible range, 0 to 1).
Because low-frequency cells were noted when childcare trajectories were cross-referenced with maternal education and KS4 results, we also created a 2-level childcare variable where we combined full-time and part-time attendance categories, where 0 indicates no or low childcare attendance and 1 indicates part-time or full-time attendance, to investigate the interaction between childcare attendance, maternal education, and children’s academic achievement.
Moderating Variable: Maternal Level of Education
Highest maternal education level was derived from a questionnaire completed during pregnancy, with answers grouped as per UK standards: less than a Certificate of Secondary Education (CSE); CSE or General Certificate of Education Ordinary, equivalent to the GCSE, with national examinations at age 16 years; vocational school; General Certificate of Education Advanced, with national examinations at age 18 years; or university degree. Using previous studies to determine appropriate cutoffs,17 we split maternal education levels into 2 categories: low (less than CSE, CSE, vocational, or General Certificate of Education Ordinary, which is the equivalent of a high school diploma or less in North America at the time mothers were schooled) and high (General Certificate of Education Advanced or university degree).
Covariates: Potential Factors Influencing Childcare Attendance
We used propensity score weighting to control for factors prior to childcare attendance that might indirectly confound the association between childcare attendance and later academic achievement. These included: birth order; mother’s pregnancy in the 4 years following the child’s birth; paternal presence in the home (2-parent family); maternal age, history of clinical mental health diagnosis, and employment status at child age 2 months; maternal and paternal highest education levels; family adversity from second trimester to child age 4 years (Family Adversity Index questionnaire, cumulative 18-item score18); socioeconomic status (Standard Occupational Classification,19 categorized as middle-high [levels 1-3] and low [levels 4-6]17); and Neighborhood Quality Index.20 Unless stated otherwise, demographic information was reported by the mother and/or father in the second trimester of pregnancy.
Covariates: Factors Concurrent With Childcare Attendance
The following covariates were considered in the multivariable models between childcare attendance and youth’s academic achievement. Although they were not involved in parental decisions to send their child to childcare, they are important characteristics linked to children’s academic achievement, and they are either concurrent with children’s childcare attendance or hypothesized to remain stable over time.
Cognitive abilities were assessed by trained psychologists at child age 8 years. The Wechsler Intelligence Scale for Children (WISC-III)21 short form was used to reduce the likelihood of fatigue affecting performance. A general cognitive abilities score was calculated to account for potential error in any given domain. We also collected child sex as recorded on the birth certificate.
At child age 2 years, mothers answered a 10-item questionnaire capturing involvement on a day-to-day basis (eg, how often do you sing together, slap the child during tantrums). Answers on a 5-point Likert scale ranged from never to every day. The sum of all items was a measure of parenting practices, with higher scores indicating more sensitive and stimulating parenting.
Because KS4 results were not available for all participants, we calculated a sampling weight using inverse probability weighting.22 First, we identified in bivariate analysis (ie, Pearson correlation, χ2 testing, and 1-way analysis of variance using 2-tailed tests) the sociodemographic characteristics that were associated with KS4 result availability (P < .10). Then, we performed a binomial logistic regression where the set of covariates previously identified was linked to KS4 results availability. Missing data on covariates were managed using multiple imputations by chained equations using the Amelia package in R.23 Covariance between children with and without KS4 results was assessed using standardized mean differences, before and after propensity score weighting. All analyses were performed with R version 3.6.1 (The R Foundation), and significance of the multinomial regression coefficient was established at P < .05.
Addressing Potential Social Selection Bias in Childcare Attendance
Propensity score weighting was used to correct bias in parental decisions for childcare attendance, according to variables identified in bivariate analyses (P < .10) on all children in the ALSPAC study, with data collected at least once between birth and age 3 years. Two propensity scores were estimated—one for the 3-level indicator of childcare attendance (low, part-time, or full-time attendance) and another for the 2-level indicator (did or did not attend)—using multinomial and binomial logistic regression, respectively, with the sociodemographic characteristics previously identified in bivariate analysis. We then calculated both scores using inverse probability weighting22 while applying the weight for population representativeness.24 Covariance balance between trajectories was assessed using standardized mean differences before and after propensity score weighting.
Addressing Population Representativeness and Social Selection Bias Simultaneously
To simultaneously control for population representativeness and social selection bias in childcare attendance, we multiplied the 2 weights previously calculated and applied their product to further analyses to provide more robust estimates.25 The population representativeness and the propensity score weights were calculated on the fully imputed sample to avoid case deletion due to missing data on covariates. Subsequent analyses were performed on the subsample with complete data on maternal education, childcare attendance, and academic achievement (n = 8936).
Question 1: Is Childcare Attendance Associated With Academic Achievement?
We used multinomial regression analysis to determine whether childcare attendance was associated with academic achievement at age 16 years while controlling for child’s cognitive abilities and sex and maternal parenting practices.
Question 2: Does Maternal Education Moderate the Association Between Childcare Attendance and Academic Qualification?
Because of low-frequency cells when the 3-trajectory childcare variable was cross-referenced with maternal education and KS4 results, we combined the part-time and full-time trajectories together. We thus used a 2-level childcare variable to investigate the interaction between childcare attendance and maternal education level on child academic achievement. We then performed a stratified analysis to examine associations between childcare attendance and academic achievement within each level of maternal education and derived the probability of belonging to each level of academic achievement.
Question 3: What is the Benefit-Cost Ratio of Childcare Attendance?
We provided a lower bound estimate of the benefit-cost ratio of full-time childcare attendance using childcare annual costs and estimates of the returns of childcare attendance on lifetime productivity.26 While childcare costs apply to all children attending, benefits may not be the same for everyone. Total cost of full-time childcare attendance is estimated at US $17 764 over 4 years27 (4-fold the annual costs of £3250 [US $4547]). Relative to not completing compulsory schooling, total lifetime productivity returns for completing Level 1 GCSE are estimated at £99 655 (US $139 430), while returns for completing Level 2 GCSE are estimated at £104 213 (US $145 808).28,29 The marginal effect of childcare attendance on degree completion is multiplied by total lifetime productivity returns divided by the total cost of childcare. We performed the same analysis according to maternal highest education level. Note that a benefit-cost ratio superior to 1 means that the benefits are higher than the costs engaged. Comparatively, a value between 0 and 1 signifies that the costs are higher than the benefits recovered. The ratio may become negative if the impact of childcare on academic achievement is negative, which implies a financial loss instead of a benefit. The consumer price index was used to adjust all values to 2019 British pounds.
Of the 14 541 mothers in the ALSPAC cohort, 12 117 (83.3%) reported on childcare attendance up to child age 3 years. On average, 3.7%, 5.9%, and 90.4% of children attended childcare full time, part time, and less than 10 hours per week, respectively (Figure 1). Our final cohort included 8936 participants with data on childcare attendance, academic achievement, and maternal education level (Table 1). Of these, 4499 (50.3%) were male.
Population Representativeness and Social Selection Into Childcare
eTables 1 and 2 in the Supplement show participant characteristics by KS4 availability results and childcare trajectory, respectively. eTables 3 to 5 in the Supplement present the binomial and multinomial associations between participant characteristics and KS4 availability results as well as childcare trajectories. Propensity score weighting significantly reduced the impact of missing KS4 data and of social selection into childcare (eFigures 1 to 3 in the Supplement).
Question 1: Is Childcare Associated With Academic Achievement?
We found a positive association between part-time and full-time childcare attendance and academic achievement. Compared with children who did not attend childcare, children who attended part-time childcare had an increased likelihood of obtaining Level 1 or Level 2 GCSE qualifications (Level 1: relative risk [RR], 1.41; 95% CI, 1.16-1.73; Level 2: RR, 1.62; 95% CI, 1.30-2.01). A similar pattern emerged for full-time attendance, where childcare attendance increased the likelihood of obtaining a Level 1 GCSE qualification (RR, 1.88; 95% CI, 1.59-2.23), but it did not reach significance for Level 2 GSCE qualification (RR, 1.15; 95% CI, 0.94-1.39). Higher child IQ was associated with a greater chance of obtaining a GCSE Level 1 or 2 qualification at the end of compulsory schooling, and parenting practices as well as female sex were only associated with Level 2 qualification (Table 2).
Question 2: Does Maternal Education Moderate the Association Between Childcare Attendance and Academic Qualification?
Maternal education moderated the association between childcare attendance and academic achievement (interactive model and stratified analyses; Table 3). Stratified analyses revealed that childcare attendance was associated with academic achievement for those with lower maternal education levels only compared with those who did not attend childcare (Level 1: RR, 1.64; 95% CI, 1.32-2.03; Level 2: RR, 1.57; 95% CI, 1.30-1.90). Presented differently, the probability of not achieving a Level 1 or 2 qualification at the end of compulsory schooling was 28.9% (95% CI, 26.8-31.0) for those who did not attend childcare and whose mothers had lower education levels (Figure 2A). For adolescents of mothers with low education who attended childcare, the probability of not achieving a Level 1 or 2 qualification was 20.3% (95% CI, 18.0-22.8), a 29.8% relative reduction. In contrast, for children of mothers with higher levels of education, childcare attendance was not associated with academic achievement; the probability of not achieving qualification was less than 10%, regardless of childcare attendance (Figure 2B).
Question 3: What Is the Benefit-Cost Ratio of Investing in Childcare Attendance?
Based on the marginal effects of full-time childcare attendance on academic achievement, we found that for each £1 (US $1.40) invested in full-time childcare, the estimated lifetime productivity returns for children achieving a Level 1 and Level 2 GCSE were £1.32 (95% CI, 0.72-1.92; US $1.85; 95% CI, 1.01-2.69) and £0.57 (95% CI, −0.26 to 1.39; US $0.80; 95% CI, −0.36 to 1.94), respectively. For children of mothers with high education, the benefit-cost ratio for achieving a Level 1 and Level 2 were −£0.87 (95% CI, −1.91 to 0.15; US −$1.22; 95% CI, −2.67 to 0.21) and −£0.48 (95% CI, −1.72 to 0.73; US −$0.67; 95% CI, −2.41 to 1.02), respectively. For children of mothers with lower education, the benefit-cost ratios were £1.29 (95% CI, 0.72-1.83; US $1.29; 95% CI, 1.01-2.56) and £1.71 (95% CI, 1.03-2.45; US $2.39; 95% CI, 1.44-3.43) for achieving a Level 1 and Level 2 GCSE qualification, respectively.
We conducted a large, prospective, linkage administrative database study of 8936 children in the ALSPAC cohort, examining the long-term association between childcare attendance and academic achievement at the end of compulsory schooling. We accounted for selection bias into childcare enrollment and used propensity score weighting to ensure that our cohort was representative of the initial ALSPAC sample. We found that for children with lower maternal education levels, childcare attendance (part time or full time) was associated with higher rates of academic achievement (Level 1 and Level 2 GCSE qualification) compared with no childcare attendance. For these children, the benefit-cost ratio of investing £1 (US $1.40) in full-time childcare was £1.71 (US $2.39) when they achieved a Level 2 GCSE qualification. On the other hand, children of mothers with higher education levels did not benefit from childcare attendance in terms of academic achievement at age 16 years, nor was childcare participation cost-effective in this subpopulation.
Our results support the idea that childcare attendance might attenuate the association of socioeconomic status with cognitive and academic trajectories and foster long-term academic achievement for children of educationally disadvantaged families. The enriched environment provided in childcare settings may benefit children’s cognitive abilities30 and early regulation skills,31 which would lead to higher levels of academic adjustment. This association might also be magnified by children’s extracurricular activities attendance during early and middle childhood.32
If anything, the benefit-cost ratios we report are conservative. Our analyses assume that Level 1 and 2 GCSE qualifications are the highest qualifications achievable by participants in the ALSPAC study. However, some participants will likely acquire further qualifications (eg, university degree) to increase lifetime productivity. Hence, the benefit-cost ratios of childcare attendance should increase with time. Second, practically all the costs were included except for the short-term negative health effects of childcare attendance on children and their parents.33 Third and importantly, the benefits of childcare attendance were only partly analyzed. We did not consider socioeconomic benefits, such as increased maternal employment,33 reduction in criminality,8 and improved children’s health, behavior, and socioemotional adjustment in the short-term and long-term.9,10,34
Nonetheless, even after accounting for childcare attendance, social disparities in academic achievement between children of mothers with low and high education levels remain. This result suggests that while subsidizing childcare in the UK for socioeconomically disadvantaged parents might increase childcare attendance and academic achievement, it might not close the academic gap between children of different socioeconomic backgrounds, for at least 2 reasons. First, as childcare is mostly private in the UK and as the quality of care provided by the private sector is likely to be correlated with socioeconomic status,35 children of low socioeconomic status are more likely to attend lower-quality childcare than others.36 The correlation between childcare quality and family income is lower in a state-run childcare or social democratic context, like that of Sweden and Denmark.37 Second, accessing childcare subsidies and finding a good-quality childcare require administrative literacy38—that is, parents’ resources and capacities to effectively interact with public administrations, community agencies, and governmental officials—to access targeted benefits.39 This additional administrative burden for disadvantaged parents may discourage them from enrolling their children in subsidized childcare. To illustrate the influence of childcare cost on children’s attendance, we closely examined the diminution of childcare hours presented in Figure 1 and found that the diminution was more prominent among children of mothers with lower levels of education (eTable 6 in the Supplement). This observation supports the idea that childcare cost is an important barrier for attendance and that directly subsidizing childcare services might be a more suitable option for offering high-quality services and facilitating access to low socioeconomic families.37
Strengths and Limitations
Our large sample size enabled an investigation into the moderating effect of maternal education. Thanks to the richness of the assessments at ALSPAC clinics, we could estimate propensity scores from an extensive set of children, parents, and sociodemographic characteristics. Propensity score weighting is stronger for controlling for social selection than the covariate adjustments reported in previous childcare studies.13 Third, the proportion of children in each trajectory was similar to the one reported in a British national report at that time.40 Fourth, linkage between participants in the ALSPAC study and participants in the UK National Pupil Database allowed us to access standardized examination results, ensuring excellence in the accuracy and reliability of our results.
Our study had limitations. One limitation is that we conducted an observational study. We tested association, not causation. However, the use of propensity scores increased the comparability of children exposed to various childcare conditions and populational representativeness, thereby assuring confidence in the interpretation of findings. Nonetheless, propensity score weighting does not account for potential unmeasured confounding factors.22 Second, due to low-frequency cells, we had to combine part-time and full-time childcare trajectories to investigate maternal education as a moderator, precluding an analysis of childcare intensity. Third, we had an important number of missing data on academic achievement, childcare attendance, and maternal education variables. However, the computation of a population representativeness allowed us to lessen the impact of this limitation. Fourth, the lack of assessment of childcare quality is inherent in the data collection of the ALSPAC study.
Promoting universal childcare with facilitated access for children of lower socioeconomic backgrounds deserves to be considered as a way to reduce the intergenerational transmission of low social capital.
Accepted for Publication: March 10, 2021.
Published Online: June 7, 2021. doi:10.1001/jamapediatrics.2021.1192
Corresponding Author: Sylvana M. Côté, PhD, Sainte-Justine’s Hospital Research Center, 3175, Côte Sainte-Catherine, Étage A, Local A-568, Montréal, QC H3T 1C5, Canada (email@example.com).
Author Contributions: Ms Larose had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Larose, Côté.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Larose, Côté.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Larose, Haeck, Barker.
Obtained funding: Côté.
Administrative, technical, or material support: Côté.
Study supervision: Haeck, Ouellet-Morin, Côté.
Conflict of Interest Disclosures: Dr Côté has received grants from the Canadian Institute for Health Research during the conduct of the study. No other disclosures were reported.
Funding/Support: The UK Medical Research Council and Wellcome Trust (grant 102215/2/13/2) and the University of Bristol continue to provide core support for the Avon Longitudinal Study of Parents and Children. This research was specifically funded by the Department for Education and Skills grant EOR/SBU/2002/121. A comprehensive list of grants funding is available at http://www.bristol.ac.uk/alspac/external/documents/grant-acknowledgements.pdf. Ms Larose was supported by the Fonds de Recherche du Québec–Santé (FRQS).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank Danielle Buch (Medical Writer and Researcher, University of Montreal, Montreal, Quebec, Canada) for critical revision and extensive substantive editing of the manuscript and tables. We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting, and the whole ALSPAC team, including interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses. Ms Buch was compensated for her work as a medical writer.
MG. High school dropouts in emerging adulthood: substance use, mental health problems, and crime. Community Ment Health J
. 2015;51(3):289-299.PubMedGoogle ScholarCrossref
K. How does early childhood care and education affect cognitive development? an international review of the effects of early interventions for children from different social backgrounds. Early Child Res Q
. 2010;25(2):140-165. doi:10.1016/j.ecresq.2009.11.001Google ScholarCrossref
WS. Meta-analysis of the effects of early education interventions on cognitive and social development. Teachers Coll Rec
. 2010;112(3):579-620.Google Scholar
WR; Family Life Project Key Investigators. Household chaos and children’s cognitive and socio-emotional development in early childhood: does childcare play a buffering role? Early Child Res Q
. 2016;34:115-127. doi:10.1016/j.ecresq.2015.09.003PubMedGoogle ScholarCrossref
HD. Concern over internal, external, and incidence validity in studies of child-care quantity and externalizing behavior problems. Child Dev Perspect
. 2017;11(2):133-138. doi:10.1111/cdep.12224Google ScholarCrossref
et al. Cohort profile: the ‘children of the 90s’—the index offspring of the Avon Longitudinal Study of Parents and Children. Int J Epidemiol
. 2013;42(1):111-127. doi:10.1093/ije/dys064PubMedGoogle ScholarCrossref
A, Van Staa
J. Socio-economic position and childhood multimorbidity: a study using linkage between the Avon Longitudinal Study of Parents and Children and the General Practice Research Database. Int J Equity Health
. 2013;12:66. doi:10.1186/1475-9276-12-66PubMedGoogle ScholarCrossref
Office of Population Censuses & Surveys. Standard Occupational Classification. HM Stationery Office; 1991.
J. WISC‐III UK Wechsler Intelligence Scale for Children: UK Manual. 3rd ed. The Psychological Corporation; 1992.
EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med
. 2015;34(28):3661-3679. doi:10.1002/sim.6607PubMedGoogle ScholarCrossref
ED. Promoting better functioning among children exposed to high levels of family adversity: the protective role of childcare attendance. J Child Psychol Psychiatry
. Published online August 21, 2020. doi:10.1111/jcpp.13313PubMedGoogle Scholar
KM. Cumulative and differential effects of early child care and middle childhood out-of-school time on adolescent functioning. Child Dev
. 2020;91(1):129-144. doi:10.1111/cdev.13136PubMedGoogle ScholarCrossref
DA. The Benefits and Costs of Head Start. National Bureau of Economic Research; 2007.
SM. La qualité des services à la petite enfance: Résultats de l’Étude Longitudinale sur le Développement des Enfants du Québec (ELDEQ). Éducation Francophonie
. 2005;2:7-27.Google Scholar
P. Public Funding of Early Years Education in England: An Historical Perspective. London School of Economics and Political Science; 2016.