The First 2,000 Days: Investing in Children’s Skills Through Early Intervention

Main article

Children from poorer backgrounds typically have lower cognitive and socio-emotional skills due to differences in the quality of their environments, and these early discrepancies can feed into worse outcomes later in life. This column details the results of a randomized control trial of the Preparing for Life program, which offered high-intensity support to parents in Ireland. The intervention generated large improvements in children’s cognitive, socio-emotional and behavioral skills. The results confirm that investment in parenting during the first 2,000 days of life can reduce the socioeconomic gradient in children’s skills, and suggest that these interventions can even be effective in countries that already have generous social welfare policies.

Children from poorer backgrounds typically have lower cognitive and socio-emotional skills. This is due to differences in the quality of the environment, with disadvantaged children facing lower family incomes, higher levels of stress, poorer parenting practices, and less academic stimulation.[1] Consequently, living in disadvantaged circumstances early in life is frequently associated with poorer health, education, and labor market outcomes (Cunha et al. 2006; Heckman, 2006; Almond et al. 2018).

These inequalities can be reduced through targeted early childhood intervention programs.[2] Such programs are effective from both a biological perspective and an economic perspective; brains are more malleable early in life (Knudsen et al. 2006) and early investment means that returns are accrued over a longer period (Heckman and Kautz, 2014). Much of the evidence in this field comes from the US and developing countries, which has sometimes led the documented socio-economic inequalities in children’s skills to be ascribed to low levels of redistribution and lack of resources, respectively. In fact, this is a global issue. Inequalities persist in Europe too, despite more redistributive social policies.

With this in mind, our study explored the role of intensive investment in parenting from pregnancy until entry into formal schooling in a highly disadvantaged community in Dublin, Ireland. The trial, known as Preparing for Life (PFL), began in 2008. It was based on the idea that providing supports to parents during this critical period of development would permanently change children’s skills and abilities. Theoretically, the program is rooted in the skill formation model (Cunha and Heckman, 2007), which states that developing children’s basic skills helps them to develop more advanced skills through a process of self-productivity, and this in turn raises the effectiveness of later investments, through a process of dynamic complementary. Thus, if the pregnancy and infancy periods are critical for boosting children’s brain development, and parenting and the quality of the home environment is strongly implicated in the development of children’s skills, then intervening early in life and focusing on parents may be an effective strategy.

The intervention: a randomized control trial of the Preparing for Life program

The PFL program was developed by the Northside Partnership as part of the Irish Government and The Atlantic Philanthropies’ Prevention and Early Intervention Program, which aimed to break intergenerational cycles of poverty. The program wanted to improve the health and development of disadvantaged children by supporting parents around parenting practices and identifying developmental milestones, and encouraging them to provide more stimulating environments for their children. We evaluated the program using a randomized control trial, whereby parents were assigned to a high or low treatment group. Figure 1 describes the program. The high treatment group received the following three intensive parenting supports by trained mentors who were specifically hired to deliver the PFL program:

• 5-year home visiting program;
• Baby massage course;
• Triple P Positive Parenting Program.

The fortnightly home visits started during pregnancy and continued until the child started primary school. The home visits were based on a curriculum of over 200 Tip Sheets which were delivered using role modelling, coaching, discussion, encouragement, and feedback. In the first year, the high treatment group also received 5 two-hour baby massage sessions to encourage early communication between parents and infants. In the second year, they also received the Triple P Positive Parenting Program (Sanders et al. 2003) which is a highly structured parenting intervention focused on positive parenting practices. The program included 5 two-hour group sessions and 3 phone calls. On average, the high treatment group received 50 home visits, 62% participated in baby massage, and 43% took part in the Triple P program.

Both the high and low treatment groups received a set of developmentally appropriate toys and books, support to participate in community-based social events and public health workshops, newsletters, birthday cards, and access to a support worker who could direct them to other services.

Figure 1 Design of the PFL Program

All pregnant women who lived in the PFL catchment area between the January 2008 and the August 2010 were eligible to join the program. We recruited parents in the local maternity hospitals and/or through self-referral. On average, parents were ~20 weeks pregnant when they joined. We used a computerized randomization procedure to assign 115 participants to the high treatment group and 118 to the low treatment group. We found that there were no statistically significant differences between the groups on the majority of baseline measures, indicating the success of the randomization procedure. Using a randomized control trial means we can be confident that any differences in children’s outcomes at the end of the trial can be attributed to the PFL program and not any underlying characteristics of the families.

We collected data at baseline and during several other follow-up surveys. We assessed the children’s cognitive development using maternal reports from the Ages and Stages Questionnaire and the Developmental Profile cognitive development score, and by direct assessment using the British Ability Scales II: Early Years Battery. We assessed their socio-emotional development using the Child Behavior Checklist, the Brief Infant-Toddler Social and Emotional Assessment, and the Strengths and Difficulties Questionnaire. 

Follow-up data was collected from 166, 150, 147 and 134 participants at 24, 26, 38, and 51 months, respectively. Attrition was largely equivalent across both groups and the groups remained balanced on baseline characteristics. In order to account for potential attrition bias, however, we estimated all results using permutation tests adjusted for Inverse Probability Weighting.

Impact on cognitive skills

We found that the PFL program had a large impact on children’s cognitive development.  In particular, we found a 10-point gap in the IQ scores of the children in the high and low treatment groups. The program shifted the entire distribution of children’s cognitive scores as shown in Figure 2. 

Figure 2 Distribution of cognitive scores for the high and low treatment groups

The program also reduced the proportion of high treatment children scoring below average and increased the proportion of children scoring above average in terms of their cognitive scores. The effects were more pronounced for children with the lowest ability (see Figure 3). Improvements were found across all types of cognitive skill including spatial ability, pictorial reasoning, and language ability. This means that the children in the high treatment group were better at solving problems, as well as understanding numbers and language. These results, which were based on direct assessment of the children’s skills at age 4, were similar to those reported by parents from age 2 onwards. 

Figure 3 Percentage Scoring Below Average on Cognitive Tests in the High and Low Treatment Groups

Note: All differences are statistically significant based on one-tailed (right-sided) conditional IPW-adjusted permutation tests with 100,000 replications.

Impact on socio-emotional and behavioral skills

The program also improved children’s socio-emotional and behavioral skills, though the effect was weaker. Children in the high treatment group were less likely to be at risk of clinically significant externalizing problems at every age (see Figure 4), and were less at risk of internalizing problems at ages two and four. This means the high treatment group children were less aggressive and less anxious compared to the low treatment group. The program also improved children’s pro-social behavior, which is an indicator of the child’s ability to make friends, though it had no impact on their peer problems, which is an indicator of bullying behavior.

Figure 4 Percentage at Risk of Clinically Significant Externalizing Problems in the High and Low Treatment Groups

Note: All differences are statistically significant based on one-tailed (right-sided) conditional IPW-adjusted permutation test with 100,000 replications.

Effects for different groups

We did not find a gender difference in the level of benefits, but the program had a larger impact for first-born children and for children from households with more financial, cognitive, and psychological difficulties.

Comparing treatment results to a national sample

As shown in figure 5, a comparison of the PFL treatment groups to a large nationally representative sample of Irish children (n=9,000) shows that the program narrowed the socioeconomic gap on some dimensions of children’s skills. While the national sample consistently scored higher than the low treatment group, the scores for the high treatment group were similar to the national sample in terms of their naming vocabulary scores and pro-social behavior scores.

Figure 5 Comparing the Scores of the Treatment Groups to a National Sample 

Note: The national sample had statistically significantly better scores than the low treatment group on all measures.  The national sample had statistically significantly better scores than the high treatment group on their picture similarity and peer problem scores, yet there was no difference for naming vocabulary and prosocial behavior. Estimated using a two-tailed unpaired t test with weights applied.

Sustained, early investment in parenting can reduce inequalities in children’s skills

Our study finds that early intervention programs can be effective in countries such as Ireland, which already provide generous social welfare policies and comprehensive supports for women and children as standard practice. The size of the effects was large and may be credited to PFL’s prenatal start, its length, its multiple connected supports, and its inclusive eligibility criteria. PFL started earlier and was longer in duration than most other early intervention programs. The PFL mentors worked with families for a substantial and critical period of their children’s lives, which resulted in a strong, high-quality mentor-parent relationship. The multi-component nature of the program may also have helped to engage families who favored one form of support over another. The program was based on the idea that providing support to disadvantaged families would increase their knowledge of appropriate parenting practices and change their attitudes and behaviors. As shown in previous studies (e.g. Doyle et al. 2015; Doyle et al. 2017; O’Sullivan et al. 2017), the PFL program was effective in supporting such changes and these incremental improvements in parenting may account for the significant improvements in children’s skills.

Much of the policy focus on early childhood interventions is based on the long-run findings from preschool programs, which target children directly and start at age three (e.g. the Perry Preschool program). Interventions that target parents and/or start in pregnancy are less intensive and therefore less costly, and they may also improve the development of all children in the family. This lower cost, and the possibility of spillover effects on siblings, may improve the economic return to programs such as PFL. Cost-benefit analyses of the center-based Head Start program by Ludwig and Philips (2007) and Deming (2009) find that improvements of 0.06-0.20 of a standard deviation in cognitive skills are enough to satisfy cost-benefit tests based on an average program cost per child of ~$US7,000. Therefore, if the significantly larger effects identified for PFL (0.77 of a standard deviation) at a cost of ~$US11,250, translate into future financial gains for program participants and society more broadly, the program is likely to generate similar positive returns.

Our study suggests that sustained investment in parenting during the first 2,000 days of life may generate large improvements in children’s ability, which are likely to have long-term implications for their development and contributions to society. More work needs to be done to assess the results of the PFL program in other contexts, but this study suggests that it may provide an effective vehicle for reducing the socioeconomic gradient in children’s skills.

This article summarizes: Doyle (2020) ‘The First 2,000 Days and Child Skills’. Journal of Political Economy, 128(6).

Orla Doyle is at University College Dublin

[1] See Becker (1965), Del Boca et al. (2014), Bradley et al. (1989), Carneiro and Heckman (2003), Cobb-Clark et al. (2019), Lupien et al. (2001) and Miller et al. (2014).

[2] See reviews by Aboud and Yousafzai (2015), Avellar et al. (2016), Britto et al. (2015), Miller et al. (2011), Mingebach et al. (2018), Peacock et al. (2013), Sweet and Appelbaum (2004), and Yap et al. (2016).