Lack of adequate preparation for primary school through pre-primary education is one of the key risk factors for poor performance in primary school (Behrman et al., 2006).* Thus, a popular approach for trying to improve outcomes in children has to do with increasing enrollment in preschool programs, and/or trying to improve the quality of existing programs. Children in low-resource settings are less likely to attend school, and they are less likely to learn when they are in the school setting – partly because they are unprepared for school when they get there.
Another approach to promoting early child development revolves around support of parents. Most existing studies of home-visiting studies in low- or middle-income countries have been smaller efficacy trials, though some recent papers have examined programs at scale (see, for example, (Yousafzai et al., 2014) and (Chang et al., 2015)). A study set in Mexico utilized the existing structure of the country’s conditional cash transfer program (Prospera) to deliver group-based parenting support and showed positive effects on child development (Fernald et al., 2017), as did a similar study in Colombia, which used a home-visiting approach (Attanasio et al., 2014). There were also modest, positive effects of an adult literacy and parental participation program in India on outcomes in older children aged 5-8 years old (Banerji et al., 2015).
In Özler et al. (2017), using a cluster-randomized controlled trial, we test the effectiveness of teacher training at informal preschools in rural Malawi on early childhood development and primary school readiness and we assess whether such school-based interventions are more effective when combined with group-based parenting training. Because the newly trained teachers and mentors in the community deliver the group-based parenting training, the approach is easily and cheaply scalable. Given the consistent effectiveness of parenting support in the promotion of early child development, and the widespread use of informal preschools within LMICs (Garcia et al., 2008), our findings have the potential for broad policy relevance within a context of extreme poverty and limited government resources.
We focus on Community-Based Childcare Centers (CBCCs) in Malawi, which are estimated to serve 580,000 children in approximately 5,000 communities (Drouin and Heymann 2010). In its wish to support these CBCCs, instead of setting up a parallel formal preschool sector, the government decided to improve the supply of play and learning materials in these centers and strengthen the capacity of teachers through additional training and mentoring to support children’s early development and learning. In addition to a comparison group that received only a standard kit of supplies from UNICEF (play and learning materials), two key interventions we studied were a second arm that received teacher training and mentoring, and a third one, in which we complemented the school-based teacher-training program with a 12-module, group-based, parenting education program for the primary caregivers of the children enrolled at the CBCC.
The study sample includes 2,120 children at 199 CBCCs in four districts of Malawi, aged 36-61 months at baseline. At the 18-month follow-up, we find that primary child outcomes improved, but only in the treatment group that received the integrated intervention – with teacher training and parenting education. In this group, children had significantly higher scores in an assessment of language skills and they exhibited more prosocial behaviors when compared with both the control group and the teacher training only group (both on the order of 0.2-0.3 SD). The gains at the child level from the added parenting education were accompanied by substantial improvements in family care indicators, e.g. how many times a day their primary caregivers read to their children or played with them.
Teacher training alone did not improve children’s outcomes, despite significant improvements relating to the classroom environment and teacher behaviors. Furthermore, a rich battery of child assessments, conducted 36 months after baseline, showed no treatment effects among the 6-8 year-old children in any treatment arm, indicating a substantial fadeout of program impacts in the integrated intervention arm.
Our findings echo those of Chile’s Un Buen Comienzo, which also found that improvements in classroom quality did not translate into improvements in child-level outcomes at the end of an intense, two-year teacher training intervention. Being assigned to higher quality classrooms (i.e. to better teachers in the absence of an intervention, as measured via classroom observations) in kindergarten has been recently shown to modestly increase math, language, and executive function test scores among children (Araujo et al., 2016). In contrast, the study by (Yoshikawa et al., 2015) in Chile and our study in Malawi highlight the difficulty of converting program-induced improvements in classroom quality into better child outcomes.
Our trial has incorporated parenting support into the context of a preschool-based quality improvement intervention, two main interventions that have not previously been tested together. While we are unable to speak to the cost-effectiveness of parenting support alone, we found promising evidence that this approach can improve child outcomes over and above teacher training – at least in the short run.
In spite of significant improvements to the classroom environment, there were no benefits of teacher training alone in terms of child-level outcomes. A plausible reading of our findings is that the associations between classroom quality and child outcomes are generally small in magnitude (Burchinal et al., 2008; Araujo et al., 2016), meaning that even moderate to large impacts on classroom quality, such as those reported in our study, may not translate to statistically significant impacts on child assessments (Yoshikawa et al., 2015). This interpretation is also supported, somewhat, by a more recent study by Wolf, Aber, and Behrman (2017) (see a brief summary of one-year findings here, I could not find a full paper of short- or longer-term findings online), who find that in-service teacher training in kindergarten has a positive but small (around 0.1 SD) effects on primary school readiness. Furthermore, these small effects are undone by an additional parental awareness intervention, which, it should be noted, is different than a parenting education intervention.
The integrated intervention of teacher training and parenting education may have reinforced messages at home that had been presented in school, suggesting a potential pathway by which benefits to language and socioemotional skills could have occurred in the short run. In fact, this is the only trial arm in which numeracy, literacy, and problem-solving activities increased both in the classroom and at home, which might explain the significant moderate effects on language skills in the short run. To avoid the fadeout in this group, perhaps regular refreshers and classroom coaching sessions could have supplemented a longer program, or children could have gone to better primary schools with improved coordination of curricula between preschool and primary school.
* All citations can be found in Özler et al. (2017).
Another approach to promoting early child development revolves around support of parents. Most existing studies of home-visiting studies in low- or middle-income countries have been smaller efficacy trials, though some recent papers have examined programs at scale (see, for example, (Yousafzai et al., 2014) and (Chang et al., 2015)). A study set in Mexico utilized the existing structure of the country’s conditional cash transfer program (Prospera) to deliver group-based parenting support and showed positive effects on child development (Fernald et al., 2017), as did a similar study in Colombia, which used a home-visiting approach (Attanasio et al., 2014). There were also modest, positive effects of an adult literacy and parental participation program in India on outcomes in older children aged 5-8 years old (Banerji et al., 2015).
In Özler et al. (2017), using a cluster-randomized controlled trial, we test the effectiveness of teacher training at informal preschools in rural Malawi on early childhood development and primary school readiness and we assess whether such school-based interventions are more effective when combined with group-based parenting training. Because the newly trained teachers and mentors in the community deliver the group-based parenting training, the approach is easily and cheaply scalable. Given the consistent effectiveness of parenting support in the promotion of early child development, and the widespread use of informal preschools within LMICs (Garcia et al., 2008), our findings have the potential for broad policy relevance within a context of extreme poverty and limited government resources.
We focus on Community-Based Childcare Centers (CBCCs) in Malawi, which are estimated to serve 580,000 children in approximately 5,000 communities (Drouin and Heymann 2010). In its wish to support these CBCCs, instead of setting up a parallel formal preschool sector, the government decided to improve the supply of play and learning materials in these centers and strengthen the capacity of teachers through additional training and mentoring to support children’s early development and learning. In addition to a comparison group that received only a standard kit of supplies from UNICEF (play and learning materials), two key interventions we studied were a second arm that received teacher training and mentoring, and a third one, in which we complemented the school-based teacher-training program with a 12-module, group-based, parenting education program for the primary caregivers of the children enrolled at the CBCC.
The study sample includes 2,120 children at 199 CBCCs in four districts of Malawi, aged 36-61 months at baseline. At the 18-month follow-up, we find that primary child outcomes improved, but only in the treatment group that received the integrated intervention – with teacher training and parenting education. In this group, children had significantly higher scores in an assessment of language skills and they exhibited more prosocial behaviors when compared with both the control group and the teacher training only group (both on the order of 0.2-0.3 SD). The gains at the child level from the added parenting education were accompanied by substantial improvements in family care indicators, e.g. how many times a day their primary caregivers read to their children or played with them.
Teacher training alone did not improve children’s outcomes, despite significant improvements relating to the classroom environment and teacher behaviors. Furthermore, a rich battery of child assessments, conducted 36 months after baseline, showed no treatment effects among the 6-8 year-old children in any treatment arm, indicating a substantial fadeout of program impacts in the integrated intervention arm.
Our findings echo those of Chile’s Un Buen Comienzo, which also found that improvements in classroom quality did not translate into improvements in child-level outcomes at the end of an intense, two-year teacher training intervention. Being assigned to higher quality classrooms (i.e. to better teachers in the absence of an intervention, as measured via classroom observations) in kindergarten has been recently shown to modestly increase math, language, and executive function test scores among children (Araujo et al., 2016). In contrast, the study by (Yoshikawa et al., 2015) in Chile and our study in Malawi highlight the difficulty of converting program-induced improvements in classroom quality into better child outcomes.
Our trial has incorporated parenting support into the context of a preschool-based quality improvement intervention, two main interventions that have not previously been tested together. While we are unable to speak to the cost-effectiveness of parenting support alone, we found promising evidence that this approach can improve child outcomes over and above teacher training – at least in the short run.
In spite of significant improvements to the classroom environment, there were no benefits of teacher training alone in terms of child-level outcomes. A plausible reading of our findings is that the associations between classroom quality and child outcomes are generally small in magnitude (Burchinal et al., 2008; Araujo et al., 2016), meaning that even moderate to large impacts on classroom quality, such as those reported in our study, may not translate to statistically significant impacts on child assessments (Yoshikawa et al., 2015). This interpretation is also supported, somewhat, by a more recent study by Wolf, Aber, and Behrman (2017) (see a brief summary of one-year findings here, I could not find a full paper of short- or longer-term findings online), who find that in-service teacher training in kindergarten has a positive but small (around 0.1 SD) effects on primary school readiness. Furthermore, these small effects are undone by an additional parental awareness intervention, which, it should be noted, is different than a parenting education intervention.
The integrated intervention of teacher training and parenting education may have reinforced messages at home that had been presented in school, suggesting a potential pathway by which benefits to language and socioemotional skills could have occurred in the short run. In fact, this is the only trial arm in which numeracy, literacy, and problem-solving activities increased both in the classroom and at home, which might explain the significant moderate effects on language skills in the short run. To avoid the fadeout in this group, perhaps regular refreshers and classroom coaching sessions could have supplemented a longer program, or children could have gone to better primary schools with improved coordination of curricula between preschool and primary school.
* All citations can be found in Özler et al. (2017).
Join the Conversation