Nurturing Parents and Children: an early stimulation and father's engagement intervention in Indonesia

Our first intervention will adapt and further test group-based Reach up and Learn (RL) to target 0 to 3 year olds from poor families in Indonesia. Poor families are identified as those eligible for the Indonesian CCT program.
Our second intervention corresponds to the adaptation of a complementary intervention that targets fathers' engagement in co-parenting. We will adapt activities from existing evidence-based programs. We will test if an intervention that targets mothers and fathers can produce complementarities to enhance child development.

Our first innovation is to adapt and further test group-based Reach up and Learn (RL) to target 0 to 3 year olds from poor families in Indonesia. RL is a proven ECD program that was first implemented in the 1980s in Jamaica using a structured early stimulation curriculum delivered through weekly home visits by community health workers (CHWs) for 2 years. Since the original individual approach can be challenging to scale-up, RL has been adapted to a group setting where it is delivered in small groups led by trained facilitators. Each group has about 8 to 12 children and caregivers, and they usually meet for 18 months. The sessions teach mothers how to promote child development using low-cost and recycled materials. Most group adaptations include monthly home-visits to ensure parents’ engagement with the curriculum at home. An evaluation of a group-based RL shows better home stimulation and cognitive and socioemotional development (Attanasio et al., 2018). Also, the group-based approach will allow us to explore social capital and network as parents’ coping strategy (peer support) in ECD, which has not been studied.
RL has been adapted and implemented in other countries including Bangladesh, India, Colombia and Peru. The curriculum adaptation incorporated local games and songs and changed pictures in the books to reflect children’s environment (Grantham McGregor and Smith, 2016). The Reach Up toolkit includes a weekly curriculum; training manual with demonstration videos that were filmed in Jamaica, Peru, and Bangladesh; and a planning manual that helps countries tailor the program to their specific needs. We will further adapt and tailor RL to reflect children’s environment in Indonesia. We will address the cultural appropriateness of the intervention and foster current parenting practices that are culturally appropriate with the input of two local ECD developmental psychologists (Smith et al., 2018; Tomlinson and Andina, 2015; Kline, 2015; Callaghan et al, 2011; Rogoff et al, 2007).

Our second innovation is the adaptation of a complementary intervention. RL mainly focuses on children's cognitive stimulation, but its impacts may be larger if complemented with components that empower the family holistically to improve the child’s home environment. RL targets children’s primary caregiver, usually mothers, to increase inputs for child development, but family interactions and resource allocation depend on both parents. So, we will adapt activities from evidence-based programs like Father Support Programme, Parents with Respectability and Program P that promote positive socio-cultural norms and behaviors around co-parenting, addressing topics like bonding/attachment, positive discipline, and gender norms and socialization (Barker et. al., 2009; Baykal et. al., 2019; Siu et. al., 2017).

Our third innovation is to use existing infrastructure and personnel in the community as the delivery mechanism for cost-effectiveness, scalability, and grassroots support. We will assess potential delivery of our proposed interventions by family wellbeing development (PKK) cadres, young family development (BKB) cadres, mothers’ class facilitators, and ECD teachers. These facilitators already operate in the community and their support will be key to ensuring community buy-in and grassroots support for the intervention. Additionally, our partnership with the government is ideal to ensure policy goal alignment and identify the initiative that the intervention can piggy-back. In assessing the delivery method, we will consider the potential crowd-out of their current activities.

Conceptual framework
Our theory of change combines theoretical frameworks in developmental psychology and economics. From developmental psychology, the family system theory emphasizes how parenting and child development involve multidirectional relations and interactions between mothers, fathers and children, which are dependent on one another (Minuchin, 1985; J. Jeong et al., 2019). Another key framework is the social learning theory (Bandura, 1977), which theorized that parents and children learn through observation and practice. The ecological systems theory (Bronfenbrenner, 1986; Bronfenbrenner and Morris, 2007) postulates that the entire ecological system in a child’s and his/her environment influences his/her growth and development. This theory motivates the design of our proposed intervention by enhancing the mesosystem and exosystem: empowering the family and using community resources, through trained facilitators and group-based sessions, to impact child learning. Thus, we expect that our proposed intervention will improve family interactions and environment to boost child development.
From economics, our study is based on the human capital production theory and the importance of early investments (Cunha and Heckman, 2007; Becker 1993; Becker 2009). Child development requires inputs from mothers, fathers, and the child’s environment, including parental knowledge and time, material resources, and other community infrastructure. Human capital formation is dynamic across the life cycle, thus early life experiences have long-term consequences in later outcomes. Resource allocation for children also involves intrahousehold bargaining theory (Browning and Chiappori, 1998). Research shows that fathers have more decision-making power than mothers, but mothers have higher preferences on child wellbeing (Lundberg, Pollak, and Wales 1997). This tradeoff is more pronounced in LMICs where resources are limited and bargaining power is more imbalanced (Jayachandran, 2015; Bjorkman and Jayachandran, 2017). Thus, an intervention that targets mothers and fathers can produce complementarities to enhance child development.
Our pilot study will use a mixed-methods approach to assess the potential impacts of multifaceted interventions with these arms: C: Control, T1: RL, T2: RL+father's involvement. To address nutrition and incentivize participation, T1-T2 will include nutrition supplementation. Our pilot will provide foundational evidence for a large-scale study that will further rigorously test multiple impacts and propose a path to scale with government partnerships.

2021-12-01

2022-06-30

Improved family inputs and home environment should support child development, and they are key first-order outcomes of our proposed intervention. To capture family inputs, enumerators will record parenting attitudes and items from UNICEF’s Family Care Indicators (Kariger et al., 2012) to capture the quality of the home environment. To measure parental knowledge (mothers and fathers) about child development, we will select items from the Knowledge of Infant Development Inventory (KIDI) (MacPhee, 1981). We will also assess parental self-efficacy using the Self-efficacy in the Caregiver Role Test– Modified (Pedersen et al., 1989; Porter and Hsu, 2003), which evaluates parental feelings while taking care of their children.

We also expect that our proposed interventions will improve children's human capital outcomes. For children’s physical health outcomes, we will measure height and weight. To capture children’s development, we will use the Ages and Stages Questionnaire.

We will also measure parents’ well-being using the Center for Epidemiologic Studies Depression Scale (CES-D) scale, bargaining power, knowledge on child development and maltreatment, and social networks. To measure women’s empowerment, we will rely on the existing body of validated instruments and behavioral games developed and tested by researchers in economics and psychology. We will review and adapt the questions compiled by Glennester, Walsh, and Diaz-Martin (2017). Overall, we will pilot and assess the reliability of the instruments, taking into account the Indonesian context. The survey instruments will also include demographic characteristics following the Indonesian Family Life Survey (Strauss et al, 2016).

Please see the previous item.

Our project will conduct both a small scale pilot RCT (to pre-test the interventions) and a large-scale RCT.
Prior to the pilot, we will perform a feasibility study using a mixed-methods approach to examine potential facilitators to deliver our proposed interventions taking into account how we can utilize existing infrastructure and personnel for cost-effectiveness and scalability.
In the small scale pilot, villages will be randomly assigned to either the control group or the following treatment arms. One treatment arm will be the adapted RL curriculum delivered in a group-based setting, with 8-12 families per group through weekly meetings ofr 6 months and one home visit per month. The other treatment arm will be the adapted RL complemented with the father co-parenting component, which will be delivered through weekend sessions for 3 months. Then villages in each treatment arm will be randomized to receive the intervention by two of the most promising facilitators, who will be identified from the mixed-methods analyses.

Based on the findings from the pilot study, we will identify the most promising facilitator and test our proposed interventions at a larger-scale. In the large-scale RCT, villages will also be randomized in to the three groups described above, however, the interventions will be delivered for 18 months.

The feasibility study part of our project will conduct a mixed-methods approach to examine potential facilitators to deliver our proposed interventions taking into account how we can utilize existing infrastructure and personnel for cost-effectiveness and scalability. In particular, we will explore the potential delivery of group-based RL by: PKK cadres, BKB cadres, mothers’ class facilitators, ECD teachers and INEY facilitators using FGDs and in-depth interviews.

For the pilot RCT, we will determine a list of eligible villages in two provinces and randomly select 24 villages. Then each of the 24 villages with be assigned to each of the three groups using randomization done in office by a computer.
For the large scale RCT, we will use triplet matching to randomly allocate villages to each of the intervention arms.

The unit of randomization is the village

Yes

For the small pilot RCT, we will have 24 villages.
For the large-scale RCT, the treatment will be clustered, with 100 villages in each arm.

For the smal pilot RCT, the sample size is 720 poor households with children ages 0-3 in 24 villages (according to administrative data, potential villages have an average of 30 CCT eligible families with children ages 0-3). For the large-scale RCT, the sample size is: 9,000 poor families in 300 villages.

For the small pilot RCT: 8 villages in the control group. 8 villages in T1: RL only (4 villages delivered by Facilitator 1 and 4 villages delivered by Facilitator 2). 8 villages in T2: RL plus father's involvement (4 villages delivered by Facilitator 1 and 4 villages delivered by Facilitator 2).
For the large scale RCT: 100 villages in the control, 100 in T1 and 100 villages in T2

The main goal of the pilot RCT is to generate proof-of-concept evidence to design the large-scale RCT. For the large-scale RCT, assuming an intra-cluster correlation of 0.1-0.2 (based on data from other developing countries for similar outcomes), 80% power, 5% type I error and 10% attrition rate per cluster, the minimum detectable effect size is 0.15-0.2 SDs for parenting and children outcomes.