Seminar Summary – Parental Beliefs and Parental Investment: Evidence from Colombia
On May 8, the Human Capital Initiative hosted Orazio Attanasio for the final session of the Human Capital Initiative Spring Research Seminar Series with a presentation on “Parental Beliefs and Parental Investment: Evidence from Colombia.” Attanasio discussed his published paper, which analyzed the results of a parenting stimulation program implemented in Colombia. Attanasio is the Cowles Professor of Economics at Yale University.
Attanasio’s study, completed in collaboration with Flávio Cunha (Rice University) and Pamela Jervis (Universidad de Chile), aimed to evaluate the importance of maternal subjective beliefs about child skill formation in determining parental investments in child development. Models indicate that parental investment depends on resources, preferences and parents’ beliefs about the process of child development. Attanasio’s work specifically focused on how to measure these subjective beliefs, validate the measures used and estimate models of parental behavior based on these beliefs. The main research questions included: Do parents understand the process of child development? And, do parents understand the role played by stimulation and interactions?
The study utilized data taken from a large clustered Randomized Control Trial (RCT) whose results were published in the British Medical Journal in 2013. The target population of this original study was low-income households with children aged 12-24 months. In this program, home visitors paid weekly visits to randomly chosen households to improve mother-child interactions and other maternal behaviors that foster the development of children’s cognitive and non-cognitive skills. This Colombia parenting stimulation program saw significant results: most importantly, increased child development outcomes. The results indicated that this increase could be attributed almost exclusively to parental investment. The program significantly shifted the distribution of parental investment, leading to child development impacts.
For researchers, the next step after obtaining these results was to understand why parental investment increased. They hypothesized that this could be caused by changing parental perceptions of the technology of skill formation; however, the technology of skill formation was not affected by the program. They decided to elicit information about skill formation beliefs directly and use this information to estimate perceptions. They aimed to identify the causal links between inputs and outcomes seen in the original RCT.
The study’s methodology involved a beliefs elicitation survey instrument. Primary caregivers were asked questions about hypothetical children to understand their beliefs about child development and parental investment. Their responses were measured against observed markers of child development; language development was used as a proxy. The researchers constructed different scenarios by varying the level of investment and the initial conditions. They were then able to assess actual and perceived returns on parental investment by comparing real and perceived technologies of skill formation and their relation to parental investment choices. The measures made sure to include variables that were salient to the parents as well – in this case, language markers. The researchers also ensured that they were accounting for confounding variables, or factors that can influence child development outside of parental involvement. This was a challenging aspect of the study, as child development and parenting are complex processes.
The researchers found that the caregiver’s beliefs were in line with a model in which the technology of skill formation follows a Cobb-Douglas specification, meaning that the more inputs that parents employed (increased investment), the greater the returns in technology of skill formation and therefore child development. They noted that actual investment depended strongly on subjective perceived returns as well as resources. Their results show that parents thought that the productivity of investment was much higher for low initial conditions (disadvantaged children) than for higher initial conditions.
Participants asked a range of questions throughout the event. An attendee inquired about generational changes in parenting, and how social norms and expectations around parenting could influence perceptions of development. Attanasio agreed that this is true, and that parenting is influenced by community structure. This was one of the factors addressed in the study’s analysis. Another participant was curious about whether siblings impacted the study results and parent’s perceptions, to which Attanasio replied that the majority of the parents in the study only had one child, as the study population targeted parents with very young children, and furthermore siblings were controlled for. Finally, another question related to whether parents were told the gender of the hypothetical child about whom they were completing the survey. Attanasio stated that gender was not included in the intervention procedures, however this would be an interesting dimension to explore further in future projects.
The next steps for the researchers include analyzing a similar study performed in India, albeit with some important differences. Though the scenarios offered were consistent, participants were asked to share the probability the child will perform a given task by a certain age, instead of asking at what age the child will perform the task. Through this newer study, the authors have for the first time collected longitudinal data on beliefs to examine how parental beliefs change over time. There are many research avenues created by both the Colombia and India studies to learn more about parental behavior and beliefs, and how this impacts child development in meaningful ways.
*
Never miss an update: Subscribe to the Human Capital Initiative Newsletter.