Skip to content. | Skip to navigation

Personal tools

Navigation

You are here: Home / Research / Selected Research / Using New Policy Parameter to Study Early Childhood Intervention for Low Birth-weight Infants

Using New Policy Parameter to Study Early Childhood Intervention for Low Birth-weight Infants

Erich Battistin examines a relatively understudied early-childhood intervention for low birth-weight infants using new policy parameter

Faculty Associate Erich Battistin along with other two co-authors just published a conference paper on IZA Institute of Labor Economics, where they examined a relatively understudied early-childhood intervention for low birth-weight infants particularly on the distribution outcomes using new policy parameter.

To learn more particularly about features of the effect distribution other than its average numbers, the study uses data from the Infant Health and Development Program (IHDP), a randomized clinical trial which provided comprehensive early intervention services for children in poor health (IHDP, 1990). Learning about features of the effect distribution other than its average is particularly relevant in this context, as poor health during childhood is an important mechanism for transmission of economic conditions.

The study further adopts a framework where potential outcomes with and without treatment are independent conditional on latent factors which affect location and scale of the outcome distribution in a flexible manner. Moreover, the study introduces a new policy parameter QCD, denoting quantiles of the conditional distribution of treatment effects given the latent factors.

Results indicate that:

1)exploratory factor analysis of all neonatal anthropometrics supports a single-factor model for latent neonatal health. Moreover, birth weight is the measurement of neonatal health with the lowest error variance (with a noise-to-signal ratio at about 5%).

2) a child’s health endowment at birth affects the location and scale of potential outcome distributions with and without treatment (a fact that a standard factor model would fail to detect).

3) within income-group variability in treatment effects is at least as important as between income-group differences in treatment effects; the larger average effects for the low-income group are driven by large returns for a minority of children in this group.

4) the income gradient in treatment effects is much stronger for close to normal birth-weight infants.

Reference: 

Battistin, Erich & Lamarche, Carlos & Rettore, Enrico, 2020. "Quantiles of the Gain Distribution of an Early Childhood Intervention," IZA Discussion Papers 13101, Institute of Labor Economics (IZA).


Navigation