Table 4.

Testing for significant interactions between parental education and socioeconomic status (asset-based poverty)

Regression number	1	2
Dependent variable	HAZ	Stunting
Country-varying controls?	Yes	Yes
Community Fixed effects	No	Yes
Education rank included?	No	No
Mother's education (years)	0.019∗∗∗	−0.006∗∗∗
	(0.002)	(0.001)
Partner's education (years)	0.011∗∗∗	−0.003∗∗∗
	(0.002)	(0.001)
Asset poor = 1	−0.135∗∗∗	0.047∗∗∗
	(0.011)	(0.004)
Mother's education * Asset poor	−0.009∗∗∗	0.002∗∗∗
	(0.002)	(0.001)
Father's education * Asset poor	0.001	0.001
	(0.001)	(0.001)
Difference between genders significant at 10% level if asset poor?	No	No
N	376,992	376,992

Notes: Cluster-robust standard errors reported in parenthesis. Clustering is at the enumeration level ∗, ∗∗, and ∗∗∗ indicate significance at the 10%, 5% and 1% significance levels respectively. The “Asset poverty” indicator is a dichotomous variable equal to one if a household’s wealth score is below the 40th percentile. The wealth score is constructed from the first principal component of a common set of asset indicators: electricity access, radio ownership, TV ownership, improved floor material, ownership of a flush toilet and ownership of a basic toilet. See the text for a description of the models and other variables used in these regressions.