The Malleability of Social Preferences: Causal Evidence for Disadvantaged Children

Motivated by the growing evidence on the importance of early investments on human capital development, home visiting programs aimed at improving parenting skills and home environment are becoming increasingly popular worldwide. Yet, the link from improved parenting skills and home environments to children’s social preferences is rarely being investigated.
This paper investigates the malleability of primary school children’s social preferences in highly disadvantaged families. Our research uses a 5-year follow-up of a randomized controlled trial (RCT, n=755) to evaluate a home visiting program in Germany (Pro Kind). Born as adaptation of the well-known Nurse Family Partnership (NFP), Pro Kind was specifically designed to improve maternal and child health, as well as parental skills of first-time disadvantaged mothers. Treated women received assistance by family midwives and social workers from pregnancy until the child’s second birthday.
In the 5-year follow-up, experiments were conducted to elicit the children’s social preferences. Using four different binary dictator games, we can directly observe children’s social preferences. Our results reveal that the home visits impacted prosocial behavior of treated mothers’ children five years after the intervention. In two dictator games the treatment significantly increased children’s prosocial behavior. The children’s willingness to share increases by 10pp while costly envious behavior decreases by 9pp. The observed treatment effects are of large magnitude, as the overall willingness to share is 29% and the overall proportion of children who behave costly envious is 39%.
Treatment effects are stronger for mothers with a very low socioeconomic status. As children whose mothers have a low socioeconomic status also tend to share less, the home visiting helped decreasing this socioeconomic sharing-gap.
Since the intervention was not targeted at the children themselves, we assume specific measures of parenting and home environment as potential mediators for the effects on children’s social preferences. In particular, we find a decrease in abusive parenting indicators and an increase in maternal prosociality indicators for treated mothers.
To further understand the findings on children’s social preferences, and the link between these preferences and the home visit intervention, we combine the outcomes from the dictator games and survey questions with administrative social insurance data of the families. We can observe mothers’ labor market participation, welfare receiving and the household size. In the next step of our analysis, we plan to use this administrative data to investigate further potential mechanisms that led to the observed changed social preferences of the children.