Intervention (Hidden)
To test whether poor parents have a lower propensity to invest in their children due to cognitive load, we resort to survey experiments, priming some parents – but not others – about money, and then presenting them with the opportunity to undertake a costly educational investment in their child. The hypothesis of interest is whether those who are primed undertake such investment to a lesser extent.
Such empirical strategy raises three main concerns. First, if there is an effect, how do we know that it is driven by cognitive load, rather than just by a negative mental state following the arrival of bad news (Lerner et al., 2014)? Second, the fact that the investment is costly may confound the effect of cognitive load with that of credit constraints. And third, if the decision is hypothetical, then it is hard to get a sense of whether it matters quantitatively.
To deal with the first concern, we randomly assign subsamples of the subject pool to other sources of negative news: either a monetary shock of much lower magnitude, or a non-monetary shock. Moreover, we also try to delve more deeply into the mechanism, by varying the source of the large monetary shocks – either related or unrelated to educational expenses –, to address the possibility of mental accounting (Thaler, 1999).
To deal with the second concern, we create a decision that involves opportunity costs, but which does not require cash on hand. Specifically, we offer parents the choice between a free educational product for a longer period, or the same product for a shorter period coupled with an immediate top up in airtime credit. The idea is that if those who are primed about money choose the short-length plan, it cannot be because they were credit constrained to choose the longer-length one.
To deal with the third concern, we present parents with a decision about a real product. We offer enrollment in an SMS campaign that delivers content to support parenting, an increasingly popular intervention to foster parental engagement (York and Loeb, 2015), for either 2 or 4 months of the school year. Beyond having real consequences, we can actually measure the impact of parents’ decisions on their children’s education outcomes – attendance, grades and drop-out rates –. The psychological impact of poverty on those outcomes is then estimated by using priming status as an instrumental variable for parents’ decisions between shorter- or longer-length plans.
Last, we also offer parents a savings technology, presenting all subjects with the trade-off between receiving an immediate 5-dollar airtime credit top-up for answering the survey, or a 10-dollar top-up one month later instead. Estimating the effects of the priming on subjects’ discount rate provides us yet another way to benchmark priming’s effect size on that educational decision.
While the cognitive load/mental bandwidth theory predicts that poverty might impose a psychological tax on poor parents’ cognition, it also predicts that the poor should perform relatively better in tasks framed in monetary terms (Shah, Shafir and Mullainathan, 2015; Lichand and Mani, 2015). By precisely exploiting this mechanism, interventions to foster parental engagement among the poor might be improved by framing communication in monetary terms.
This paper also tests this hypothesis, by cross-randomizing how the educational investment is presented for parents (framed as either “good for your child” or “with potential to increase future wages”). We test whether the monetary framing increases uptake of the longer-length plan, particularly for the subjects primed about money.