Infant Simulators and Behavioral Economics
A recent study published in the Lancet has received a lot of media attention over the last two weeks. The study represents the first randomized control trial (RCT) assessing an infant simulator program (life-like robotic baby) on longer term teenage pregnancy outcomes. The program evaluated was the Virtual Infant Parent (VIP) Programme in Western Australia. The results of the paper show that those high school girls who spent time caring for the simulator (treatment group), actual became pregnant at a higher rate than those in the control group. Plus, after controlling for differences in socio-economic status (treatment group participants were on average poorer that control group), the authors assert that the girls in the treatment group were at a greater risk of an unplanned pregnancy than those in the control group.
Lets break down this result through the lens of behavioral economics. Through this lens, we can see that the goal of the infant simulator is to correct for the Illusion vs Reality or Today vs Tomorrow bias (broken down in the BERI White Paper), by giving these adolescent girls a taste of what motherhood may be like. With a hope that if these biases are corrected, then the pregnancy outcomes will align with stated intentions. The Lancet study cites a synthesis report showing previous studies found mixed results regarding the impact of infant simulator programs on self-reported beliefs about the demands of parenthood. While the Lancet study does not measure stated intentions directly, the evidence from the cited synthesis report combined with the reported pregnancy outcomes, suggest that infant-simulators are ineffective at addressing these behavioral biases of adolescent girls.
Another potential behavioral reason this infant-simulator program was ineffective is that it does not address the Self vs Other bias. Succinctly summarized by Julie A Quinlivan, “it takes two to tango. The intervention was directed at teenage girls and neglects [potential] fathers.” Therefore, even if the intervention was successful in addressing the previously mentioned biases of adolescent girls, their pregnancy outcomes may not be any different. This is because the intervention did not address biases in the potential fathers, and the impact these biases have on the actions of the adolescent girls in the program.
It is estimated that 21 million adolescent women in developing countries will become pregnant this year, and almost half of those pregnancies will be unintended. Thus, there is a great need to find cost effective / scalable ways to help prevent these unintended pregnancies.