Kentucky SFEM MLEs.

Both behavioral and experimental economics have a lengthy research history concerning the presence of and ability to support cooperation in infinitely repeated games. A workhorse model which has emerged for inferring players’ strategies from their actions in finitely repeated games, is applied to data from randomized controlled trials conducted in Jordan and Kentucky; an attractive economic application portending substantive public health and other policy consequences. We find that treatment significantly increases unconditional cooperation (from 27% of controls to 45% of treated subjects in Kentucky, and 38% to 85% in Jordan, with an intermediate level of 76% associated with partial treatment in the latter), and despite all attending cultural and other differences between these two geographies, treatment in each yielded significant and unimpeachable improvements in outcomes; improved program attendance and weight loss. Such harmonious results across vastly disparate geographies is highly unusual and correspondingly impressive. For the first time, we are showing increased cooperative behaviors causally associated with these trials.