Humans versus Chatbots: Scaling-up behavioral interventions to reduce teacher shortages

Empirical results in economics often stem from success in controlled experimental settings, but often fail when scaled up. This study presents a behavioral intervention and a scalable equivalent aimed at reducing teacher shortages by motivating high school students to pursue an education degree. The intervention was delivered through WhatsApp chats by trained human promoters (humans arm) and rule-based Chatbots programmed to closely replicate the humans program (bots arm). Results show that the humans arm successfully increased high-school students' demand for and enrollment in education majors, particularly among high-performing students. The bots arm showed positive but smaller and statistically insignificant effects. These findings indicate that a relatively low-cost intervention can effectively reduce teacher shortages, but scaling up such interventions may have limitations. Therefore, testing scalable solutions during the design stage of experiments is crucial.