The limits of nudging: Results of a randomized trial of text messages to promote home-based caregiving and reduce perceptions of stigma for COVID-19 patients in Kenyan informal settlements.

During the early stages of the global COVID-19 pandemic, governments searched for effective means to rapidly disseminate information about how to prevent the disease and care for sick household members. In June 2020, the government of Kenya considered sending text messages, a behavioral nudging approach, to inform and persuade the public to practice home-based care for those who were infected. We conducted a randomized evaluation of simple informational messages compared to messages targeting personal and social benefits for those receiving the messages. We hypothesized that those that received messages tailored around social or personal benefit would be more likely to undertake the promoted behaviors of isolating if infected with COVID-19 and intending to care for an infected family member. While fear and perceptions of stigma were widespread, more than two-thirds of respondents in the control condition expressed an intention to care for an infected family member at home. Despite greater recall of the personal benefit message, which used reciprocity as its key behavioral lever, intentions to provide care at home and perceptions of stigma did not differ across study groups. Rather, capabilities such as wealth and having sufficient room at home were the key determinants. While text messages as behavioral nudges may be useful for some behaviors, policymakers should consider a broader range of tools for behaviors that are influenced by people's capabilities, since even low-cost interventions may crowd out the time and energy needed for other responses during an emergency.