Leah R. Rosenzweig

“Strategies to increase covid-19 vaccination in kenya: does social pressure play a role?” WITH K. CARNEY, M. KREMER, E. Maffioli & W. Wong. in progress

Intensive public health policies can sometimes be regarded as intrusions on personal freedoms. In public health emergencies governments often send healthcare providers to local communities to offer vaccination, but this carries the risk of public backlash. This study examines the impact of such a policy on immediate and medium-term vaccine uptake during the COVID-19 pandemic in Kenya. To assess whether vaccination was influenced by social pressure from healthcare providers, we adapted a technique from studies on charitable giving and voting and randomized whether the healthcare provider visits were announced in advance, thereby allowing vaccine avoidance without the social costs of saying no. We find that sending healthcare providers to communities and visiting households to invite individuals to get vaccinated increased vaccination by 7.4 cumulative doses per 100 people on the day of the intervention (SE = 0.011, p<0.001), representing an 8% increase over the baseline number of cumulative doses in the control group (89.6 per 100 people). The effect persisted in the three months following the intervention, indicating that the campaign vaccinated individuals who otherwise would not have been vaccinated, rather than merely accelerating the timing of vaccinations. Announcing the visit increased the probability of receiving an additional dose by 3.8 percentage points, equivalent to a 16% increase over the baseline rate in the unannounced group (23.1%). This suggests that the increase in vaccination rates was not driven by social pressure. We estimate the cost per marginal dose to be $42, and propose strategies for improving cost-effectiveness of future interventions.

“Conversations with a concern-addressing chatbot increase covid-19 vaccination intentions among social media users in kenya and nigeria” With M. Offer-Westort. [policy brief, virtual panel] Forthcoming at Journal of Politics.

This study leverages a two-stage response-adaptive experimental design to explore how tailored digital messaging can influence public attitudes toward health interventions, with broader implications for political communication. In public health settings, quickly scaling effective communication is critical; adaptive designs enable efficient learning of relevant treatments. For this study, we recruited 22,052 Facebook users in Kenya and Nigeria to engage in conversations on Messenger about their concerns regarding COVID-19 vaccines. We optimized messaging using an adaptive algorithm, then experimentally evaluated the optimized concern-addressing messaging. We find that the optimized concern-addressing messaging increases COVID-19 vaccine intentions and willingness by 4-5% compared to a control condition, and by 3-4% compared to a public service announcement. We observe the largest treatment effects among those most vaccine hesitant at baseline. Personalized digital messaging interventions offer a scalable communication tool to encourage compliance with public health programs.

“Can SMS Interventions Increase Vaccination? Evidence from the COVID-19 Pandemic in Kenya” With K. Carney, W. Wong, F. Akech, J. Otieno, E. Maffioli. Journal of Development Economics (2025).

Researchers, practitioners, and policymakers often view Short Messaging Service (SMS) interventions as quick and inexpensive tools to influence individual decisions. SMS text messaging is particularly attractive in countries where much of the population lives in remote areas but mobile phone technology is becoming increasingly widespread. In partnership with the Ministry of Health in Kenya, we sent a series of SMS messages—incorporating behavioral nudges and financial incentives—to encourage individuals overdue for their second COVID-19 vaccine dose to complete their two-dose primary series. Applying principles from behavioral economics, we conducted four experiments between July 2022 and January 2023, sending randomized SMS messages to 51,903 individuals encouraging them to complete their vaccination. We found no evidence that these messages, designed to nudge behavior or offer financial incentives, significantly increased vaccination completion rates. A follow-up phone survey with a subsample of 105 respondents revealed that, while 91% of people remembered receiving the message, 84% of the sample reported being either not worried or only a little worried about the virus. Our data suggest that, despite recipients recalling the specific SMS, a single message with less salient content may not be effective in improving health outcomes.

“If we move, it moves with us:” Physical distancing in africa during covid-19 With C. Bicalho & M. Platas. World Development (2021). [Appendix, Data, Tumbl] WAshington POst

Until a vaccine is widely available, physical distancing is central to curbing the spread of the novel coronavirus. In this paper, we use an online survey of respondents in Kenya, Nigeria, and Uganda conducted in April 2020 to understand knowledge, beliefs, and behavior related to physical distancing. We find that, while there is widespread knowledge that physical distancing reduces the spread of the virus, respondents underestimate their peers’ support for policies designed to enforce physical distancing, expect others not to practice physical distancing, and do not maintain physical distance themselves. However, more than half of respondents wrote a message to encourage others to practice physical distancing. Findings from survey experiments suggest that making salient the social and material costs for not keeping physical distance are insufficient to encourage compliance. Given the gap between individual’s own attitudes and expectations of others’ attitudes toward lockdown policies, we suggest that providing information on the extent of public support for physical distancing, in citizens’ own words, could be useful in encouraging compliance.

Sierra Leone With L. Tsai

With the Institute for Governance Reform (IGR) and Sierra Leone’s Directorate of Science, Technology and Innovation (DSTI) we created a nationally representative panel of citizens who were first surveyed April 11-18 [Wave 1 report] to help inform the government’s lockdown and aid distribution decisions.

Here is a Science Magazine article about our work.

Uganda with C. Bosancianu, A Garcia-Hernandez, M. Humphreys, P. Kiwanuka-Mukiibi, M. Platas, & L. Tsai

The Covid-19 pandemic has already disrupted community life and will likely have lasting effects on social and political dynamics. This project aims to identify and track over time citizens’ information about the virus, their compliance with Covid-19 mitigation policies, and local community norms of behavior using repeated phone surveys of a representative sample in Kampala, Uganda.

Data dashboard is available here!

Nigeria with J. Phillips, M. Platas, L. Tsai

With the Nigeria Center for Disease Control and Innovations Poverty Action (IPA) Nigeria we conducted a phone survey of citizens in Lagos State to understand their information about Coronavirus, attitudes toward the government’s response and willingness to comply with government policies.