Combatting COVID-19: Measuring and Changing Beliefs, Knowledge and Behaviors

We implement over-the-phone interventions to 1) encourage social distancing by accelerating changes in community norms, and 2) improve knowledge about COVID-19

Social Distancing Treatments:
-SD1: Community Support for Social Distancing. We will ask individuals whether they themselves support social distancing, and use this information to calculate the fraction of households in the community who support social distancing. Then, in a later phone call, we will ask individuals to guess the share of households in the community who support social distancing. Individuals who underestimate the true share of households in the community that support social distancing will be given information on the true (higher) share of support for social distancing, and individuals correctly guessing the true share will be told that their guess is correct.
-SD2: Community leader support for social distancing. We will survey community leaders and ask them to endorse social distancing in their communities. In this treatment, we will inform households by phone call that their leaders support social distancing in their communities.

Knowledge Treatments:
-K1: Knowledge Incentives. We will randomly offer a subset of respondents 5 Mozambican meticais (MT) for every correct knowledge response on a subsequent phone survey. We will examine the effect of the treatment on future knowledge and behavior. If they answer all 40 questions correctly, respondents can earn a maximum of 200MT (approx. US$2.86).
-K2: Tailored Feedback. We will randomly give tailored feedback to a subset of respondents based on their response to COVID-19 knowledge questions, by informing them of a subset of their correct responses and correcting a subset of their incorrect responses. We will examine if tailored feedback improves relevant knowledge and behavior in a subsequent telephone survey.

2020-08-26

2020-10-04

For Social Distancing interventions:
-An indicator for the respondent practicing social distancing.

For Knowledge interventions:
-Index of COVID-19 overall knowledge

For Social Distancing interventions:
The primary outcome will be an indicator for the respondent practicing social distancing. It will be constructed from two component indicators: the own report of practicing social distancing, and others’ report of the respondent’s practicing social distancing. The primary outcome will be equal to one if both the own report and others’ report of practicing social distancing is equal to one, and zero otherwise. Some detail below and see Analysis Plan for further details.
-Others’ and self-report of social interactions: Others’ reports of social interactions with respondent. Taken from the question ask for up to 10 other study participants in one’s social network and neighbors within 200m: Did you talk to / stand within 1.5 meters / shake hands or otherwise touch (insert name) in last 14 days?
-Index of household social distancing behaviors: Ask about whether or not the household attends social gatherings, leaves the household area, avoids crowded areas, keeps a distance of 1.5 meters from others, frequently wash hands, informs others if sick, and more…
◦ Is this something you think people should be doing?
◦ Is this something your household has been doing for the last 7 days?

For Knowledge interventions:
-Index of COVID-19 overall knowledge - number of correct responses to knowledge questions about coronavirus symptoms, prevention, how it spreads, and who is most at risk; household social distancing and self-prevention behaviors. Confidence with which responses are given.

We randomize our over-the-phone interventions to test if we can 1) encourage social distancing by accelerating changes in community norms, and 2) improve knowledge about COVID-19 via incentives and tailored feedback. We consider the intent-to-treat (ITT) effect of the randomized interventions on a standardized version of our outcomes: indices of perceived social distancing norms and household social distancing behavior.

All interventions will be implemented in the Round 2 Survey; this is because some treatments require input from the Round 1 Survey and will allow for comparison across treatments. Based on our power analysis, we limit to four treatment arms to detect effects of reasonable size. Thus, we will cross-randomize Social Distancing Treatments and Knowledge Treatments. Within each treatment family, the control group will be 40% of the sample and each of the three treatment arms will be 20% of the sample. Therefore, 16% of the sample will be a strict control group, neither receiving a Social Distancing treatment nor a Knowledge Treatment.

Randomization done in office by a computer

Household

No

2,000 Households

2,000 Households

2,000 Households

The study has sufficient power to detect effects of reasonable size in the analyses for our primary hypotheses. For these power calculations, we consider the intent-to-treat (ITT) effect of the randomized interventions on a standardized version of our outcomes: indices of perceived social distancing norms and household social distancing behavior. The study sample will be drawn from the 3,135 individuals for whom we have phone numbers. We assume a conservative response rate of 54%—drawn from a follow-up telephone survey conducted in 2019—which we aim to improve through repeated calls at different times of the day. We calculate the minimum detectable effect (MDE) of comparing the control group to one of the three treatment groups (i.e., each group is one fourth of the responding sample). The MDE is 0.1929, meaning that our study is sufficiently powered to detect a difference of 0.2 standard deviations in our standardized outcome measures between the control group and each treatment group.

Accelerating Changes in Norms about Social Distancing to Combat COVID‐19

Learning about COVID-19: Improving Knowledge via Incentives and Feedback

Can informing people of high community support for social distancing encourage them to do more of it? We randomly assigned a treatment correcting individuals' underestimates of community support for social distancing. In theory, informing people that more neighbors support social distancing than expected encourages free-riding and lowers the perceived benefits from social distancing. At the same time, the treatment induces people to revise their beliefs about the infectiousness of COVID-19 upwards; this perceived infectiousness effect as well as the norm adherence effect increase the perceived benefits from social distancing. We estimate impacts on social distancing, measured using a combination of self-reports and reports of others. While experts surveyed in advance expected the treatment to increase social distancing, we find that its average effect is close to zero and significantly lower than expert predictions. However, the treatment's effect is heterogeneous, as predicted by theory: it decreases social distancing where current COVID-19 cases are low (where free-riding dominates), but increases it where cases are high (where the perceived-infectiousness effect dominates). These findings highlight that correcting misperceptions may have heterogeneous effects depending on disease prevalence.

Replication package: https://doi.org/10.7910/DVN/TMARZT

Yang, Dean, James Allen IV, Tanya Rosenblat, James Riddell IV, Hang Yu, and Arlete Mahumane. 2024. "Correcting Misperceptions about Support for Social Distancing to Combat COVID-19". Economic Development and Cultural Change. https://doi.org/10.1086/727192

https://doi.org/10.1086/727192

Interventions to promote learning are often categorized into supply- and demand-side approaches. In a randomized experiment to promote learning about COVID-19 among Mozambican adults, we study the interaction between a supply and a demand intervention, respectively: teaching via targeted feedback, and providing financial incentives to learners. In theory, teaching and learner-incentives may be substitutes (crowding out one another) or complements (enhancing one another). Experts surveyed in advance predicted a high degree of substitutability between the two treatments. In contrast, we find substantially more complementarity than experts predicted. Combining teaching and incentive treatments raises COVID-19 knowledge test scores by 0.5 standard deviations, though the standalone teaching treatment is the most cost-effective. The complementarity between teaching and incentives persists in the longer run, over nine months post-treatment.

Replication Package: https://doi.org/10.7910/DVN/BCMVJT

Allen IV, James, Arlete Mahumane, James Riddell IV, Tanya Rosenblat, Dean Yang, and Hang Yu. 2023. "Teaching and incentives: Substitutes or complements?". Economics of Education Review 91: 102317. https://doi.org/10.1016/j.econedurev.2022.102317

https://doi.org/10.1016/j.econedurev.2022.102317