The Popeye Effect: Using Media to Improve Nutrition

How to cite

Sungu, Alp; Aouad, Ali; Kesebir, Selin; Ramdas, Kamalini; Swaminathan, Soumya. 2023. "The Popeye Effect: Using Media to Improve Nutrition." AEA RCT Registry. AEARCTR-0011156. https://doi.org/10.1257/rct.11156-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Alp Sungu* , Ali Aouad , Selin Kesebir , Kamalini Ramdas , Soumya Swaminathan

* Primary investigator

Registry ID
AEARCTR-0011156
DOI
https://doi.org/10.1257/rct.11156-1.0
First registered
Last updated
Status
In development
Trial dates
2023-04-05 – 2024-10-15
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/11156

Abstract

Finding cost-efficient ways to increase healthy eating in low-income communities is a pressing global challenge. Our study aims to reduce junk food consumption and increase protein-rich food consumption in a low-income settlement in Mumbai, India, by providing information messages. In a randomized field experiment, we send WhatsApp-based informational videos and observe the recipients' real shopping transactions by deploying scanner devices in local groceries. To examine what information works, we randomly manipulate the content of the video messages. The first arm receives informative videos emphasizing the health benefits of a nutritious diet, the second arm receives a similar video emphasizing the benefits in terms of economic success and performance, the third arm receives placebo (neutral) videos, and the fourth arm receives no messages.

Intervention

The intervention will be based on WhatsApp messages with a video intervention. Individuals will be randomly assigned to an information message or no message, control arm. The information message arm will be further divided into three arms. Our first round of messages will include a recording of an Indian public figure delivering the information via a short video. The first arm will receive information messages about the health benefits of a nutritious diet. The second arm will receive information messages about the success benefits of a nutritious diet. The notion of success is conveyed by emphasizing productivity gains and economic well-being dimensions. The third arm is a placebo control arm, where we will motivate individuals to buy the same set of products, but we omit additional information on health or success. Once a respondent is in a trial arm, she will stay in the same arm throughout the study. We will send multiple video messages to each respondent, where we vary the format (e.g., speaker, script, background) but maintain consistent emphasis (health, success, or placebo) for each arm.

Experimental design

In a field experiment, we will provide information to randomly selected low-income individuals via WhatsApp messages, focusing on the value of a healthy diet. We investigate the effect of such video messages on individuals' food habits. We first examine the effect of providing information. Next, we examine the differential effect of emphasizing different dimensions of information: health and success. Each respondent will receive a series of video messages.

Design details

Randomization method
The randomization will be established in an office computer before sending the messages.
Randomization unit
The randomization is at the individual level. One individual will receive multiple videos throughout the study, but will remain in the same arm throughout.
Clusters
We expect to observe 15,000 shoppers prior to us starting the experiment.
Observations
The design is not clustered. We will collect panel data that will cover the weeks before, during, and after the video messages.
Arms
Individuals in our experiment will be assigned to a video message arm with 60 percent probability. Remaining 40 percent will receive no video messages. Those who receive a video message will be assigned to a health, success or a (placebo) control message arm with equal probabilities. Therefore, each of these subgroups will constitute 20% of the total participant pool. We hereby, pre-specify the total number of participants and the duration of the experiment. Expected Participants: We aim to engage about 15,000 individuals in our study. Sixty percent of these will be randomly assigned to one of the three video message groups as outlined above. Based on our preliminary data, we estimate that we will be able to monitor approximately 35% of these participants during the three-week experimental period. Experimental Duration: The study will run for three weeks, during which time the video messages will be sent out. Following the last video message, we intend to collect data for at least an additional three weeks to assess the potentially persistent effects of the messages. If funding allows, we may extend the data collection period beyond this period to measure medium to long-term outcomes. Please note that our participant pool consists of adults who: i) have been observed at least once in our partner grocery stores, and ii) use WhatsApp. Our aim is to monitor individuals across a network of stores within the same community, which is why we have installed our devices in these locations. Unfortunately, due to high setup costs, we were unable to conduct a pilot study in a different community. Additionally, inviting individuals from our current network of stores to participate in a pilot study would directly reduce the sample size available for the main study.
Primary outcomes
Food purchases.
Secondary outcomes
Nutritional purchases.

Cited by

No citation data available for this registration yet.

Keywords