In-Group Competition for Incentives

How to cite

Olabisi, Michael; Maredia, Mywish; Liu, Jiawen; Ajibade, Toyin; Ajeigbe, Hakeem. 2022. "In-Group Competition for Incentives." AEA RCT Registry. AEARCTR-0009654. https://doi.org/10.1257/rct.9654-1.3

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

Michael Olabisi* , Mywish Maredia , Jiawen Liu , Toyin Ajibade , Hakeem Ajeigbe

* Primary investigator

Registry ID
AEARCTR-0009654
DOI
https://doi.org/10.1257/rct.9654-1.3
First registered
Last updated
Status
In development
Trial dates
2022-07-18 – 2022-08-22
Countries
Private
JEL codes
C93, J20, J33, O12, O14
Registry record
http://www.socialscienceregistry.org/trials/9654

Abstract

Governments in developing countries spend billions every year to promote technological adoption. A neglected but important question is how to effectively motivate the agents hired for such outreach projects, especially as digital extension services become more mainstream. We will conduct an experiment that puts hired agents for a mobile phone-based service into two treatment groups: agents working towards individual goals, and others competing for a share of the group-goal in their locality. The goals are based on the number of users that agents can register for KasuwaGo, an app to help farmers, sellers and resellers to trade with one another. We will compare performance means and dispersion for the two groups by measuring (a) the number of agents that achieve the goals. (b) the time taken to reach the goal, and (c) the number of registered users per agent.. The results could help to improve agricultural extension and other piecework-based programs in developing countries.

Experimental design

To study the goal structures that are more effective, we will conduct an experiment trial allots hired extension agents for a mobile-phone based service to two treatment groups: some agents are given individual goals, while others compete for a share of the goal for the group in their locality. The goals are based on the number of users that agents can register for KasuwaGo an app to help farmers, sellers and resellers to trade with one another. We will compare performance means and dispersion for the two groups.

Design details

Randomization method
Randomization done in office by a computer.
Randomization unit
The randomization and sampling strategy is based on practical considerations following the design of the already planned recruitment, training, and outreach of the KasuwaGo app. Given the locality-based recruitment of agents and outreach planned by the project, we plan to use localities as the units of randomization. Each locality from where the youth agents will be recruited from will be stratified into large, medium, and small market size. In each training session, 50% of localities represented will be randomly assigned to T1 and 50% to T2, each stratified by the market size.
Clusters
70 localities/towns
Observations
220 hired registration agents
Arms
35 localities treatment, 35 localities control
Primary outcomes
(a) the number of agents that achieve their set goal. (b) the time taken to reach the goal, and (c) the number of registered users per agent.

Cited by

No citation data available for this registration yet.

Keywords