NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Charitable giving by the poor. A field experiment
Initial registration date
March 03, 2018
March 05, 2018 10:33 AM EST
This section is unavailable to the public. Use the button below
to request access to this information.
Other Primary Investigator(s)
WZB Berlin & UCL
Additional Trial Information
The studies of altruism so far mostly concentrated on middle or higher middle incomes, neglecting the poor, although the lowest income groups are often shown to contribute large shares of their income to charitable causes. In a large-scale natural field experiment with clients of a micro-lending company in Kyrgyzstan, we study charitable giving by the poor. Our subjects own, on average, a debt equal to a monthly salary. In a 2x2 design, we set to understand the following: How do they respond to a fundraising campaign that collects money for local projects? How do they react to price incentives, that is matching of donations?
The population under study are clients of a micro lending organization who, on average, own a debt equal to a monthly salary. In addition to their main business, the micro lending organization supports a local charitable fund. The fund realizes charitable projects that aim at helping the people in the country. Specifically, now, they plan to implement nine local projects related to water supply, local infrastructure and school reconstructions. In order to implement the projects they run a fundraising campaign, among others, asking the customers of the micro lending organization for support.
Intervention Start Date
Intervention End Date
Primary Outcomes (end points)
(1) Response rate - donation dummy (yes/no) per customer or number of donations/number of customers per office
(2) Average positive donation
(3) Return from campaign - average donation (including zeros)
(4) Share of customers informed about the campaign
Primary Outcomes (explanation)
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
2x2 design with following treatments:
1) The first branch is about matching donations: One half of the clients will be informed that a lead donor has already contributed a large amount of money, the other half that a large donor will match their donations one by one. Both informations are true, as large amounts of money have already been contributed by large donors, and we will match the donations in the treatment condition 1:1. (Since we expect the individual donations to be in a range of cents, we assume the total donation in the range of at most €4000-6000. The clients will be informed about the maximum cap at half a million of local currency= around €7,000.)
2) The second branch is about local benefits: One half of the clients will not get any additional information, another half will be informed that "If clients of [name of the company] from your region will donate the highest amount per client, the next project that will be funded from the fund will aim at helping your region! Thus by donating now you also help projects from your region to be funded." This is also going to be implemented ex post. There will be no specific priority implemented for the group that is not informed about priority. Implementation: the clients will be informed about the fundraising campaign by their loan officer or while visiting the office. A telephone survey will be conducted with a subsample of the clients in order to control the share of informed clients. Loan officers will receive incentives to reach a high rate of informed clients (punishment for low rate).
Experimental Design Details
randomization done in office by a computer: highest possible balance on observable characteristics at the level of office was be attempted. Observable characteristics at the office level include: region, number of clerks, average age, experience and gender of clerks, number of clients, ethnic shares among the clients, average portfolio etc.
Was the treatment clustered?
Sample size: planned number of clusters
Sample size: planned number of observations
over 160,000 individuals
Sample size (or number of clusters) by treatment arms
26 offices in each of the 2x2 treatments
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Minimum Detectable Effect Size for Main Outcomes 10% (details outlined in the additional document).
INSTITUTIONAL REVIEW BOARDS (IRBs)
WZB Research Ethics Committee
IRB Approval Date