“If I look at the mass, I will never act. If I look at the one, I will.” — Mother Teresa
A Brief Overview of the Original Experiment
The main research question of our selected paper is how deliberation on donations could impact people’s sympathy and generosity toward victims. One particular study was to investigate how the identifiable victim effect alone, and combined with high level statistics, could influence people’s donation decision. The origin hypothesis made in the paper Small et al was that showing statistical information in conjunction with an identifiable victim will reduce giving relative to just showing the identifiable victim.
In the original study, the research group recruited students at a university in Pennsylvania, and asked if they would complete a short survey in exchange for $5. A total of 159 individuals participated in the experiment.
The group first asked each participant to complete a survey about their use of various technological products and give each participant $5 in reward. Then each participant received a donation letter from an organization called Save the Children, and was asked if they would donate any of the $5 dollar they received to the organization after reading the letter. The participants were asked to indicate on paper the amount they chose to donate, and to include it with any money they donated in an envelope.
The key treatment in the original study is the donation letter each participant received. There are three conditions in the experiment:
Group 1 — Identifiable Victim: The charity request has a picture of a little girl and read a brief description about her.
Group 2 — Statistical Victim: The charity request described factual information taken from Save the Children website about the problems of starvation in Africa.
Group 3 — Identifiable Victim with statistical information: This letter added victim statistics to the charity request described in Group 1 for Identifiable Victim.
The independent variable is the charity letter presented to the participants. The dependent variable is the amount of money people donated after reading the charity letter. The key measurements taken in the original study is the mean donation to separate and joint presentation of victim types.
Replication Experiment Recruiting techniques
To replicate the experiment, we recruited 20 participants into our study. We posted an online recruiting message on Wechat Moment, a facebook-like social media feed, and received half of the total sign-ups. The other half joined the experiment through friends’ referral. Each participant was promised an unknown sum of Wechat digital currency (Wechat lucky money) after completion of the experiment. Participants entered the experiment by scanning one of the QR codes. The left QR code corresponded to the identifiable victim group. The right QR code corresponded to the identifiable victim & statistics group. Both groups were engaged and moderated through Wechat’s online messaging threads.
Experiment Procedure
20 participants were split into two groups. Half of the participants received a donation message with a description of an identifiable victim, and the other half received a donation message with a description of an identifiable victim and victim statistics. After the participants have finished reading the donation message, we asked all participants to fill out an online survey to indicate the amount of WeChat lucky money (in CNY) that they are willing to donate to help out this cause. We thanked the participants for their participation and distributed the promised WeChat red pocket at the end of the experiment.
We kept the same hypothesis from the original study in our experiment. The key treatment in our experiment is the donation message that the participants receive. We dropped the statistical victim group and conducted our experiment with 2 treatments on the donation message:
Group 1 — Identifiable Victim: The donation message contains a brief description of a girl named Kamali in Mali, West Africa. Given that we are using WeChat messaging threads to communicate the message, we choose not to include a picture of the identifiable victim and to keep the message concise with concrete details about the identifiable victim.
Group 2 — Identifiable Victim with statistical information: The donation message contains the same description of the identifiable victim Kamali in Group 1 with factual information about the problems of water contamination from CharityWater.org.
The independent variable in our experiment is the donation message each participant receives. The dependent variable is the amount of money people are willing to donate after reading the message. The key variable we are measuring in the experiment is the amount of donation in each of the control groups. More specifically, we are looking to compare descriptive statistics about the amount of donations generated in the two treatment groups.
Donation Message for Group 1 — Identifiable Victim:
Thanks everyone for your participation. Everyone will receive a red pocket at the end of the experiment.
The red pocket might be trivial to you, but it means a lot to a girl named Kamali who lives in Mali, West Africa. This money can provide her family with clean water resources that can last for 3 days. The community Kamali lives in doesn’t have access to clean water and the villagers are forced to use contaminated sources that they share with neighboring communities. Kamali’s brother died recently due to diarrheal disease.
Today, you have the chance to donate any amount of the red pocket (integer amounts are not required) you received to the villagers in Mali to help girls like Kamali.
Please indicate the amount of money you are willing to donate in the survey provided.
Donation Message for Group 2 — Identifiable Victim:
Thanks everyone for your participation. Everyone will receive a red pocket at the end of the experiment.
The red pocket might be trivial to you, but it means a lot to a girl named Kamali who lives in Mali, West Africa. This money can provide her family with clean water resources that can last for 3 days. The community Kamali lives in doesn’t have access to clean water and the villagers are forced to use contaminated sources that they share with neighboring communities. Kamali’s brother died recently due to diarrheal disease.
780 million people (Nearly 1 in 10 people worldwide) don’t have access to an improved water source.
In Africa alone, women spend 40 billion hours a year walking for water.
801,000 children younger than five years old die annually from diarrhea.
Today, you have the chance to donate any amount of the red pocket (integer amounts are not required) you received to the villagers in Mali to help girls like Kamali.
Please indicate the amount of money you are willing to donate in the survey provided.
Participants characteristics
Among 20 participants,9 are male and 11 female. Majority of the participants are university students and young professionals aged from 20 to 40. All participants are either pursuing or have received a formal college degree. They are digital natives and culturally aware individuals. Most of them are economically able to donate for philanthropic causes.
Summary Statistics (Unit: CNY)
Figures
Results
The data derived from our experiment have shown that the identifiable victim effect is more salient when presented without statistics. In this experiment, participants in the identifiable victim group were told a real victim story whereas participants in the identifiable & statistics group were told both a victim story and statistics. The data show that the former group generated more donations than the latter. When presented exclusively with a real victim, the donation messaging garnered greater aid. When a real story was complemented with statistics on a large, vaguely defined group, participants turned out to donate less. The data outcome confirmed our hypothesis that showing statistical information in conjunction with an identifiable victim will reduce giving relative to just showing the identifiable victim.
A real victim story adds both specificity and credibility to the human hardship, making people donate more. On the other hand, high-level statistics, when presented with the victim story, might undercut the specificity.
Another finding from our experiment was that the identifiable victim effect still holds in context of online fund-raising and digital currency. Our experiment was moderated entirely on WeChat. Donations were transacted digitally. Since some people perceive digital currency and physical currency differently, we were uncertain how the experiment setting could affect the experiment outcome. Our data suggested that even when the experiment environment was switched to an online, digital setting, the identifiable victim effect could still be salient. Alt our original hypothesis.
Trial and Errors: Takeaways
One key takeaway was that the framing of the donation messaging could have an interactive effect on the experimental outcome.
When we first designed our donation message, we indicated the amount of reward money that the participants would receive in our messaging. After we ran our procedure on Group 1 (Identifiable Victim), we discovered that the majority of participants ended up donating the entire amount (10 CNY). A potential explanation for such generosity was that the disclosed reward was too small. Relative to receiving this little amount of money, people would feel better off just donating it to a philanthropic cause. As a result, almost every participant donated 10 CNY, with just one exception. The uniform data derived from our first trial did not provide enough variance for us to analyze.
We, therefore, revised our original message and concealed the reward money. After this minor adjustment, we noticed more variance from our final data; participants donated various amounts in both groups.
In future research, we should identify other cognitive biases and effects (such as endowment effect, anchoring effect, and participation bias) and minimize their interactive effects on similar behavior-based experiments.