Does the framing of information on inequality in absolute or relative terms affect altruism?

2022-11-14

2023-03-31

Measures of altruism, which are the amounts indicated for donation in the dictator games with low and high stakes, respectively.
We use this measure to address our main question ‘Does the framing of information on inequality in absolute or relative terms affect altruism?’ by testing whether there are significant differences between the groups who heard the treatment in absolute or in relative terms.

The game is a simple (non-incentivized) dictator game with low and high stakes that allows us to observe the differences in the amount participants would be willing to donate depending on the overall value that is at stake. We use it to derive a measure of altruism. An example of the text used in the low stakes questions is as follows, illustrating with the Mozambique case:
“Imagine that you were given 200 MZN that you could keep or donate (a share or the entire amount) to build a well-functioning hospital or health centre in the village I described. You see on your answer sheet that there are ten circles with a 20. They represent ten notes of 20, for a total of 200 MZN. In this case, how much would you choose to keep and how much would you choose to donate?”
Our experiment includes an information treatment, read to participants before they play the dictator games. All participants receive information about the difference in wealth of the poorest person and the richest person in one village. For example, in the case of Mozambique enumerators read the following: “I will now give you some information about a different village in Mozambique, for example, Alipe. This information is about a village in Mozambique. Many people live there.” To half of the participants in the session we frame this information in absolute terms (our absolute treatment) and show a diagram to highlight the absolute differences: “As you can see in this drawing, the richest person has 6 goats more than the poorest person.” To the other half of the participants we frame the information in relative terms (our relative treatment), again showing a diagram to highlight the relative differences: “As you can see in this drawing, the richest person has 4 times more goats than the poorest person.” This allows us to assess whether providing information about inequality in relative or absolute terms influences the participants’ answers in the dictator games.
Additionally, before receiving the information treatment and playing the dictator games, we ask participants a simple question to infer whether they think about inequality in absolute or relative terms. We then examine whether this indicator is correlated with the amounts chosen for donation in the dictator games.
The design of the study is based on between-subject comparison. In each country, we initially run a total of 40 sessions, 20 in rural settings and 20 in urban settings (see more details on the sampling in section 3) and we will add an additional round of 40 sessions, 20 in rural settings and 20 in urban settings, in the South of Vietnam. Each of the treatments is randomly allocated to half of the participants in each session.

The randomization is done within the experimental site. Half of the participants are randomly selected to go to a different room (based on a random count).

The treatment is randomized at the individual level. Each of the treatments (absolute and relative) is randomly allocated to half of the participants in each session.

No

1,040 individuals in Mozambique and 2,080 individuals in Vietnam (total of 3,120 individuals).

1,040 individuals in Mozambique and 2,080 individuals in Vietnam (total of 3,120 individuals).

In each data collection effort (Mozambique, North of Vietnam and South of Vietnam), 520 individuals receive the absolute treatment and 520 individuals receive the relative treatment.