Public good games and attitudes towards vaccination: group size and safe action

Respondents participate to 3 games: (G1) “risk elicitation game”, (G2) “binary threshold public good game” and (G3) “belief elicitation game about G2”.
The first two treatments consist in randomly grouping the respondents in groups of small size and of large size (5 people or 50 people). The third treatment consists in a modification of the binary threshold public good game in a way such that the “participate” action is always beneficial (instead of being “risky” if the participation threshold is not reached by the group) while keeping the group size as large as in the second treatment.

2021-10-25

2021-10-31

For Game 1, the outcome is the amount invested by the respondent. For Game 2, it is whether the respondent chooses to participate or not. For Game 3, it is the number of others in her group she believes decided to participate in Game 2.

Respondents are asked whether they are vaccinated or not against Covid19 and what were their motivation to do so. They are also asked what their expectations about the situation in Italy in the next future are.

Respondents participate to an online survey. They are first presented some questions about demographics and location. Then, they are asked to play 3 games. The treatment randomly divides respondents in 3 treatments (T1, T2, T3) and intervenes before Game 2, so Game 2 and Game 3 are different across treatments. However, Game 3 is equal for treatment T2 and T3.

Respondents complete a variety of pre-treatment covariates about their age and location.
Then, they are presented 3 games.
The first game is a “risk elicitation” game, where participants are asked to choose the amount of money they want to invest in a risky project (i.e., a project that succeeds 50% of the times).
The second game is a “binary threshold public good game” where the public good is enjoyed by all participants of the group if the participation threshold is reached (3 out of 5 or 25 out of 50, respectively in T1 and T2-T3). The game is of the type “volunteer’s dilemma” or “Kitty Genovese bystander effect”.
Respondents are randomly assigned to one of the following treatments:
1. treatment T1 “small and risky”: participants are told that they will be randomly assigned to a group of 5 people and participate to a public good game, where the public good is obtained if at least 3 out of 5 people decide to participate;
2. treatment T2 “large and risky”: participants are told that they will be randomly assigned to a group of 50 people and participate to a public good game, where the public good is obtained if at least 25 out of 50 people decide to participate;
3. treatment T3 “large and safe”: participants are told that they will be randomly assigned to a group of 50 people and participate to a public good game, where the public good is obtained if at least 3 out of 5 people decide to participate. The difference with the previous treatment is that the “participate” action is always beneficial for the respondent who chooses to participate, even if the threshold is not reached.
The third game consists in asking the participants to guess how many of the others in their group have participated.
Finally, we ask respondents whether they are vaccinated against Covid-19, their attitude towards vaccination and their expectations about how the situation will evolve in Italy.

Randomization is done in office by a computer.

Randomization is at the individual level. Specifically, each individual is randomly assigned to one of the 3 possible treatments (with even probabilities).

No

1,500 individuals representing the Italian population for demographic characteristics and geographical location.

1,500 individuals.

Individuals are randomly (and evenly) assigned to their treatment, so we expect around 500 individuals in each of the 3 treatments.