Experimental Design Details
The task the workers are asked to do is a survey that gives a description of a hypothetical situation in which two children have agreed to gamble with their pocket money so that one will end up losing their pocket money to the other. The workers are then asked to decide whether they will allow the children to gamble or if they will prevent the gamble. The description of the hypothetical situation is divided into 4 different treatments yielding different descriptions of the nature of the gamble between the children.
(1) - The children have decided that they will bet a week's pocket money on a game of dice. Both roll a dice and whoever gets the highest roll will win the whole amount.
(2) - The same situation as in (1), except that the children have already rolled the dice and given the whole amount to the one with the highest dice roll.
(3) - The children have decided that they will play a game with a week's pocket money at stake, and whoever wins will get the whole amount.
(4) - The same situation as in (3), except that the game has already been played and the winner been given the whole amount.
Additionally, the hypothetical situations described to the spectators vary over 5 different age treatments, which vary the age of the children involved in the gamble. The children are varied between being 6, 9, 12, 15 or 18 years old.
They also answer to what extent they agree to different statements about paternalism, and also several background questions.
After having completed the survey the workers are told that they will be matched with another worker who has completed the same assignment. The workers are then asked whether they prefer a risky payment scheme where one of the workers will be randomly drawn to receive 5 USD, while the other will receive 1 USD, or a safe payment scheme where each worker will receive 2 USD for completing the assignment. They are told that the chosen payment scheme will be implemented with a certain probability.
The workers are matched in pairs of two where both preferred the same payment scheme, and 90% of the pairs are then matched with a spectator. The spectators are randomly assigned to six different treatments.
(1) Ex ante: The spectators are informed that both workers preferred the risky payment scheme, and they are then asked to decide whether the safe or the risky payment scheme should be implemented.
(2) Ex post - alternative A: The spectators are informed that both workers preferred the risky payment scheme, and they are then asked to decide whether the safe or the risky payment scheme should be implemented. If they choose to implement the risky payment scheme, worker A will be the lucky worker receiving the high payment.
(3) Ex post - alternative B: The spectators are informed that both workers preferred the risky payment scheme, and they are then asked to decide whether the safe or the risky payment scheme should be implemented. If they choose to implement the risky payment scheme, worker B will be the lucky worker receiving the high payment.
In addition, treatment (4), (5) and (6) are the same as (1), (2) and (3), except that the spectators are not informed about the preferred payment scheme of the workers.
Updated 31.10.17:
Three hypotheses for the comparison of the Norwegian and the American spectators:
- Fewer Americans violate Ex Ante Pareto.
- Fewer Americans choose Ex Post Egalitarian distribution.
- The share of egalitarian spectators violating Ex Ante Pareto is lower in the US than in Norway (a = share violating Ex Ante Pareto, b = share choosing egalitarian distribution Ex Post, the share we refer to is a/b).