Experimental Design
The participants are first asked how much each member of their household earned in 2014. Based on this information, we calculate their household income for 2014, and show participants how much it adds up to. We then ask them where they think their household stood in the income distribution. The exact question they answered was: "In your opinion, what percentage of US households earned less than your household in 2014?" If participants correctly guess the percentage (within three percentage points), they receive a bonus payment of 10 cents.
Using Census data on the US household income distribution, we can determine whether participants accurately evaluated their position in the income distribution. This allows us to give participants in the treatment group some information on their actual standing in the income distribution. Subjects in the control group, on the other hand, do not receive any information.\medskip
We divide participants into two groups: those who over-estimate their position in the income distribution, and those who under-estimate it. We expect that very few people will guess their exact position in the income distribution, and we therefore discard those observations completely. For clarity's sake, we call participants who over-estimate their position in the income distribution over-estimators, and participants who under-estimate their position under-estimators. This terminology will be used in the rest of the paper.
Half of the over-estimators do not receive any information about the accuracy of their estimate, while the other half receive the following message:
Actually, you overestimated your relative position in the income distribution. In reality, you are relatively poorer than you thought. In other words, you are closer to the bottom of the income distribution than you thought. You currently earn significantly less than what you would need to be at the position you thought you occupied.
Similarly, half of the under-estimators do not receive any information about the accuracy of their estimate, while the other half receive the following message:
Actually, you underestimated your relative position in the income distribution. In reality, you are relatively richer than you thought. In other words, you are closer to the top of the income distribution than you thought. You currently earn significantly more than what you would need to be at the position you thought you occupied.
All participants are then asked how satisfied they are with their position in the income distribution. This question allow us to check, separately for over-estimators and under-estimators, whether the treatment has any effect on the participants' economic satisfaction. We also ask participants in the treatment group to give us a new estimate of their position in the income distribution, now that they have received some information about the accuracy of their estimate. This question enables us to see to what extent participants updated their beliefs regarding their position in the income distribution, after receiving the treatment. All these questions are used as manipulation checks in the analysis, to make sure that our information treatment changed people's perception of their relative income.