How Elastic are Preferences for Redistribution? Evidence from Randomized Survey Experiments
This paper analyzes the effects of information about inequality and taxes on preferences for redistribution using randomized online surveys on Amazon Mechanical Turk (mTurk). About 5,000 respondents were randomized into treatments providing interactive information on U.S. income inequality, the link between top income tax rates and economic growth, and the estate tax. We find that the informational treatment has very large effects on whether respondents view inequality as an important problem. By contrast, we find quantitatively small effects of the treatment on views about policy and redistribution: support for taxing the rich increases slightly, support for transfers to the poor does not, especially among those with lower incomes and education. An exception is the estate tax---we find that informing respondents that it affects only the very richest families has an extremely large positive effect on estate tax support, even increasing respondents' willingness to write to their U.S. senator about the issue. We also find that the treatment substantially decreases trust in government, potentially mitigating respondents' willingness to translate concerns about inequality into government action. Methodologically, we explore different strategies to lower attrition in online survey platforms and show our main results are robust across methods. A small follow-up survey one month later reveals that our results persist over time. Finally, we compare mTurk with other survey vendors and provide suggestions to future researchers considering this platform.
|Date of creation:||Apr 2013|
|Date of revision:|
|Contact details of provider:|| Postal: |
Phone: 44 - 20 - 7183 8801
Fax: 44 - 20 - 7183 8820
|Order Information:|| Email: |
When requesting a correction, please mention this item's handle: RePEc:cpr:ceprdp:9438. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: ()
If references are entirely missing, you can add them using this form.