Lately though Karlan’s work has turned to US politics. In a recent paper he wrote with a number of others, the trust of New Haven children was tested. The paper received a recent write up in Slate. The abstract describes the experiment below:
We conducted experiments during trick-or-treating at Halloween, four days prior to the 2008 presidential election. We decorated one side of a porch with Obama material and the other with McCain material. Some children are asked to choose a side to get an equal quantity of candy, whereas other children are offered more candy to go to the McCain side. At the candy table, each child chooses between a clear plastic bag and a brown paper bag, thus revealing their level of trust or comfort with ambiguity. We find that, in a predominantly liberal neighborhood, children choose the Obama table and continued to do so even upon the promise of more candy at the McCain table. We also find that Obama supporters, identified as those who choose the Obama table, are more likely than to take the brown bag of candy than the McCain supporters, identified as those who choose the McCain table.
These results mimic results from the General Social Survey in which supporters of Kerry over Bush in 2004 are more trusting.
I think the main lesson we can garner though is to not trust economists with candy.
h/t to Chris Blattman