Instead of trick or treat, how about treatment or control? We conducted two new studies on my porch this year for Halloween. Unfortunately, the mayor of New Haven recommended that people delay trick-or-treating post-Sandy even though the neighborhood was in good shape. This caused lots of confusion, and a turnout of half of the normal turnout of 600 or so kids. So sample size is down, standard errors up.
Alas, two nice results. Both written up in one-page one-graph papers.
First: replication is undersold in social sciences. Huge shame. So Experiment No. 1 simply redid an experiment from 2008, to see if we would get the same result. We found in 2008 that younger children, but not older children, had inelastic demand for Barack Obama, with respect to the quantity of candy. We replicated this exact result.
Second: In Experiment No. 2, we find that presenting a visual cue of Michelle Obama (compared to Ann Romney) led older children but not younger children to be more likely to choose fruit over candy.
These were conducted with the help of Yale’s Students for Proven Impact Club.