Last Halloween, Yale PhD student Santosh Anagol conducted a clever social experiment on the 200+ trick-or-treaters who visited his New Haven home. Anagol presented each child with two bowls, one covered and one uncovered, containing blue and red markers.

Each child first chose a bowl, and then guessed what color marker Anagol would pull (blindly) out of that bowl. If they were right, they’d get a big candy bar instead of a small one.

In the uncovered bowl, the child could estimate, visually, the proportion of red and blue markers and make an educated guess. But with the covered bowl, of course, the odds were hidden.

So which kids would go for covered bowl, and thus risk their chance of winning big candy bar? Those with the best costumes, apparently.

After analyzing the data, Anagol and his economist colleagues found that the kids with the most common costumes—princess, witch, and Spiderman—were 30 percent less likely to choose the risky bet, the covered bowl. Moreover, they found that kids in “fantasy” costumes, such as dragons or goblins, were 7.9 percent more likely to choose the covered bowl, though that result didn’t reach statistical significance.

It’s a cute study, but what does it mean?? Does it mean that there’s a link between creativity and risk behavior? That you should scrutinize your kid’s choice in costume? That Harry Potter leads to gambling? That Spiderman is a whus?

Unfortunately, the researchers didn’t make any grand conclusions from the data, stating only that “the concept of ambiguity, in both the theoretical and empirical literature, remains ambiguous.”

Happy Halloween!

(Hat tip: Boston Globe)