Research & Projects
Research & Projects
Joint work with Robin Musolff
Differences in factual beliefs between members of opposing political parties are sometimes attributed to politically motivated reasoning - an asymmetric reaction to information whereby friendly political messages are weighed more heavily. Using a series of survey experiments, we find that these underlying differences in factual beliefs are exaggerated, and that despite the expectations of both the general population and academic experts, politically motivated reasoning is not widespread among US partisans. Instead, we find it is limited to highly polarized Democrats and unlikely to explain differences in beliefs. Our experimental design stands out in that it includes easy to understand informative signals, and a direct measure of belief movement in a non-binary environment. Our results suggest that whatever differences in beliefs exist among US partisans originate elsewhere.
Joint work with Ro'ee Levy and Moses Shayo
Coverage of the ongoing Israel-Gaza war varies enormously across the two sides of the conflict. The two sides also seem to hold sharply different beliefs about the facts of the conflict, with each doubting the other’s claims regarding civilian casualties. Are these purely a product of ignorance and biased local news outlets, or is there also a supply side element to the phenomena? How would their perceptions and behaviors change if they were exposed to sort of news they usually do not see?
We conduct a large scale experiment among Israeli Jews and Jordanian Arabs to ask three questions relating to the ongoing Israel-Hamas war. First, do individuals try to avoid information about victims from the other side? If so, why? Second, if individuals were exposed to such information, would this affect their empathy towards the other side, and their views on the conduct of the war? Third, is the effect different for people who seek to avoid news about outgroup victims?
Grant, Kajii and Polak (2000) show that certain families of non-standard utility revert to expected utility when combined with intrinsic preferences for information. I show this reversion can be circumnavigated by forgoing “Compound Independence”, a common but surprisingly limiting assumption.