Vision scientists Johannes Haushofer and Nancy Kanwisher claim, in a recent PNAS paper, that “Both sides retaliate in the Israeli-Palestinian conflict.” They use time series analysis to examine the time series of Israeli and Palestinian attacks, in the form of killings and (on the Palestinian side) the firing of Qassam rockets from Gaza into Israel. They find that both sides show retaliatory patterns in their attacks — in contrast, I guess, to the work of some other scholars, who find Israel retaliatory and Palestine unconditionally violent.
This is the kind of creative use of quantitative talent that I’d like to see more of from psychology. I’d have liked to see more direct engagement with the studies they criticize — their capstone finding is that the firing of Qassam rockets increases after Palestinians are killed, but they also find that killings on both sides follow a retaliatory pattern, which I gather other studies don’t. Presumably some detail of either data sampling or statistical analysis is relevant here, but without examining the other literature I can’t tell what. I’d also like to know if there’s a plausible modeling scheme that would make Israelis seem unconditionally violent — that is, just how flexible is this sort of modeling relative to this sort of data, in terms of the conclusions you can extract?