Why most polls on a no-fly zone over Ukraine are a no-go (so far) | No. 187 – March 13, 2022

Polls are the result of a complex data-generating process that can be influenced by randomness, unpredictable biases, and, yes flaws in human judgment

One of the reasons empirical journalism is useful when studying people and society is that data offer us an objective barometer with which to measure some phenomenon or — to use the social science jargon — “quantity of interest.”

At least, that’s what data are supposed to do. But it is worth constantly reminding yourself that not all data are created equal. It is the reality of science — an iterative process of experimenting and refining — that we often have to peer into the process that created the numbers we are looking at to know if they are truly 100% reliable.

I have recently found myself reminding people to consider this so-called “data generating process” in public opinion polling, as I increasingly see pundits and journalists citing polls that (claim to) show American support for enacting a no-fly zone over Ukraine. Such a zone would, presumably, give NATO aircraft the authority to shoot down Russian jets flying in Ukrainian airspace. That would almost certainly escalate Ameri…