Cause No. 3: Bias

Many early pandemic predictions pointed toward a similar solution: a left-of-center policy agenda. A she-cession justified universal day care and paid family leave; an eviction tsunami justified stronger legal protections for renters; state and local distress seemed to require what Republicans called “blue-state bailouts.” But if this trend suggests bias at work, where was it coming from?

Goldin believes part of why many forecasts were incorrect is that much of the relevant research was produced by advocacy organizations. The McKinsey report on women leaving the workforce, for instance, was co-published by LeanIn.Org.

Similarly, the Aspen eviction study was co-written not only by researchers from think tanks and academic institutions, but also by three leaders of advocacy organizations. And those authors made judgment calls that perhaps depicted a bleaker landscape than was warranted. As Panfil explained, “The Aspen Institute study, in coming up with the largest number, the 40-million number, included not just people who said they had no confidence in making rent but also people who said they had moderate confidence in making rent.”

I don’t mean to suggest anything more sinister than motivated reasoning. Some advocates may have regarded the coronavirus pandemic as an opportunity to shoehorn in important social policies that they felt were long-justified, and, to a certain extent, they saw in the data what they wanted to see. As Hepburn, the Princeton sociologist, argued, the numbers generated by Aspen may have been useful “from a lobbying standpoint,” and Panfil noted that perhaps “it was helpful to the movement of activists who were pushing for relief measures to be put into place to cite some of these larger figures.”