Thread by @HistDem

Somehow I missed this subtweet. @frankmcsherry finds it "mystifying" that "(some) demographers" (i.e. me) have "contempt for the the privacy of their subjects." In his blog quotes, he screenshots my tweets and characterizes them as "just embarrassing."/1

https://twitter.com/frankmcsherry/status/1375939165351776266

In my tweets, I was objecting to @john_abowd's characterization of a 45% match rate between his so-called database reconstruction and the actual data on four characteristics as "highly accurate."

Since then, I have demonstrated that the supposed reconstruction had a match rate only slightly better than one would expect by chance.
assets.ipums.org/_files/mpc/wp2…

But the larger point is that the obsession of some computer scientists with this issue is out of proportion to the actual risks, and ignores the compelling need for accurate data.

Over 60 years of public-use decennial census summary files and microdata, there is not one single case of anyone being identified by someone outside of the Census Bureau.

The Census Bureau agrees that that "confirmation of reidentified responses requires access to confidential internal Census Bureau information … an external attacker has no means of confirming them”

And even if somehow they were reidentified, it would cause minimal harm. The Census contains no questions useful for identify theft, and mostly consists of basic information that would be much more easily obtained through other sources.

So the risks are essentially nonexistent, but the harm caused by the elimination of accurate data about the population are huge.

I do not have contempt for the respondents, but as a rational person I recognize that they are not endangered by the publication of census tables.