@datahistorian.bsky.social
20 followers 20 following 55 posts
Retired History Prof, UW-Milwaukee, https://uwm.edu/history/about/directory/anderson-margo/ History of Data, Politics of Numbers. Latest book [with William Seltzer]: Use and Misuse of the U.S. Census
Posts Media Videos Starter Packs
datahistorian.bsky.social
Excellent. Look forward to wide promotion of this effort.
datahistorian.bsky.social
Can PRB, or a coalition of state data centers pick up the list?
datahistorian.bsky.social
Where in law is there a right for a person/institution/business to know who is accessing their "data"?
datahistorian.bsky.social
The data community needs to weigh in here, not just on “privacy,” but on data “stewardship,” the way data curators see their ethical responsibilities. The administration has a completely different understanding, namely that all “data” is fair game for their agenda.

www.nytimes.com/2025/08/28/b...
Trump Is Claiming Mortgage Fraud to Attack Enemies. Is Your Information Public?
www.nytimes.com
datahistorian.bsky.social
More tinkering: When to report. Who does what, BLS or Census Bureau. What about private sector surveys? What else could one use to measure unemployment, a new question once the Social Security Act [1935] authorized unemployment “insurance.”
More to come...
datahistorian.bsky.social
The “Current Population Survey” was born and has been around ever since. We have a “data series”!
datahistorian.bsky.social
4. Get the tabulation report out fast, preferably 2-3 weeks after the data collection.
5.Who’s going to pay for it? For ~10 years, a government “experiment,” originally the “monthly report on the labor force,” late 1930s-1940s, to work out kinks. [Skipping WWII here, a big part of the story.]
datahistorian.bsky.social
3. Let’s get the definition straight: a person was only unemployed if in the right demographic [not too old, not too young], and actually looking for work the week the sample surveyors came around.
datahistorian.bsky.social
2. Probability sampling [a hot new thing at the time], would permit frequent simple measurement [monthly!], so we’ll need a staff of “field interviewers.”
datahistorian.bsky.social
1. Unemployment fluctuates a lot with the cycles of the larger economy, and needs frequent measurement. A census is the wrong instrument to measure it. Too expensive, too slow for the reporting.....
datahistorian.bsky.social
The “data” were unsatisfactory, irregularly reported, long story short, the Depression accelerated technical work on the issue. [A truncated version of the issues…]
datahistorian.bsky.social
A question for the “occupied” population went on the 1880 census, usa.ipums.org/usa-action/v..., asking the “number of months this person has been unemployed during the census year.” Stayed with changes until 1930 census, including a Congressionally mandated “unemployment census” in 1930 and 1931.
IPUMS USA: descr: MOUNEMP
usa.ipums.org
datahistorian.bsky.social
Unemployment, labor force participation, is hard to measure.

Who does it apply to? Adults? Of what ages? People with their own “businesses”? What about farmers? What about “homemakers?” ....
datahistorian.bsky.social
A series of posts on the history of data.
Like the bridges and roads one drives on, one doesn’t need to know about the bridge builder, unless the bridge falls down, fails. Then all of a sudden it’s really important, which is what happened Friday August 1, 2025, with the monthly unemployment report.
datahistorian.bsky.social
Data series, like the unemployment rate, or the consumer price index, are “infrastructure” – need to be “built,” engineered, “usable,” are the work of many hands, built over time, generate debate and commentary about how to build, which allows historians like me to report on the building process....
datahistorian.bsky.social
Doge went after administrative data, not survey data, so the threats have been in different places than the ones I expected…until yesterday, so now the game is on….
datahistorian.bsky.social
You can “other” them into categories of “exclusion” –e.g. proposals to fiddle the constitutional basis of apportionment. Add a “citizenship” question to the 2030 census. “Exclude” “non citizens” from House apportionment calculations. One version of that. www.congress.gov/bill/118th-c...
www.congress.gov
datahistorian.bsky.social
Or if you stigmatize the person/institution/organization counted in the classification process.....
datahistorian.bsky.social
Questions…

What next? More leaks/revelations from WSJ or other outlets?

Why now? What were Bondi/Bongino thinking by the initial effort to shut stuff down? Why didn’t they continue stonewalling?