John Allen Paulos wrote A Mathematician Reads A Newspaper in 1996. However, the newsroom innumeracy he identified then still exist today. Exhibit A: the New York Times, 27 October 2022.
A new series of House polls by The New York Times and Siena College across four archetypal swing districts offers fresh evidence that Republicans are poised to retake Congress this fall as the party dominated among voters who care most about the economy.
The polling data the writers linked to do not support the claim.
First, these are very small data sets. Each of the four contests is (probably) within the margin of error (MOE), which is huge: more than 5% for each race.
No GOP candidate leads.
In two races — Kansas 3rd and Pennsylvania 8th — the Democratic candidates have a lead that may be slightly greater than the margin of error, depending on the rounding used to report their lead.
If this November’s congressional elections were held today, which candidate would you be more likely to vote for…
Second, the writers claim that the GOP “dominated among voters who care most about the economy.”
That may be true, but the published cross-tabs provide no data for that claim. Given that it would be a sub-sub group, the MOE would be even greater than for the published data. It is highly unlikely that the polling data support that claim.
This fiction aligns with the paper’s October 17th narrative that “independents, especially women, are swinging to the G.O.P.”
Last week, that “independent women” data point had an MOE of 12%. Today, the page contains only the main survey MOE of 4%.
Here’s the thing: the @nytimes poll last week was a statistical tie.
It had a *4% margin of error* on the full results and 12% on the MOE on the 💩 like ‘independent women shift red.’
— 🏳️🌈 Kathy E Gill 😷 🇺🇦mastodon.social/@kegill (@kegill) October 27, 2022
The election is about the future of the country. Major news media like the NYT seem to be cheering for our downfall by their news choice and slant.
By the way? All of the writers and editors and managers at the Times REALLY need to read A Mathematician Reads the Newspaper.
#scitech, #media, #statistics
Kathy Gill
Daily posts, 2022-2023 (279/365)