A continuously-updated post collecting times I've made errors
The point of this post is to collect errors that I’ve made. If you believe I’ve made an error, tell me, and if you’re correct, I’ll add it here. This does not count spelling errors, typos, or differences in the interpretation of ambiguous language.
Erroneous Claim: ‘The weakening link between age and crime is not just a reflection of the homicide rate.’
Detail: I made that claim in this Substack post. I was later directly messaged with a referral to additional data and a means of decomposing the sources of the change and, as it turns out, I was wrong and I could have known better had I been familiar with the methods. A proportion indistinguishable from all of the change in the age at peak perpetration was attributable to the decline of the homicide rate in the data I used.
Update Status: Writing a revised post is on the list, but not a high priority. The offending post still contains the erroneous inferences, but it also contains good data and the inference is not very harmful or that much of a stretch, so I’ve left it up.
Coding and Calculation Errors: On X/Twitter, I posted a series of graphs of results regarding the breeder’s equation and I was negligent, so the presented numbers were incorrect.
Detail: I borrowed code and misinterpreted it while coming down from a bout of pneumonia, leading to calculation errors and nonsense numbers.
Update Status: The offending posts have been removed so no one else can be misled by my error. An update has been posted.
Misconception: On X/Twitter, I posted that the FDA’s invisible graveyard was, at one point, filled primarily with gay men.
Detail: This is a widespread belief that has been encouraged by films like Dallas Buyers Club, and I negligently believed it without looking into the topic more deeply.
Update Status: I quote tweeted the original claim—which I left up—with a correction.
Misconception: On X/Twitter, I suggested that hospital desegregation led to reductions in Black postneonatal mortality based on data from a study of desegregation in Mississippi.
Detail: This study has misled many people into believing desegregation improved infant mortality for Blacks in the U.S., but this wasn’t true. Including year fixed effects, the original source’s event-study estimates are nullified or reversed.
Update Status: I wrote an article that, among other things, detailed this error.
Misconception: On X/Twitter, I suggested that Nigerians were highly-successful in the U.S. in terms of occupational and educational attainment, as well as that they had IQs that were on par with American Whites.
Detail: Nigerian Americans are not exceptional in these regards. They bring many doctors to the U.S., but they are not exceptionally high-income as a group, nor do they score on par with Whites in IQ tests, and their occupational attainment in subsequent generations is not on par with American Whites.
Update Status: I wrote an article that detailed the relevant statistics on Nigerians, correcting the error.