Re:The ACLU Has Lost Its Way
by joshuamorton on Tuesday May 10, 2022 @02:15PM
(#31330461)
Yes, and in the case of small N (which, like, is the case here), those are problematic to disclose!
[ Parent
| on HN ]
How so? You'd have to figure out the N in a population of hundreds of thousands (the state's prison population) without any other variables to go on.
Re:The ACLU Has Lost Its Way
by PuppyTailWags on Tuesday May 10, 2022 @04:43PM
(#31332444)
It breaks down the N by location, which... what defines a location here? Is it a state? Is it a single prison? For a small N, narrow location is definitely a de-anonymization factor.
There's no mention of any narrow locations. The numbers are at the state level per the request in the article.
Re:The ACLU Has Lost Its Way
by tsimionescu on Tuesday May 10, 2022 @03:09PM
(#31331255)
How would knowing "1 Transwoman is currently housed in a female prison on the territory of Washington state" in any way risk causing harm to that transwoman, or help you identify her in any way?
It doesn’t. I think GP is confusing small cell counts in micro data (a real problem) with small numbers in aggregate data (sometimes a problem).
In this case, just a count and prison is not a privacy issue as someone would have to already know the individual is trans to identify them. And that’s the only information contained in the data release.