3 ms·
Understanding this study is very difficult because they haven't released all of the details about statistical procedures, recruitment, raw data, etc, but I woul
by lalaland1125 7y ago
Understanding this study is very difficult because they haven't released all of the details about statistical procedures, recruitment, raw data, etc, but I would be worried that this study will have similar issues to the recently published Santa Clara study (both use the same lab test and share several authors).
I highly recommend that people read https://statmodeling.stat.columbia.edu/2020/04/19/fatal-flaws-in-stanford-study-of-coronavirus-prevalence/ https://statmodeling.stat.columbia.edu/2020/04/19/fatal-flaw... for information about for information on how the Santa Clara study messed up. TLDR: Incorrect statistics, incorrect analysis, and incorrect sampling probably led to a vast overestimate of both prevalence and certainty in that high prevalence estimate.
- tempsy 7y agoThe huge difference is that the LA one claims to have tested a random sample of people while the Stanford one did not.
- tunesmith 7y agoDoesn't sound like that is a sufficient difference, however, if they also screwed up the stats and the analysis.
- eightysixfour 7y agoSome of the same authors, possible/likely they are using the same test.
- longtimegoogler 7y agoThat isn't huge. The biggest issue with the Stanford study was that they didn't take into account the Confidence Intervals of the Specificity.