the most recent numbers i've seen from new york city are over 25% now, with the same 3-4% margin of error, suggesting it could be higher than 30% - and getting closer to my deductions.

there was also an antibody study that suggests 30% in a small area of boston, which i'm going to agree looks a little bit shady. but, you have to understand that error works in both directions. pointing out that the sample may be unrepresentative necessarily means that it's just as likely to be an underestimate as it is to be an overestimate. so, that particular neighbourhood might have seen substantive spread through a specific grocery outlet, for example - and the true number in that region could be 40% or higher.

we need more testing...

but, don't ignore the data or write it off as "unreliable". that's bad form.

they keep testing people, and while the results are variable, the conclusions are consistent - way, way, way more people have already gotten this thing than previously realized.

https://www.sciencemag.org/news/2020/04/antibody-surveys-suggesting-vast-undercount-coronavirus-infections-may-be-unreliable