It’s even worse, they did some post adjustment afterward to make up for the fact that the volunteers weren’t properly distributed geographically. Makes sense, but the further people were from the test sites the more likely they tested positive, so they end up adjusting everything upward.
Why is that an issue? Again, self-selected volunteers. It’s already a problem that people who think they might have had something would be more likely to seek confirmation. But that seems doubly true if you had to travel an hour to a test site.
The good news is there are many other serological studies happening now, so we should get some better data.