Sat, Apr-18-20, 12:03
|
 |
Senior Member
Posts: 15,075
|
|
Plan: mostly milkfat
Stats: 190/152.4/154
BF:
Progress: 104%
Location: Ontario
|
|
Larger numbers having had the virus suggests lower mortality as a percentage--but I wonder what it says about true current prevalence?
I'm a little suspicious about the recruiting method;
Quote:
The study in Santa Clara County recruited participants largely using Facebook ads targeted by zip code to sample various parts of the community. The study tested 3,330 adults and children.
"They had about five different tents set up and you just kind of drove through and they pricked your finger, took a blood sample and took down your information," one of those participants, Tony Huston, told CNN Chief Medical Correspondent Dr. Sanjay Gupta on his podcast, "Coronavirus: Fact vs. Fiction."
|
Just as testing for the virus itself can be non-representative where it's rationed for those who need it most--people who are sick or otherwise likely to be infected--choosing to participate in the study gives another self-selection bias--just about everybody probably wants to know if they've developed antibodies, but people with good reason to suspect they might have been sick with corona may be more likely to go get tested. Needs more randomization. |