« The percent with antibodies could be 3 percent or it could be zero. That's what the Stanford study really said. | Main | To justify zero false-positive assumption, Stanford study needs a reference sample 10 times bigger »


Feed You can follow this conversation by subscribing to the comment feed for this post.


Would the time factor make a difference.
On any given day x % could test positive and y% negative
Which and what percentage should you test every day?

If you repeat every day what happens re. Errors?


Time: If there is a reason why when someone tests affects the test result, then that factor has to be accounted for. But that doesn't seem like it applies here. Why are you thinking in that direction? (If you're talking about reporting time versus testing time, that's a different issue.)


Let's leave the time aspect for now. Lets take your example at end and think of it applied 5 times to a manufacturing operation. 5 QC tests to eliminate errors before final output. If we start w1000 units what would be tge number of units at end falling into each of your categories


Time: Are you suggesting that another objection to the "test is inaccurate" excuse is that you can do repeated testing?


Not really. But the likely required administration of testing, and or possible pattern of testing in certain environments will likely lead to repetition. If that is true (and it opens the debate) it would mean that accuracy would also need to measured as part of a process.
(Note some manufacturers state their tests are to be used this manner)

Not challenging your statement re politicians and standalone accuracy.

The e.g given more might apply to someone testing a set group over a period of time.


This will fill some testing gaps https://doi.org/10.1101/2020.04.21.20068858 for


Time: thanks for that paper. I only read the abstract and will read the whole paper once I find some time. This study is very valuable, and its value lies in generalizing to the population. It seems like they should bring on a statistician to help generalize the data. From the abstract, it does not appear that they attempted to measure selection bias.

The comments to this entry are closed.

Get new posts by email:
Kaiser Fung. Business analytics and data visualization expert. Author and Speaker.
Visit my website. Follow my Twitter. See my articles at Daily Beast, 538, HBR, Wired.

See my Youtube and Flickr.


  • only in Big Data
Numbers Rule Your World:
Amazon - Barnes&Noble

Amazon - Barnes&Noble

Junk Charts Blog

Link to junkcharts

Graphics design by Amanda Lee

Next Events

Jan: 10 NYPL Data Science Careers Talk, New York, NY

Past Events

Aug: 15 NYPL Analytics Resume Review Workshop, New York, NY

Apr: 2 Data Visualization Seminar, Pasadena, CA

Mar: 30 ASA DataFest, New York, NY

See more here

Principal Analytics Prep

Link to Principal Analytics Prep