4 ms·
"Everywhere was growing with roughly the same exponent before lockdown." No, they weren't. You're just asserting that this is true. There's effectively no way
by timr 7y ago
"Everywhere was growing with roughly the same exponent before lockdown."
No, they weren't. You're just asserting that this is true. There's effectively no way to know what the doubling rate was in a place like Boise before the national lockdown, because the numbers were in the low single digits:
https://covid-19.direct/metro/Boise?tab=glance https://covid-19.direct/metro/Boise?tab=glance
But even ignoring that...the site you're linking to shows a wide range of slopes on the per-state graph. You just think they look the same, because the log Y-axis compresses dramatic differences in scale, particularly near the origin. New York, NJ, Michigan and Louisiana are well above the diagonal line, while states like Wyoming, Montana, Vermont, Maine etc. are all well below it.
- SpicyLemonZest 7y agoOne of the things we urgently need to figure out is how to prevent bad data visualization from leading to bad conclusions. Statistical malpractice isn't just a meme phrase anymore.
- ajross 7y agoIdaho on 20 March was at 23 confirmed cases (above the "low single digits" metric you picked, but you can pick other points too and get the same result). On 31 March it hit 515. That's a doubling period of less than three days. It was spreading every bit as fast as New York was. It looks like exactly the same chart with a different constant factor. Yet you insist that these identical results are somehow due to different underlying behavior? Why? Where's the research showing that? Because the pretty obvious hypothesis is that the disease was spreading the same way because it's the same disease and doesn't care about whether you live in Boise.