Mar 18, 2020
4:55:08pm
4:55:08pm
Selection bias and the size of n in this case means it's hard to tell how close any of its estimators are to the real
value. Add to that the level of care they received, which I'm pretty certain it doesn't scale.
In the future, I'm sure a scientist will do a true random sample that isn't just looking at old people on a cruise.Or we could also look at the rates in South Korea, where they're testing their population aggressively. It's also flawed, but there's no good reason to take 712 people on a cruise as a the proxy for the rest of humanity and ignore other datapoints.
- Previous username
- mjo
- New username
- Dani Rojas
- Bio page
- mik3
- Joined
- Aug 14, 2008
- Last login
- Jan 5, 2021
- Total posts
- 0 (0 FO)
Messages
Time
3/18/20 7:46pm
3/18/20 11:46pm
3/19/20 6:17am
Because death is a low frequency event, you need a very large sample size for "statistically credible" data
3/19/20 8:14am
3/19/20 8:17am
You do realize that I'm an actuary and understand mortality rates and what "statistically credible" means?
3/19/20 8:28am
3/19/20 9:15am
3/19/20 9:28am
3/18/20 3:54pm
3/18/20 3:58pm
3/18/20 4:07pm
3/18/20 7:48pm
3/18/20 4:00pm
3/18/20 4:15pm
3/18/20 4:16pm
3/18/20 5:07pm
3/18/20 11:17pm
3/18/20 5:50pm
3/18/20 6:05pm
3/18/20 6:25pm
3/18/20 7:44pm
3/18/20 6:09pm
3/18/20 6:12pm
3/18/20 6:10pm
3/18/20 7:39pm
3/18/20 7:44pm
3/18/20 8:33pm
3/18/20 8:24pm
3/19/20 12:02am