Timeline for What can explain large Odds Ratios 95% confidence intervals in multilevel binomial logistic regression (glmer)?
Current License: CC BY-SA 4.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 28, 2022 at 8:10 | history | edited | AnaG | CC BY-SA 4.0 |
deleted 125 characters in body
|
| Oct 20, 2022 at 7:46 | vote | accept | AnaG | ||
| Oct 16, 2022 at 18:34 | comment | added | Russ Lenth | ... and the SE for comparing two coins with 24 flips each is sqrt(.25/24 + .25/24) = 0.144. And here, you are allowing for additional random effects and a much more complex design. You can't make a silk purse out of a sow's ear. Be realistic | |
| Oct 16, 2022 at 18:28 | comment | added | Russ Lenth | In the end, you've got what you've got. Unless you want your software to lie about the precision of your results, if the SEs are too big, then you need more data. Even in the simplest coin-flip experiment, 48 flips is a pretty small sample size for estimating P(heads); the SE is about 0.072. | |
| Oct 16, 2022 at 11:14 | answer | added | dipetkov | timeline score: 2 | |
| Oct 13, 2022 at 12:51 | history | edited | AnaG | CC BY-SA 4.0 |
Adding a link to the dataset
|
| Oct 11, 2022 at 14:28 | comment | added | AnaG | Thanks @mdewey, I am using an equivalent gender ratio so I have 48 individuals (clusters), 24 are men and 24 are women | |
| Oct 11, 2022 at 13:16 | comment | added | mdewey | How many men and women do you have? | |
| Oct 11, 2022 at 12:59 | answer | added | Shawn Hemelstrand | timeline score: 2 | |
| S Oct 11, 2022 at 12:09 | review | First questions | |||
| Oct 11, 2022 at 12:15 | |||||
| S Oct 11, 2022 at 12:09 | history | asked | AnaG | CC BY-SA 4.0 |