Timeline for Can we really assign random slopes to between-subject effects in a mixed effects model?
Current License: CC BY-SA 4.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 9, 2020 at 6:49 | comment | added | CatM | @HereroskedasticJim Thank you so much for your prompt response! I was also biased to think that between variables were not an option as a random slope because of reading the same sources as the original poster. This was quite an interesting thing to adjust to, thanks! | |
| Aug 9, 2020 at 6:36 | comment | added | Heteroskedastic Jim | @CatM If for any reason, I thought some coefficient might vary by a grouping structure and the model wasn't too complex to fit, I'd try a random slope on just about any variable. | |
| Aug 9, 2020 at 5:44 | comment | added | CatM | @HeteroskedasticJim Jim Sorry for bringing this back, but I came across it now as I was looking at the same issue. How do you know when to include between variables as random slopes? Would you always go for the maximal random structure as suggested by Barr and decide based on model comparison? | |
| Jul 20, 2020 at 5:50 | comment | added | Robert Long | (+1) great answer ! | |
| May 19, 2019 at 8:27 | comment | added | Dimitris Rizopoulos | @IsabellaGhement I wouldn’t call it “random slopes” for F and M, you have two separate “random intercepts” for F and M. Put otherwise you just say that the random intercepts for F and M have different variances. And if you would allow the error variance to be different between F and M, this would correspond to a sub-group analysis fitting separate models for F and M. | |
| May 18, 2019 at 22:14 | comment | added | Heteroskedastic Jim | @IsabellaGhement yes, it can be re-written that way. It's flexible that way. | |
| May 18, 2019 at 22:11 | comment | added | Isabella Ghement | Could the fixed effects portion of your model be re-written as 0 + F + M? I think that would make it easier to then understand that we're allowing the F and M variables to have random slopes. | |
| May 18, 2019 at 22:02 | comment | added | Isabella Ghement | There are other places on this forum where your answer would be helpful, like here: stats.stackexchange.com/questions/245960/…. | |
| May 18, 2019 at 21:57 | comment | added | Heteroskedastic Jim | @IsabellaGhement you're welcome. | |
| May 18, 2019 at 21:54 | comment | added | Isabella Ghement | Thank you for this beautiful answer, @HeteroskedasticJim! 👍 | |
| May 18, 2019 at 21:53 | vote | accept | Isabella Ghement | ||
| May 18, 2019 at 21:26 | history | answered | Heteroskedastic Jim | CC BY-SA 4.0 |