Explaining Variance Explained in Mixed Models
Learning Objectives:
#1. Understand what variance explained means
#2. Why variance explained is more complicated for mixed models
#3. Be able to use model.comparison in flexplot
This video is part of my multivariate playlist: This is part of a playlist on multivariate statistics: • Multivariate Statistics
Here are other videos about mixed models:
• Mixed Models, Hierarch...
• Mixed Model Notation -...
Пікірлер: 26
you are a supeeer annoying but also addictive teacher thanks x
@QuantPsych
4 ай бұрын
umm....thanks?
Thank you so much. Teachers often just skip over what "variance explained" mean. Also, at 1:58, I think you mean't to say "Because they are heavier"
@FilippoGianburrasca
5 ай бұрын
* fatter
I'm running RStudio 2021.09.2 Build 382 "Ghost Orchid" Release for Mac OS, and flexplot 0.13.3 and I get the following message when I run the baseline you mention in this video: Error in reformulate(paste0("(", vapply(findbars(f), deparse1, ""), ")"), : unused argument (env = environment(f)). Any thoughts?
Thanks for the nice explanation, Dustin! When I tried to follow this approach I was not shown the R-squared change values. Could this be because I was using the "glmer.nb" function and that the R2 calculation is tricky in that case?
Hi, I got Error "$ operator not defined for this S4 class" when I try compare.fits function. How can I fix that?
What does it mean the ghost.line black line in the compare.fits you are doing ?
7:53 you ran the model at line 56 but then you commented it like it was the model at line 60, which you didn't run (if I'm not mistaken)
model.comparison() from the RRPP package wont let me use an lmer object - do you know why? it says 'Objects must be lm.rrpp fits' :-(
Thank you very much. Can you make a video abot sample size in mixed effects models? thank you
8:40 why it doesn't report also the improvement in the slope variance (i.e. how each cluster slope differs from the grand slope)?
Hi all. I'm running Flexplot version 0.19.1 so quite a bit newer than what Dustin is running here. If you DO NOT have the r squared change output, you'll need to detach the lmerTest package. At least when I detached lmerTest, I got the r squared change output.
I do not see the felxplot in r package. Is it removed?
@QuantPsych
9 күн бұрын
No, you have to install from github.
Your videos are so helpful - and enjoyable - but ... " package ‘flexplot’ is not available (for R version 4.0.2)"
@QuantPsych
3 жыл бұрын
It must be installed from github: install.packages("devtools") devtools::install_github("dustinfife/flexplot")
Hi..can you make short video on how REML estimates variance components in mixed model.. i mean whats goes inside of REML. hoe does it treats the data sets..Thanks...SUbscribed your channel.
I can't load data(math) in R. It gives the message: object 'math' not found.
@tatjanajak
3 жыл бұрын
Sorry, i take this back. The math data is the part of the flexplot project. Just needed to install it properly.
@QuantPsych
3 жыл бұрын
Glad you figured it out!
Voice is not syncronized
Hi, when I use model.comparison for this example, it prints everything except for the $r_squared_change section. Do you know why this may be?
@QuantPsych
2 жыл бұрын
Sorry! It's fixed in the most recent iteration of flexplot. devtools::install_github("dustinfife/flexplot", ref="development")
@yutingcui2485
2 жыл бұрын
@@QuantPsych So sorry, I got the same print in using model.comparison function, the results didn't show R2 change. How do solve it?
What does it mean the ghost.line black line in the compare.fits you are doing ?