Skip to main content
14 events
when toggle format what by license comment
Mar 24 at 1:45 vote accept Jeff
Mar 23 at 23:59 history edited Jeff CC BY-SA 4.0
original question was poorly phrased
May 30, 2022 at 19:42 history edited Jeff CC BY-SA 4.0
said slope, meant intercept
S May 26, 2022 at 1:05 history bounty ended CommunityBot
S May 26, 2022 at 1:05 history notice removed CommunityBot
May 22, 2022 at 3:01 history tweeted twitter.com/StackStats/status/1528209322165035008
May 22, 2022 at 1:06 answer added Sextus Empiricus timeline score: 6
May 22, 2022 at 0:47 comment added usεr11852 Just to state the obvious: The sum is practically $0$. I just ran 200 simulations with the proposed design for the GLMM and I got an $N(-0.000155, 0.001411)$. Sure, it is not as consistently close to zero as the one for the LMM but I wouldn't lose sleep about it... Also while the text of the question refers to two logistic models, the R code refers to a LMM vs a GLMM... Also because in the text you mention fixed-effects: Removing the intercept makes the glm and the glmer estimates of x much more similar (assuming the glm has a as.factor(group) covariate too).
May 21, 2022 at 22:50 comment added Jeff @EdM it is true for any simple linear/logistic models with one random slope. i have added an example.
May 21, 2022 at 22:49 history edited Jeff CC BY-SA 4.0
added example
May 18, 2022 at 12:51 comment added EdM It will be hard to answer this question without seeing more details of the two models. Please edit the question to show the function calls and model summaries, via the code {} tool on the editing toolbar. It would be even better if you could provide a reproducible example for others to play with.
S May 17, 2022 at 23:25 history bounty started Jeff
S May 17, 2022 at 23:25 history notice added Jeff Draw attention
May 4, 2022 at 14:27 history asked Jeff CC BY-SA 4.0