Mixed Model Notation - A Simple Explanation

Рет қаралды 24,073

3 жыл бұрын

This is part of the multivariate playlist: • Multivariate Statistics
If you missed part 1 of the video, visit this link: • Mixed Models, Hierarch...
Learning Objectives
#1. Understand subscript notation in Regression
#2. Understand subscript notation for mixed models
#3. Understand Random effects anova, random intercept, random slopes, random slopes/intercept
#4. Understand what fixed/random effects mean
#5. Understand the gammas

Пікірлер: 30

@alifarahani4398 Жыл бұрын

This guy is AMAZING. The best video you can find on the internet on this concept.

@khosbeh5467 3 ай бұрын

GREAT work. Fantastic presentation. You make complicated concepts so understandable. I have been an instructor for 37 years and I admire your abilities and execution. KEEP UP the great work.

@QuantPsych 3 ай бұрын

Thanks!

@seaz9992 Жыл бұрын

These videos are so well done and informative. I took MLM a bit ago and forgot some of the notation, this was very helpful!

@Robin-wo8gj Жыл бұрын

Thank you so much Dustin ! Your videos are a lifesaver.

@tahamansoor6548 Жыл бұрын

Awesome!! This was so complicated for me before this video. You made it so easy. Thanks!

@LeoWei-dd9to 2 жыл бұрын

Your video is so helpful! Thank you!

@nghieply7951 2 жыл бұрын

Thank you for the video. This is what i need to understand about LMEM

@brandyhorne1997 2 жыл бұрын

You are awesome - thank you!

@lenag3329 3 жыл бұрын

awesome explanation, thank u!

@lucagiraldi372 3 жыл бұрын

Amazing video! at <a href="#" class="seekto" data-time="910">15:10</a>, when you talk about the "-1", this is to indicate that the intercept should not be random, not that the slope should not be random, isn't it?

@rHk217 Жыл бұрын

I was thinking the same thing.

@fjumi3652 Жыл бұрын

Thank you for this video! :D :D :D

@romainouldammar3660 2 жыл бұрын

Very good video! Thank you. In the followings models (Alcuse~1+AGE_14+(1|ID ); Alcuse~1+AGE_14+(AGE_14|ID)), does the random intercept (for both models) correspond to the variation of alcohol consumption at Age 14 ? If yes, could you explain why the estimation differs (0.5966 vs 0.6355)?

@thelateknights 2 жыл бұрын

Is there a certain number of levels that should be present to use a mixed model? Like if we look at gender assigned at birth, there would be 2 levels, would it be appropriate to include that in a mixed model, or just dummy code that?

@vincentcabibel9838 3 жыл бұрын

Hi, very great as usual ! I'm trying to use flexplot to visualize my lmer models however I consistently get a "$ operator not defined for this S4 class" error, even when using your code (without any change) at <a href="#" class="seekto" data-time="630">10:30</a>. Any idea of what's happening ?

@QuantPsych 3 жыл бұрын

When was the last time you updated flexplot? I think in the Fall, I fixed all those bugs. Try installing from github.

@vincentcabibel9838 3 жыл бұрын

@@QuantPsych Installed right before using few days ago, but I tried on another computer and it worked. Seems there is an issue with the computer then. Thanks for the great package !

@fazlfazl2346 6 ай бұрын

@0.36 I have a question. We usually have two parts for regression. One is before fitting the model and the second is prediction using the fitted model. What is confusing for me is that (Depression_i = B0_i + B1*Stress_i) seems to be something that we will use during prediction because at prediction we are interested in a single individual. While fitting the model we take all individuals together to fit the regression line so we do not think in terms of single individuals (the ith individual) but all the individuals. So @0.36 seconds, what part of the model are you referring to? If we are fitting the regression line using OLS, how does the subscript I matter while fitting using OLS? Please clarify.

@kashyappandya6357 2 жыл бұрын

As per explanation at 5.15 of video, for intercept : that people who come to respective therapist have already average score of depression is 11 and for slope : where average or general slope is +0.5, but in given group slope is -0.25, which shows efficiency or effectiveness of therapy(for example, may be), which is low compare to average slope across group. am I right? please somebody answer...