Level 1 variables predict level 2 var...
Message/Author
 Li Lu posted on Monday, February 04, 2013 - 12:45 pm
Hi all,
I am learning Mplus and have used HLM.
I recently learned that Mplus can use level 1 variables (e.g., individual IQs) to predict level 2 variables (e.g., group performance).

How can this be done? Would you kindly share me some codes or resources to look into it?

Thanks!
 Bengt O. Muthen posted on Monday, February 04, 2013 - 6:10 pm
It is the level-2 random intercept part of an individual's variable that predicts the level-2 variable. See UG ex9.4 where y plays this role.
 Abbie Lambert posted on Sunday, March 27, 2016 - 1:44 pm
Hello:
I am very new to MPlus and it was recommended on other sites for Multilevel modeling. One question I have that I am having a hard time finding a reliable answer to is - Will MPlus allow for different sample sizes between variables? My IV (employee level) is rather large at approx 3,000 while my DV (restaurant level) is around 250? If this is not allowed in MPlus, what procedures would you recommend I use to have the IV sample size match the DV sample size.

Thank you so much!
 Bengt O. Muthen posted on Sunday, March 27, 2016 - 5:53 pm
Sounds like you are saying that you have 250 clusters and a total of 3,000 observations with employees nested within restaurants. That's a typical data setting for multilevel analysis. But maybe I am misunderstanding you.
 Matt Boulter posted on Thursday, June 20, 2019 - 3:32 am
Hello there,

I am attempting a cross level interaction which a level-1 variable (x) predicts a level-2 variable (y, this variable is measured at level-1 but I want to represent it at level-2 eg. group agg. of cohesion), and this effect is moderated by a level-2 moderator (W), reading through your work this is my interpretation of how to model this, is this correct?:

%WITHIN%
s| y on x;

%BETWEEN%
y s on w;
y with s;

Thanks!
 Bengt O. Muthen posted on Saturday, June 22, 2019 - 6:28 am
Yes.