Interpretation when DV is categorical PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 tino posted on Wednesday, January 07, 2015 - 5:03 am
Hello,

my question refers to the interpretation of the regression coefficients for a twolevel regression model when the dependent variable is dichotomous and a logit link is used, as in UG example 9.3.

(1) Assuming that the between-group regression coefficients refer to the (continuous) random intercept of u (and given that the solution is unstandardized), would this interpretation be correct:

A one unit increase in W is associated with a 1.26 units increase in the mean score of y between groups

(2) In contrast, given that the within-group regression coefficients refer to the log-odds of the dichotomous dependent variable (and given that the solution is unstandardized), would this interpretation be correct:

A one unit increase in X2 is positively associated with u, given log-odds for the event = .533.

Many thanks for your answer
 Bengt O. Muthen posted on Wednesday, January 07, 2015 - 5:30 pm
1. You want to change "mean score of y between groups" to "cluster-mean of y".

2. Correct, but you can say more than just giving the sign - the value .533 can be restated as a change in the odds of u=1 vs u=0 (using odds = exp(logodds))when x changes one unit.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: