Mplus Discussion > Growth Modeling of Longitudinal Data >
 Daniel Rodriguez posted on Thursday, November 07, 2002 - 5:03 pm
I have a question about interpreting change in the log odds when dealing with a non-linear model. I released the final 2 of my four time points, improving the fit of my measurement model to the data. My factor loadings from the slope factor are 0, 1, 1.543, and 2.055, for times 1-4, respectively. My Beta (slope mean) is .134. Also, is this interpretation affected by the fact that my data is measured on four unequal time points? Time points1-3 are spaced by 6 month intervals. Time four occurs 1 year after time 3. Unfortunately, the linear model (0,1,2,4) didn't fit the model quite as well as when I freed the final two time points. I'd appreciate your response. I am much obliged for your help.
 bmuthen posted on Thursday, November 07, 2002 - 8:32 pm
When your time scores are linearly developing like 0, 1, 2, 3, the change from time to time is constant, and is equal to the beta-mean. If time scores are 0, 1, 2, 4, the change from the 3rd to the fourth time point is (4-2)*beta-mean. The change from 1st to 4th is (4-0)*beta-mean. With estimated time scores, you use the estimated values in an analogous way.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message