When to run conditional models PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Jeannine Brant posted on Friday, February 15, 2008 - 11:36 am
Does it make sense to run a conditional model when the unconditional model is not significant but the variance is significant? I would think that the conditional model can help explain the variance.

For example, the best model fit for the depression variable in my study is quadratic (slope 0.645, quadratic -0.275) but the slope and quadratic parameters are not significant. The variance, however, is significant for intercept (p=0.002), slope (p=0.027), and quadratic (p=0.005). Conditioning the model on gender resulted in significant slope (-4.880, p<0.05) and quadratic (1.756, p<0.05) parameters. When I graphed the trajectories, the unconditional model was of course slightly convex curvilinear. For females, the curve was slightly concave curvilinear. For males, the curve was highly convex curvilinear. Am I interpreting this correctly?
 Linda K. Muthen posted on Friday, February 15, 2008 - 3:28 pm
I would look at the conditional model. It may be you did not have enough power in the unconditional model. I assume that you saw a quadratic curve in your observed data.
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