Anonymous posted on Tuesday, June 07, 2011 - 3:01 pm
I need to analyze incremental validity of an assessment measure (X) using Mplus. To do so, I begin the model with a single predictor variable, then add a second variable to test its incremental validity in predicting the criteria over the first variable. An adequate incremental validity of the X will be signified by a statistically significant difference between the values of adjusted R2 for the first and second steps of the model. How do I calculate the change in R-square and, more importantly, the significance of that change in Mplus ?
I actually don't know. You need the asymptotic covariance of both (adjusted) R2's and take into account that they are not independent. Anyone?
Anonymous posted on Tuesday, June 07, 2011 - 7:08 pm
Is there any other way to test incremental validy of a measure, besides testing the significance of R-Square change?
Anonymous posted on Tuesday, June 07, 2011 - 7:40 pm
For example, in one article (In press, not published yet), the authors analyzed incremental validity of a newly-developed assessment measure by using multilevel random intercept-only regression modeling. I beleive this method is explained in your user book under EXAMPLE 9.1( TWO-LEVEL REGRESSION ANALYSIS FOR A CONTINUOUS DEPENDENT VARIABLE WITH A RANDOM INTERCEPT). In your opinion, is this a right way to test incremental validity? Any other suggestions?