For a grant application, we are interested in comparing models for the
predictive power of risk and protective factors across ethnic group (white,
black, hispanic, native america). The data are from a longitudinal complex
survey design with N of about 5,000 of children in foster care in the US.
There are essentially 3 waves of data, but three are of interest.
Most key variables are continuous, but some are binary.
We are interested in two types of models:
1. Models for predicting wave 4 functioning from wave 1 risk and protective
factors. These would use survey regression procedures
2. Growth curve models for predicting rate of change (slope) from wave 1
risk and protective factors. These would use xtmixed (or gllamm) with
weights and clustering, with robust standard errors.
In particular, our question concerns ethnic group similarities and
differences in predictive models.We are willing to view whites as a
reference group and examine differences in models for other groups 1 by 1.
As I see it, this is essentially an interaction of ethnicity X predictor.
The problem is that we have multple predictors and want to ultimately do a
multivariate model including a number of precdictors. HYet, models with many
interactions are essentially impossible to interpret.
I can only see looking at interactions one predictor at a time, to see which
predictors differ by ethnic group. Then constructing a model separately for
each ethnic group. Not that satisfactory.
Questions:
A. If we use my strategy, is there any reasonable way to compare the final
models as a whole? If so, can we compare particular paths.
B. Is there a better alternative?
C. How might one estimate power, to at least have something to put in a
"power" section of the grant? BTW, at this point, we can find no info on
design effects in the sample.
Any ideas welcome.
Thanks.
Stephen Soldz
Director, Center for Research, Evaluation, and Program Development Boston
Graduate School of Psychoanalysis
1581 Beacon St.
Brookline, MA 02446
ssoldz@bgsp.edu
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/