st: RE: Analyzing NAEP Data in Stata

["Verkuilen, Jay" <JVerkuilen@gc.cuny.edu>]

Bob Fitzgerald wrote:

>>I would be interested in hearing from any list member who has used
Stata 
for NAEP student-level analyses.<<

I did a little bit with it in the past but ended up using HLM 6.0 for
most of the work (for reasons separate from the analysis). 


>>The NAEP student data (National Assessment of Educational Progress),
are 
collected during a large national assessment conducted about every two 
years in the U.S. NAEP items are constructed using IRT methods, and the 
item selection is based on a matrix sample. The NAEP data contain five 
plausible values for each response.<<

Correct, which means you need to deal with the fact that the achievment
variables are multiply imputed (that's what the "plausible values" are).
Other people have mentioned the facilities that exist in Stata already. 



Rodrigo Alfaro wrote:

>>There is a very good information on those formulas and more at
http://www.stat.psu.edu/~jls/mifaq.html. You will note that it is "good"
to know the procedure that generated your datasets. Schafer (1997)
suggests consistency between the imputation-model and the
analysis-model. For example, if in the imputation-model you controlled
by family characteristic but in your analysis you could not omit those
vars.<<

Interesting that Schafer said that. From what I understand of Rubin's
position, it is a good idea to include extra variables even if you don't
intend to analyze them. For instance, there are variables the
administrators of NAEP have that are not available generally and I
believe they include them in the multiple imputation. 

Jay

*
*   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/