1. Pick one variable (at the baseline measurement) in the ATLAS data
set.
-
compute the intraclass correlation based on the output from proc mixed
and proc glm. (Note the value of the intraclass correlation may differ
between MIXED and GLM because of different estimation strategies dealing
with unequal sample sizes). INTRA.SAS
-
Compute the Variance inflation factor.
-
Briefly describe why the intraclass correlation is important.
2. Pick one dependent variable in the atlas data set and run the following
models. (SAS programs to conduct each of these analysis were given out
in class)
-
Proc GLM repeated measures analysis across the 6 waves with only complete
cases. REPGLM5.SAS
-
Proc mixed analysis with 6 waves, including missing data, but not nesting
of schools in conditions. CONTRAST.SAS
-
Proc mixed analysis with 6 waves, including missing data, and including
nesting of schools within conditions. CONTRAST.SAS
-
Plot the means from C and briefly describe the results. Use the contrasts
tests to interpret the results.
Note:You will need to change the sas program to read new variables
based on which dependent variable you select. You will need to:
-
Add the new variables to the keep statement.
-
Recode the scores for the d and cs, and f and es waves.
If you have trouble with this, please ask me, Chondra, Jeanne, or Jeewon
for help.
Back
to the Prevention Statistics Homepage |
|