st: Problem Generating ICC Using xtmixed

[Dylan Galos <galos002@umn.edu>]

I'm using xtmixed (using the item as the dependent variable and the 
subject as the group variablle) and the postestimation estat icc to 
calculate ICCs for a test-retest reliability study of several 
Likert-type items. For only one item, the LR test (vs linear regression) 
had a chibar2 value of 0,  p = 1, and the icc failed to calculate with 
error r(498) and the error message "Maximum number of iterations 
exceeded" (the xtmixed model itself took 3 iterations to converge). As I 
looked through the data, the responses changed between time1 and time2 
for many of these subjects (it's probably unreliable), only in this one 
item, and I'm not sure what a valid next approach would be to estimate 
this, or how to interpret this result/error.

If it helps, my syntax was:

xtmixed var || surveyid:, mle
estat icc

Thank you for your help.

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/