st: xtmixed: Cross-level interaction only with random coefficients?

[enfinity@web.de]

Hi there:

I have a question regarding cross-level interactions in multilevel models:

The majority of examples that I find on the internet and in books places the level 1 covariant in the random-coefficient part of the xtmixed command, such as the following example:

xtmixed y level1 level2 level1Xlevel2 || country: level1, nolog covar(un)

My question is, why is this necessary? When I run the model as simple random-intercept version without random slope, what would be wrong with that:

xtmixed y level1 level2 level1Xlevel2 || country: , nolog covar(un)

Thanks for any comments on this issue... I just try to understand the logic behind it.

Thanks!
*
*   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/