Multilevel modeling and heteroscedasticity

[andreek]

Hi,

I would like to check for assumptions in a multilevel model. I know that
for multicollinearity it is possible to run a normal regress command
since it is not based on the estimates, and that I can test for normal
distribution in standardized residuals with the xtmixed command. But
what about checking for heteroskedasticity? How do I do this in
multilevel modeling?

The model I want to check is
. xtmixed health1 eduyrs woman age_41_55 age_56_70 age_71_85 age_86_105 feelabout_income2 feelabout_income3 feelabout_income4 sclmeet discuss_personal control_work ESS2_dummy ESS3_dummy ESS4_dummy ESS5_dummy ESS6_dummy unemployment gini finland norway sweden ireland unitedkingdom belgium germany netherland portugal spain poland slovenia edu2004 edu2006 edu2008 edu2010 edu2012 finlandeducation norwayeducation swedeneducation irelandeducation unitedkingdomeducation belgiumeducation germanyeducation netherlandseducation portugaleducation spaineducation polandeducation sloveniaeducation finlandgini norwaygini swedengini irelandgini unitedkingdomgini belgiumgini germanygini netherlandsgini portugalgini spaingini polandgini sloveniagini [pw=weight] || cntry_level3: , || level2: , ml variance


Regards Andrea

[Carlo Lazzaro]

If Andrea is using Stata 13.1 (as, without further clarifications, he is supposed to), he may want to take a look at the example n. 7 under the entry -mixed-, page 309-311, Stata 13.1 .pdf manual.
Best regards,
Carlo

[Richard Williams]

Andreek, within the space of a half hour you seem to have started 4 threads that are identical or nearly identical. Try not to do that! If you've made a mistake, I believe you can either edit your post or at least add a correction to it.