Check fixed effects assumptions for likert scale variables

[Fa Anaira]

Hello everyone,

I am working with an ordinal dependent variable, which I created a dummy for (1-3=0 and 4-6=1 named h004_bin) so became binary and a binary variable is the independent variable. Now I want to check the zero conditional mean assumption which is the linearity of residuals as I read on here.

my regression looks the following:

xtreg h004_bin covid i.nettocat i.geslacht i.lftdcat i.burgstat i.oplmet i.year, fe robust

covid is the binary dependent variable and the i.XX are my controls

How can I check if my data meets the assumption?


Thank you in advance! I appreciate every help

[Carlo Lazzaro]

Fa:
do you mean that you're dealing with a linear probability regression?
If that were the case, as you've invoked non-default standard errors, you should switch from -hausman- to the community-contributed module -xtoverid- to test whether -re- (instead of -fe- is the way to go). Please note that, being a bit old-fashioned, -xtoverid- does not support -fvvarlist- notation; therefore, you have to prefix your -xtreg,re- code with -xi:-.
That said, a panel dataset with an ordinal categorical variable is usually a job for -xtologit- that, unfortunately, provides the -re- specification only.