Hi,

Hi,


I want to run a multi-level ordinal logistic regression in which people have completed an assessment ('assessoutcome') at different stages of their career. The outcome is 1-3 on an ordinal scale. They do the assessment at careerlevel1 through to careerlevel8

I have done the following, nesting assessment outcome by id:

meglm assessoutcome 'variable' || id:, family(ordinal) link(logit) or

One of the problems is that people enter the cohort at different times and so have different numbers of assessment outcomes, e.g. one person will enter the cohort at the point they are completing the last assessment (careerlevel8) so we do not have any data of assessment outcomes prior to this.

In order to account for this, will a crossed effect model like the one below work?


meglm assessoutcome 'variable' || _all:R.careerlevel || id:, family(ordinal) link(logit) cov(unstructured) or


I hope that makes sense and as always grateful for your help,
Carla