I did a replication analysis of the Reise study( look at attached file) using a different data set. My output and data shares the same analysis trend. In the original study, the author used PROMIS DEPRESSION as the scale. For my replication, I used PROMIS ANXIETY. So please write in the context of anxiety instead of depression. 
I need you to write a paper on: 
1) what is IRT, why is it important or what are the advantages of IRT in health outcome research 
2) IRT models have basic assumptions, and PROMIS ANXIETY is calibrated using Graded Response Model. But what are some controversies on GRM in healthoutcome constructs? 
3) talk about the differnces in common irt models construct vs. anxiety constructs
4) what is graded response model? what is log logistic model?
5) conclusion: since the yield the same outcome, would you recommend LL model for future research or is it ok to continue with grm?