I am doing a research in the chance of EBITDA reporting for companies. As my dependent variable is binary and my data is complete for the companies from 2015-2022 I was thinking to do a fixed effects logistic regression. However a lot of my data is omitted because of all positive or negative values. This is because a lot of companies are not changing their reporting over the years. I was thinking to do the same but with random effects or to do a conditional logistic regression (clogit) to prevent that a lot of data is not used. Which of these seems the best for me, and how can i see this. However i would still like to have fixed effects for year and industry, is their a possibility for this?
Attached Files
