Dear all,
so far I have been a read-only member and have several times profited greatly from your knowledge. Now I have a statistical problem with the data I'm working on and hoping for your help.
The problem goes as follows: I intend on doing a probit regression and both the concerning variables are binary. I want to analyse the effect of experiencing a loss during the financial crisis (loss) on the future participation of the stockmarket (market_part). However, these two variables are closely linked as experiencing a loss only occurs when you participated in the market in the first place.
What would be the best statistical procedure to handle this?
Would the use of instrumental variables such as ivreg2h by Lewbel be suitable?
Worth mentioning is probably also the fact, that I'm working with imputed data.
Thank you in advance for your help. If anything is unclear please let me know.
Eva