Regression on time passed until certain threshold is "re-reached" I would like to test the time/duration it takes for a certain variable to reach a certain level (e.g. time passed after companies reach a pre-crisis level at point x again in their recovery process from the crisis. Lets say the ROA was precrisis 5% for company A and I would like to collect the data on the following periods and measure how long it took until the 5% threshold was first surpassed again for each company)
I have read that often Survival models, Hazard Models and Cox Models are used. However given my quite large inexperience with running own regressions I was wondering if also just a simple linear regression or any other model (as I do not need survival probabilities or similar) would be suitable without coming at the cost of reduced "true" effects?
Many thanks in advance!
 A: This sounds like you have what's called "repeated events" of the same type.
One of the major advantages of survival analysis is that it naturally handles the situation when you haven't yet seen the next event happen. For example, say it's been 2 weeks since the last supermarket visit for a participant in the study but you have stopped collecting data. You don't want to throw away the information that the time was at least 2 weeks since the last visit.* Just throwing away that information with a standard regression could lead to bias in any results that you get.
It's certainly worth learning survival analysis for this type of study. The word "survival" (or "reliability") represents where a lot of the interest in this type of analysis originally came from, but the applications are much more general to whenever you care about the time to or between events.
The main vignette for the R survival package has a section on how to set up and analyze data for repeated events.

*When all that you have is a lower limit for a time span, that's called "right censored." I still remember how confusing I found that terminology when I started this type of work.
