(I apologize if this question is off-topic.)
I would like to apply a method developed for time-series cross-section data to a balanced panel dataset to estimate the effect of a treatment. What I mean by that is, instead of using a standard panel data model (e.g., Pooled OLS, fixed-effects, etc.), the method I would like to use "breaks" the panel data into cross-sections and analyze the data that way (i.e., through multiple cross-section comparisons). I want to believe there is no problem analyzing my balanced panel data that way, but perhaps I am missing something?
As a side note, I am using the matching method developed by Imai et al. 2018 (that actually uses a nonparametric generalization of the difference-in-differences estimator), implemented through the 'PanelMatch' package for R. The authors of the package are currently working on how to implement that method to continuous treatments (which would be the ideal case for me), but again, I wonder if I could apply that method to a balanced panel data ignoring the fact that I have a continuous treatment. Thank you in advance.