How to calculate the t-statistics over many estimated coefficients (for the same variable, but across years) in r?

Question

I am working on a paper replication project and i need some advice on the following issue. The baseline regression is in the form:

model <- feols(y_t2 ~ x1_t1 + x2, data = dat, fsplit = ~year)

This regression is ran across years. Assume there are in total 20 years, there are in total 20 annual cross-sectional regressions with 20 estimated coefficients for the key independent variable x1_t . How could I calculate the t-statistics of the 20 estimated coefficients in R, ie, how do i get the correct estimated standard errors of the 20 coefficients (for each regression there is an estimated coefficient and standard errors computed)?

Any help is highly appreciated!

Answer 1

Using CO2 (which comes with R) and also lmList from nlme (which comes from R so no installation required) we can calculate the t values of conc by Plant like this:

library(nlme)
coef(summary(lmList(uptake ~ conc | Plant, CO2)))[, , "conc"][, "t value"]

giving

      Qn1       Qn2       Qn3       Qc1       Qc3       Qc2       Mn3       Mn2 
2.3727920 3.2797927 2.9829021 2.7215324 3.1578555 3.2725825 1.7544113 1.9517375 
      Mn1       Mc2       Mc3       Mc1 
2.9368841 0.6942998 0.8113596 1.3932651