hey 👋 I’m Gosha

my tools:

currently working on:

usefulr's People

Contributors

Watchers

usefulr's Issues

Overhaul tests for all the new functionality

All the tests are outdated and should be expanded and re-written

Move all output formatting from analyses() to table_out()

Currently treat_only, stars and round_digits are arguments for usefulr::analyses() which does not make sense, since the function is mainly for estimation. Consider moving this functionality and all output formatting to usefulr::table_fun()

This most likely will mean that the following chunk should be moved to usefulr::table_fun()

out <-
    dplyr::mutate(estout,
                  printout =
                    ifelse(is.nan(estimate), "-- [--]",
                           paste0(fround(estimate, digits = round_digits),
                                  ifelse(stars, add_stars(p.value), ""),
                                  " [", fround(std.error, digits = round_digits), "]")),
                  estimate =
                    round(estimate, digits = 3), std.error = round(std.error,
                                                                   digits = round_digits),
                  p.value = round(p.value, digits = round_digits))

out <- dplyr::select(.data = out,
                                  term, estimate, std.error,
                                  printout, p.value)

Rename functions

Not sure is the function naming is very systematic. In particular I don't like

table_fun()
clusterboot_mfx()
vcov_cluster()
ocme_mod()

Consider adding travis checks to the package

Extend robust = TRUE option in usefulr::analyses()

Currently robust = TRUE in combination with cluster = NULL is only supported for model = "lm" in usefulr::analyses(). Given how much Stata-people like , robust, should probably extend this to other models.

Recommend Projects