reg_felm: Run a regression and return the results in a tibble. Useful...

In pbaylis/baylisR: Patrick Baylis' R Functions

Description

Run a regression and return the results in a tibble. Useful for running multiple regressions with purrr::pmap. All inputs are fed directly into as.formula, and should follow the relevant specifications.

Usage

reg_felm(lhs, rhs, fe = "0", iv = "0", clus = "0", data_str,
  tidy = FALSE)

Arguments

lhs	Character vector with left hand side variable.
rhs	Character vector with right hand side variable(s).
fe	Character vector with fixed effects. Optional.
iv	Character vector with instruments. Optional.
clus	Character vector of clustering variable(s). Optional.
data_str	String to be evaluated to produce dataset. Use single quotes inside double quotes to avoid need for escaping.
tidy	TRUE/FALSE. Return a tidied data.frame rather than a fitted felm object? Optional.

Value

tibble with summary statistics and fitted output (felm or tidied data.frame).

Examples

# Load data
data("Wages", package="plm")
setDT(Wages)

reg_tibble <- as_tibble(
  expand.grid(lhs = "lwage", rhs = c("exp", "exp + wks"),
              fe = "married", clus = "0", data_str = c("Wages", "Wages[bluecol == 'no']"),
              stringsAsFactors = F))

reg_output <- cbind(reg_tibble, purrr::pmap_dfr(reg_tibble, reg_felm))

# Can also save regression output as a tidied data.frame
reg_output_tidy <- cbind(reg_tibble, purrr::pmap_dfr(reg_tibble, reg_felm, tidy = T))

# These can be used directly in stargazer or huxtable
stargazer::stargazer(reg_output$fit)
huxtable::huxreg(reg_output$fit)
huxtable::huxreg(reg_output_tidy$fit_tidy)

# Easy to select which regressions to display
stargazer::stargazer(reg_output %>% filter(rhs == "exp + wks") %>% pull(fit))
huxtable::huxreg(reg_output %>% filter(rhs == "exp + wks") %>% pull(fit))

pbaylis/baylisR documentation built on Nov. 8, 2019, 6:20 p.m.