robustse: Estimate robust standard errors for different types of...
In cimentadaj/cimentadaj: My various R functions

Description Usage Arguments Value Examples

Estimate robust standard errors for different types of coefficients

1	robustse(x, coef = c("logit", "odd.ratio", "probs"))

`x`	A linear or glm models
`coef`	A character vector indicating the type of coefficient. The function accepts 'logit' for log coefficients, 'odd.ratio' for odd ratios and 'probs' for probability scaled coefficients.

A matrix containing the estimated coefficient, robust standard errors, z values and p values

library(haven)
dat <- read_dta("http://www.stata-press.com/data/r9/quad1.dta")
mod1 <- glm(z ~ x1 + x2 + x3, dat, family = binomial)

robustse(mod1, coef = "logit")
robustse(mod1, coef = "odd.ratio")
robustse(mod1, coef = "probs")

# See http://www.jorgecimentada.com/?p=214 for a detailed example.