glm_table: Produce a formatted generalized linear model table

In audreyrenson/clinRes: Convenience functions for the NYU Lutheran Clinical Research Offic

Description

This function returns a formatted table of generalized linear model results.

Usage

glm_table(fit, digits = 4, p.digits = digits + 1, se = "normal",
  intercept = FALSE, fmt = FALSE, fun_coef = if (class(fit)[1] == "lm" ||
  fit$family$link == c("identity")) I else exp, id, ...)

fmt.glm_table(glm_tbl, tbl_colnames = c(colnames(glm_tbl)[1], "95% CI",
  "p-value"), tbl_rownames = rownames(glm_tbl), digits = 2,
  p.digits = digits + 1, p_fmt = function(p)
  format.pval(round(as.numeric(p), p.digits), eps = 0.001),
  beta_fmt = function(b) format(as.numeric(b), digits = digits, nsmall =
  digits), ci_fmt = function(col1, col2) paste0(beta_fmt(col1), " to ",
  beta_fmt(col2)))

Arguments

fit	An glm fit object of class glm.
digits	Integer. The number of significant digits to return in coefficients
p.digits	Integer. The numeric of significant digits to return in p-values
se	Character. The type of standard error to use, "normal" or "robust".
intercept	Logical. Whether or not to include the intercept row in the returned table.
fmt	Logical. If true, passes the returned table to fmt.glm_table, returning a publication-ready table.
fun_coef	Function to apply to the coefficients. By default, exp if link is "log" or "logit", otherwise identity.
...	Arguments passed to fmt.glm_table.

Value

A matrix with coefficients, 95% confidence intervals, and p-values.

Examples

## Basic logistic regression example
n <- 50
x <- rnorm(n)
y <- sample(c(TRUE, FALSE), size=n, replace=TRUE)
fit <- glm(y~x, family=binomial)
glm_table(fit)

## Use se="robust" to get risk ratios from poisson regression
fit <- glm(y~x, family=poisson)
glm_table(fit, se="robust")

## Get a nicely formatted table
glm_table(fit, fmt=TRUE)

audreyrenson/clinRes documentation built on Feb. 14, 2020, 10:27 a.m.