tern (version 0.9.8)

h_odds_ratio: Helper functions for odds ratio estimation

Description

Functions to calculate odds ratios in estimate_odds_ratio().

Usage

or_glm(data, conf_level)
or_clogit(data, conf_level, method = "exact")

Value

A named list of elements or_ci and n_tot.

Arguments

data

(data.frame)
data frame containing at least the variables rsp and grp, and optionally strata for or_clogit().

conf_level

(proportion)
confidence level of the interval.

method

(string)
whether to use the correct ("exact") calculation in the conditional likelihood or one of the approximations. See survival::clogit() for details.

Functions

• or_glm(): Estimates the odds ratio based on stats::glm(). Note that there must be exactly 2 groups in data as specified by the grp variable.

• or_clogit(): Estimates the odds ratio based on survival::clogit(). This is done for the whole data set including all groups, since the results are not the same as when doing pairwise comparisons between the groups.

See Also

odds_ratio

Examples

# Data with 2 groups.
data <- data.frame(
  rsp = as.logical(c(1, 1, 0, 1, 0, 0, 1, 1)),
  grp = letters[c(1, 1, 1, 2, 2, 2, 1, 2)],
  strata = letters[c(1, 2, 1, 2, 2, 2, 1, 2)],
  stringsAsFactors = TRUE
)

# Odds ratio based on glm.
or_glm(data, conf_level = 0.95)

# Data with 3 groups.
data <- data.frame(
  rsp = as.logical(c(1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0)),
  grp = letters[c(1, 1, 1, 2, 2, 2, 3, 3, 3, 3, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3)],
  strata = LETTERS[c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2)],
  stringsAsFactors = TRUE
)

# Odds ratio based on stratified estimation by conditional logistic regression.
or_clogit(data, conf_level = 0.95)