gkappa: Generalized kappa

In raredd/ragree: Rater agreement

Description

Calculates generalized kappa for k-raters

Usage

gkappa(ratings, data, id = "id", rater = "rater", score = "score")

Arguments

ratings	n * m matrix or data frame; n subjects, m raters
data	data in long format with three columns: ids, raters, scores
id, rater, score	column names corresponding to IDs, raters, and scores

Value

A list containing the generalized kappa coefficient, the estimated variance, the test statistic, and corresponding p-value comparing the statistic to a standard normal distribution.

	$method	agreement method
	$ragree.name	method type
	$subjects	number of subjects
	$raters	number of raters
	$categories	number of categories
	$value	generalized kappa statistic
	$variance	variance
	$statistic	test statistic
	$stat.name	test statistic name
	$p.value	p-value comparing statistic to standard normal distribution
	$detail
	...$p_hat_j	proportion of all classifications assigned to category j; the pairwise agreement
	...$p_bar	overall agreement
	...$p_bar_e	expected chance agreement

Author(s)

Robert Redd [e-mail]

Examples

## long format using data argument
gkappa(data = disease.l, id = 'patient', rater = 'rater', score = 'category')

## wide format using ratings argument
gkappa(ratings = disease.w)

## example from irr package
library('irr')
data(diagnoses)

gkappa(ratings = diagnoses)

## fleiss's variance calculation is slightly different
kappam.fleiss(diagnoses)

raredd/ragree documentation built on March 25, 2021, 1:42 p.m.