fmsb (version 0.5.2)

Kappa.test: Calculate Cohen's kappa statistics for agreement

Description

Calculate Cohen's kappa statistics for agreement and its confidence intervals followed by testing null-hypothesis that the extent of agreement is same as random, kappa statistic equals zero.

Usage

Kappa.test(x, y=NULL, conf.level=0.95)

Arguments

x

If y is not given, x must be the square matrix that the rows and columns show the ratings of different rater (or repeated measure) and the values indicate the numbers of data having that combination. If y is given, x must be the result of ratings by the first rater (or first time measurement).

y

If given, y must be the result of ratings by the second rater (or second time measurement). As default, it is not given.

conf.level

Probability for confidence intervals for kappa statistics. Default is 0.95.

Value

Result$statistic

Z score to test null-hypothesis.

Result$estimate

Calculated point estimate of Cohen's kappa statistic.

Result$conf.int

A numeric vector of length 2 to give upper/lower limit of confidence intervals.

Result$p.value

The significant probability as the result of null-hypothesis testing.

Judgement

The judgement for the estimated kappa about the extent of agreement, given by Landis JR, Koch GG (1977) Biometrics, 33: 159-174: If kappa is less than 0, "No agreement", if 0-0.2, "Slignt agreement", if 0.2-0.4, "Fair agreement", if 0.4-0.6, "Moderate agreement", if 0.6-0.8, "Substantial agreement", if 0.8-1.0, "Almost perfect agreement".

References

Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics, 33: 159-174.

See Also

Kappa

Examples

 res <- Kappa.test(matrix(c(20, 10, 5, 15), 2, 2))
 str(res)
 print(res)
 Kappa.test(c(1, 1, 3, 1, 1, 2, 1, 2, 1, 1), c(2, 1, 3, 1, 3, 2, 1, 3, 3, 3))