vcd (version 0.1-3.2)

Kappa: Cohen's Kappa and weighted Kappa

Description

Computes two agreement rates: Cohen's kappa and weighted kappa, and confidence bands.

Usage

Kappa(x, weights = c("Equal-Spacing", "Fleiss-Cohen"), conf.level = 0.95)

Arguments

x

a confusion matrix.

weights

either one of the two options or a user-specified matrix with same dimensions as x.

conf.level

level for the confidence intervals.

Value

• An object of class kappa with three components:
• Kappa
  Kappa statistic, along with Approximate Standard Error (ASE) and 95% confidence bounds.
• Kappa.Weighted
  idem for the weighted kappa.
• Weights
  weight matrix used.

Details

Cohen's kappa is the diagonal sum of the (possibly weighted) relative frequencies, corrected for expected values and standardized by its maximum value. The equal-spacing weights are defined by $1 - abs(i - j) / (r - 1)$, r number of colums/rows, and the Fleiss-Cohen weights by $1 - abs(i - j)^2 / (r - 1)^2$. The latter ones attach greater importance to near disagreements.

References

\itemCohen, Jacob (1960): A coefficient of agreement for nominal scales. Educational and Psychological Measurement. \itemEveritt, B.S. (1968): Moments of statistics kappa and weighted kappa. The British Journal of Mathematical and Statistical Psychology.

See Also

agreementplot

Examples

data(SexualFun)
Kappa(SexualFun)