confIntICC: Confidence intervals for intraclass correlation in interrater...

In felix-hof/biostatUZH: Misc Tools of the Department of Biostatistics, EBPI, University of Zurich

Confidence intervals for intraclass correlation in interrater reliability

Description

Compute, under suitable assumptions, confidence intervals for interrater reliability for continuous measurements presented to two or more raters.

Usage

confIntICC(dat, conf.level = 0.95, psi.re.0 = c(0, 1))

Arguments

dat	Data frame that contains the columns score, pat, rater.
conf.level	Confidence level for confidence interval.
psi.re.0	2-d vector specifying the interval [psi_0, psi_1] on p. 621 of Rousson et al. (2003).

Details

This function computes all the confidence intervals that are discussed in Roussen et al. (2003). In applications, the interval under the "trained rater" assumptions is often suitable.

Value

A list containing:

icc(2, 1)	ICC(2, 1): Intraclass correlation from a two-random effects model.
icc(3, 1)	ICC(3, 1): Intraclass correlation from a model with fixed rater effect.
psi_r/e	The value \psi_{r/e} computed from the actual data.
ci.trained.rater	Confidence interval under the trained rater assumption, see Rousson et al. (2003), Section 4.
ci.low.asy.corr	Lower bound of asymptotically exact confidence interval, see Rousson et al. (2003), Section 3.
ci.low.fix.rater	Lower bound of confidence interval under the fixed rater assumption, see Rousson et al. (2003), Section 5.

Note

The function computeICCrater computes ICCs relying on a mixed-model formulation, and is therefore able to handle unbalanced data. On the contrary, the confidence intervals in the function confIntICC are computed using sums of squares, and the data must therefore be balanced. See the example below.

Author(s)

Kaspar Rufibach
[e-mail]

References

Rousson, V., Gasser, T., and Seifert, B. (2002). Assessing intrarater, interrater and test-retest reliability of continuous measurements. Statist. Med., 21, 3431–3446.

Rousson, V., Gasser, T., and Seifert, B. (2003). Confidence intervals for intraclass correlation in inter-rater reliability. Scand. J. Statist., 30, 617–624.

See Also

computeICCrater, Aalpha, computeICCrater

Examples

## Generate dataset. Data must be balanced!
set.seed(1977)
n <- 40
r1 <- round(runif(n, 1, 20))
dat <- data.frame(
    "score" = c(r1, r1 + abs(round(rnorm(n, mean = 1, sd = 3))), 
        r1 + abs(round(rnorm(n, mean = 1, sd = 3)))), 
    "pat" = rep(c(1:n), 3),
    "rater" = rep(1:3, each = n)
)
confIntICC(dat, conf.level = 0.95, psi.re.0 = c(0, 1))

computeICCrater(dat)

felix-hof/biostatUZH documentation built on Sept. 27, 2024, 1:48 p.m.