Confidence Interval Approach for the Number of Subjects Required for a Study of Interobserver Agreement with Four Outcome Categories

Share:

Description

This function provides detailed sample size estimation information to determine the number of subjects required using the confidence interval perspective to sample size estimation for κ. This version assumes that the outcome has four categories.

Usage

1
CI4Cats(kappa0, kappaL, kappaU=NA, props, raters=2, alpha=0.05)

Arguments

kappa0

The preliminary (anticipated) value of κ.

kappaL

The desired expected lower bound for a two-sided 100(1 - α) % confidence interval for κ. Alternatively, if kappaU is set to NA, the procedure produces the number of required subjects for a one-sided confidence interval.

kappaU

The desired expected upper confidence limit for κ.

props

The anticipated prevalence of the desired trait. Note that the elements of the four element vector must be non-negative and sum to one.

raters

The number of raters that are available. This function allows between 2 and 6 raters.

alpha

The desired type I error rate.

Details

This function provides detailed sample size estimation computation for studies of interobserver agreement with four outcomes. This function employs the confidence interval perspective, determining the correct sample size that provides the specified expected confidence limits. Sample size estimation is based on the precision of the estimate, instead of a simple hypothesis testing perspective. Note that a warning message is provided if any of the expected cell counts are less than 5.

Value

N

The calculated sample size.

kappa0

The specified anticipated value of κ.

kappaL

The specified expected lower limit.

kappaU

The specified expected upper limit.

props

The anticipated proportions of individuals with the outcomes of interest.

raters

The number of raters.

alpha

The desired type I error rate.

ChiCrit

The critical value that is required for sample size estimation. It is typically not required and is not displayed in the summary output.

Author(s)

Michael Rotondi, mrotondi@uwo.ca

References

Donner A, Rotondi MA. (2010). Sample Size Requirements for Interval Estimation of the Kappa Statistic for Interobserver Agreement Studies with a Binary Outcome and Multiple Raters. International Journal of Biostatistics 6:31.

Altaye M, Donner A, Klar N. (2001). Procedures for Assessing Interobserver Agreement among Multiple Raters. Biometrics 57:584-588.

Donner A. (1999). Sample Size Requirements for Interval Estimation of the Intraclass Kappa Statistic. Communication in Statistics 28:415-429.

Bartfay E, Donner A. (2001). Statistical Inferences for Interobserver Agreement Studies with Nominal Outcome Data. The Statistician 50:135-146.

Donner A, Eliasziw M. (1987) Sample size requirements for reliability studies. Statistics in Medicine 6:441-448.

See Also

Power4Cats

Examples

1
2
3
4
5
6
## Not run: Suppose an investigator would like to determine the required sample size to test 
kappa0=0.4 with precision of 0.1 on each side, in a study of interobserver agreement. 
Further suppose that the prevalence of the traits are 0.30, 0.2, 0.2, 0.3.
## End(Not run)

CI4Cats(kappa0=0.4, kappaL=0.3, kappaU=0.5, props=c(0.30, 0.2, 0.2, 0.3), alpha=0.05);