diffpci {diffdepprop} | R Documentation |
This function gives 12 different two-sided confidence intervals. Data are assumed to be of a fourfold table, which contains the numbers of concordance and the numbers of discordance of two dependent methods. The following intervals are listed: Wald, Wald with continuity correction, Agresti, Tango, Exact (Clopper Pearson and mid-p), Profile Likelihood, Wilson (without and with continuity corrections) and nonparametric approaches using rank methods (with normal and t-approximation).
diffpci(a, b, c, d, n, alpha)
a |
first number of concordant paires as described above |
b |
first number of discordant paires as described above |
c |
second number of discordant paires as described above |
d |
second number of concordant paires as described above |
n |
number of observed objects |
alpha |
type I error; between zero and one |
Details are given for each function separately.
A matrix containing the method, the difference estimator and the corresponding confidence limits.
Daniela Wenzel, Antonia Zapf
Newcombe, R.G. (1998). Improved confidence intervals for the difference between binomial proportions based on paired data. Statistics in Medicine 17. 2635-2650.
Clopper, C. and Pearson, E.S. (1934). The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika 26, 404-413.
Vollset, S.E. (1993). Confidence intervals for a binomial proportion. Statistics in Medicine 12. 809-824.
Lange, K. and Brunner, E. (2012). Sensitivity, Specificity and ROC-curves in multiple reader diagnostic trials-A unified, nonparametric approach. Statistical Methodology 9, 490-500.
Fleiss, Joseph L. et al. (2003). Statistical Methods for Rates and Proportions. Wiley.
# a=59, b=23, c=3, d=37, n=122, type I error is 0.05
diffpci(59,23,3,37,122,0.05)