| mrmcAnalysisOR {mitoticFigureCounts} | R Documentation |
OR's method to estimate the MRMC variance for a given theta. This function requires an nModalities by nReaders covariance matrix to estimate covOR = (cov1, cov2, cov3), as described by Hillis (2014, SIM, Section 2.2)
mrmcAnalysisOR(theta.hat, cov.hat, sgl = "corres")
theta.hat |
[2, nReaders]: Performance estimates for two modalities by all readers |
cov.hat |
[2 * nReaders, 2 * nReaders]: Covariance matrix between each reader x modality |
sgl |
ch: Flat indicating whether or not to do single modality analysis only using single modality data |
list
F : F statistic
df.H : degrees of freedom difference of concordance
p : P value
se.dif : difference of variance
theta.i [2] : reader-averaged HarrellsC form modality A & B
df.sgl [2] : degrees of freedom for modality A & B
se.i [2] : Variance form modality A & B
covOR [6] : Components of variance of the Obuchowski and Rockette method
cov1, cov2(pooled over modalities), cov3,
varC, cov2(modalityA only), cov2(modalityB only)