CRAN release

Updated code in testCompareR_internal.R to correct errors in statistical methodology.

Specifically, lines

945 pvalue1 <- 2 * (1 - pnorm(w1, 0, 1))
946 pvalue2 <- 2 * (1 - pnorm(w2, 0, 1))
953 pvalue3a <- 2 * (1 - pnorm(Mcc1, 0, 1, ...))
954 pvalue4a <- 2 * (1 - pnorm(Mcc2, 0, 1, ...))
959 pvalue3b <- 2 * (1 - pnorm(M1, 0, 1, ...))
960 pvalue4b <- 2 * (1 - pnorm(M2, 0, 1, ...))

have been corrected to

945 pvalue1 <- 1 - pchisq(w1, 1, ...)
946 pvalue2 <- 1 - pchisq(w2, 1, ...)
953 pvalue3a <- 1 - pchisq(Mcc1, 1, ...)
954 pvalue4a <- 1 - pchisq(Mcc2, 1, ...)
959 pvalue3b <- 1 - pchisq(M1, 1, ...)
960 pvalue4b <- 1 - pchisq(M2, 1, ...)

because the test statistics approximate the chi-squared, rather than the normal distribution (see https://doi.org/10.1186/s12874-020-00988-y).