Minimum Credence Test

R script for Minimum Credence statistical test for comparing mixed samples. The test evaluates maximum possible population standard deviation from a reference sample and then applies this estimate to mixed samples for comparisons. In case the reference sample is less than 8, chi-squared approximation is used to find confidence interval for population standard deviation. Otherwise, the test bootstraps corrected median absolute deviation of the reference sample to obtain bias-corrected accelerated confidence interval for population standard deviation. This script is adapted for RFA soil analysis, but can be used elswhere if method=0. If method=1, the test extracts GOST metrological data (between-lab analytical error) from the excel file, and must be corrected for every specific analysis. The test implies homogeniety of variances among groups, and the studied parameter must average additively (i. e. arithmetic mean of individual samples is equal to mixed sample) in order to use mixed samples.For any questions/remarks please contact me at: [e-mail]