A method and system for determining whether the number of samples taken from a population of units where the distribution of X and Y variables are unknown by evaluating ratio mean measurements on a computer to determine a confidence interval. The method comprising: inputting samples from the total population with each unit sample having at least two variables X and Y; redefining the multivariate data that comprises the two variables X and Y; estimating the mean; computing the standard error; using a bootstrapping method, generating boot strap samples, computing a Z distribution based upon the bootstrap samples; and computing the confidence interval. The system comprises a processor for performing the steps of the method.