CORRELATION ANALYSIS: THE BOOTSTRAP APPROACH

For a general class of problems, the bootstrap method of resampling is one of the possible methods of constructing tests of significance. The sampling distribution of a test statistic for an experiment compiled by the bootstrap approach requires no reference to the population distribution and therefore no requirement that it should conform to a mathematically definable frequency distribution. Algorithms for the bootstrap distribution of correlation coefficients are presented and implemented. As an illustrative example, a critical value for Pearson’s product moment correlation coefficients and Spearman’s rank correlation coefficients are produced for a given set of data