Appendix B — Technical Note to the Professional Reader

The material presented in this book fits together with the technical literature as follows: Though I (JLS) had proceeded from first principles rather than from the literature, I have from the start cited work by Chung and Fraser (1958) and Meyer Dwass (1957) They suggested taking samples of permutations in a two-sample test as a way of extending the applicability of Fisher’s randomization test (1935; 1960, chap. III, section 21). Resampling with replacement from a single sample to determine sample statistic variability was suggested by Simon (1969). Independent work by Efron (1979) explored the properties of this technique (Efron termed it the “bootstrap”) and lent it theoretical support. The notion of using these techniques routinely and in preference to conventional techniques based on Gaussian assumptions was suggested by Simon (1969) and by Simon, Atkinson, and Shevokas (1976).