Description and analogy in the practice of statistics

Abstract

Formal and practical aspects of the role of statistical description in exploratory investigation are discussed in two well-known situations: the Wilcoxon two-sample test and 2 × 2 tables.