The principle that, for optimal retrieval, documents should be ranked in order of the probability of relevance is discussed. (Author/KP)

Describes a sampling demonstration, using random numbers generated on a microcomputer, that helps to make the central limit theorem concrete and understandable. (Author/JDH)

Problems involving the relative positions of four soccer teams and possible point allocations for the teams are analyzed. (DT)

Describes a mathematical model used in statistical prediction. Model is derived from known independent continuously variable quantities having the same distribution. (RS)

This article evaluates power and Type I error probabilities for a modified Fisher exact test for homogeneity of response probabilities in 2 x 2 tables. The continuity-corrected Fisher statistic is more powerful than the uncorrected Fisher exact test, while it is less often nonconservatively biased than the Pearson chi-square. (Author/RD)

The commonly accepted idea that the only rational measure of the worth of a gamble is its expected value or some subjective counterpart such as expected utility is examined. Also discussed are the changes called for in theories of rational choice when prescriptions of rational models violate common sense. (Author/GK)

Lambda and kappa coefficients of nominal scale association are developed for research hypotheses that involve predictions of modality, agreement, or some theoretically specified configuration. The proposed new coefficient is offered as an alternative to Goodman and Kruskal's lambda. (Author/CTM)

Describes a modified version of the tree diagram used to make probability more meaningful. (MKR)

One interpretation of the Pearson product-moment correlation ("r"), correlation as the probability of originating from common descent, important to the genetic measurement of inbreeding, is examined. The conditions under which "r" can be interpreted as the probability of "identity by descent" are specified, and the…

A k-dimensional multivariate normal distribution is made discrete by partitioning the k-dimensional Euclidean space with rectangular grids. The probability integrals over the partitioned cubes forms a k-dimensional contingency table with ordered categories. A loglinear model with main effects plus two-way interactions provides an approximation for…