Exact conditional tests of independence in cross-classification tables are formulated based on chi square and other statistics with stronger operational interpretations, such as some nominal and ordinal measures of association. Guidelines for table dimensions and sample sizes for which the tests are economically implemented on a computer are…

Studied thirty variables that may affect crime using cross-validation procedures. When Sample 1 predictions were subjected to validation in Sample 2, the number of predictors was reduced from seven to four for delinquency, from five to four for criminal offenses, and from five to three for juvenile status offenses. (BH)

Discusses the suitability of randomization tests when working with a nonrandom sample. Offers a rationale for the noninferential use of statistical tests. (JOW)

This paper derives optimal rules for sequential mastery tests. In a sequential mastery test, the decision is to classify a subject as a master or a nonmaster or to continue sampling and administering another random item. The framework of minimax sequential decision theory (minimum information approach) is used; that is, optimal rules are obtained…

Efron's Monte Carlo bootstrap algorithm is shown to cause degeneracies in Pearson's r for sufficiently small samples. Two ways of preventing this problem when programing the bootstrap of r are considered. (Author)

It is demonstrated that tests of homogeneity of independent correlation coefficients based on the simple forms of the normal and t approximations to the distribution of the correlation coefficients are comparable in terms of robustness, size and power. (Author)

This study explores a strategy for improving the stability of variance component estimates when only small samples are available, using a series of small, less complex generalizability (G) study designs as a surrogate for a single large design. (Author/BW)

Examples are given to show that content sampling may be considered a form of operationalism in defining constructs. Five conditions are proposed as necessary if one is to accept the use of a measuring instrument as a valid operational definition on the basis of content sampling alone. (Author/RC)

Identifies conditions under which equivalence is likely attained with simple random sampling and randomization in psychotherapy efficacy studies of kind examined in recent meta-analyses. Views consequences of nonequivalence as manifestations of Simpson's paradox. Expects misinterpretations of estimates of relative efficacy of treatments. Compares…

This paper describes the need for statistically based survey design. It briefly summarizes the U.S. Environmental Protection Agency's Environmental Monitoring and Assessment Program (EMAP) and demonstrates how its design is tailored for the selection of a probability sample of lakes on which to make measurements of lake conditions. (LZ)

Reports a reanalysis of data collected in a person x task x occasion rater or method G-study design (M. Ruiz-Primo and others, 1993), and brings this reanalysis to bear on the interpretation of task-sampling variability and the convergence of different performance-assessment methods. (SLD)

Presents an approach to graphical differential item functioning (DIF) based on a sampling-theory approach to expected response functions. Applied the approach to a set of pretest items and compared results to traditional Mantel Haenszel DIF statistics. Discusses implications of the method as a complement to the approach of P. Pashley (1992). (SLD)

Describes a study of commercial Web sites that analyzed the correlation between domain names and regional allocation of IP addresses. Explains Internet addresses and names, discusses results which showed that domain names don't always correspond to the location of the host, and offers suggestions for international sampling. (Contains 25…