Multivariate interpretation of laboratory tests used in monitoring patients

This study demonstrates an approach to the problem of minimizing false-negative and false-positive laboratory findings. In this approach, we consider the fact that results of laboratory tests are correlated, utilize within-person test results to interpret current results, and minimize the impact of multivariate conservatism by examining test results in small groups. The procedure requires panels of tests to be divided into related subpanels, testing each subpanel independently, and using the Bonferroni inequality to determine whether any of the observed values for a given subpanel is "out-of-range." The procedure is demonstrated, and its limitations are observed and discussed.

Error Probabilities in Educational and Psychological Research

The well-known problem of cumulating error probabilities is reconsidered from a general epistemological perspective, namely, the concepts of severity (Popper) and of fairness of tests. Applying these concepts to hypothesis-testing research leads to a reevaluation of the relative importance of the probabilities of Type 1 and Type 2 errors connected with those statistical hypotheses that have been derived from the substantive ones. It is shown that not only Type 1 but also Type 2 errors can cumulate. This cumulation is discussed for various basic types of empirical situations in which substantive hypotheses are examined by means of statistical ones. A new adjustment strategy based on the Dunn-Bonferroni inequality for planned tests is proposed and applied to some empirical examples.