Techniques for classifying individuals into mastery or non-mastery status based on criterion-referenced test performance are investigated. Extensions of mastery classification techniques developed by Macready and Dayton are presented. The techniques are based on latent class and quasi-independence models. (Author/JKS)

A defense of the use of standards with criterion-referenced testing is made in response to Glass's article (TM 504 031). (JKS)

The use of setting standards for criterion-referenced tests is defended in a response to two papers by Gene Glass and Nancy Burton. (JKS)

The problem of determining an optimal passing score for a mastery test is discussed, when the purpose of the test is to predict success on an external criterion. For the case of constant losses for the two possible error types, a method for determining passing scores is derived. (Author/JKS)

The problem of establishing appropriate passing scores is one of evaluation rather than estimation and not amenable to exact solution. It must therefore be approached by (1) identifying criteria for judging the acceptability of the passing score, (2) collecting the data appropriate to assessing each relevant criterion, and (3) judging how well the…

The utility of setting standards for educational decisions, even though those standards may be somewhat arbitrary, is defended in this response to Glass's article (TM 504 031). (JKS)

The use of cut-off scores with criterion referenced tests is defended in this response to two papers by Gene Glass and Nancy Burton. Suggestions for setting cut-off scores are made. (JKS)

A procedure is described for the determination of individual mastery scores when instructional units are sequenced in a linear hierarchy. It can be used when data for the entire sequence have been obtained. The process entails working backwards, from the final test to the first. (Author/JKS)

Addressing Glass' argument (EJ 198 842) that a lack of interrelater reliability is an inherent deficiency in the Nedelsky technique, poor rater training and the need for a group decision procedure are presented as standard setting problems. (CM)

Describes a procedure, called a full Bayesian procedure, for making decisions in connection with criterion-referenced measurements. The procedure uses continuous utility functions instead of a dichotomized utility structure and combines the posterior distribution for a certain person with utility functions for "advance" and "retain" decisions…

The discussion on criterion-referenced (CRT) and norm-referenced achievement tests (NRT) is divided into two parts: definitions and use. The authors contrast CRTs, tests which compare an individual's performance to some specified behavioral criterion of proficiency, and NRTs, tests which compare an individual's score to scores of others. They also…

Three methods for setting performance standards for criterion referenced tests--theories, expert judgments, and practical necessity--are reviewed. The author concludes that the use of performance standards is not a promising vehicle for making societal decisions. (Author/JKS)

A set of guidelines for evaluating criterion-referenced tests is presented. Additionally, 11 sets of extant criterion-referenced tests are evaluated using these guidelines. (JKS)

The Army Skill Qualification Tests (SQT's) were designed to determine levels of competence in performance of the tasks crucial to an enlisted soldier's occupational specialty. SQT's are performance-based, criterion-referenced measures which offer two advantages over traditional proficiency and achievement testing programs: test content can be made…

After reviewing the history and evolution of criterion referenced tests and citing the writings of other researchers, the author summarizes the current state of criterion referenced tests. They are used to measure an individual's status with respect to a well-defined behavior domain; the behavior domain can be well-ordered, structured, or…