The Scholastic Aptitude Test (SAT) is shown to be a better predictor of first semester college grade point average than the American College Test (ACT) at a highly selective midwestern university. (JKS)

Grades in a college freshman course in finite mathematics were predicted from IQ, high school math grades, college board scores, and high school class ranks. The squared multiple correlation for the full model was .611. (JKS)

The concurrent criterion validity of a modified pass/fail grading system in a medical school was assessed. The investigators examined possible differences between honor society and nonhonor society students on three medical school entry variables and on medical board scores. (Author/JKS)

The predictive validity of the Admissions Test for Graduate Study in Business (ATGSB) and alternative predictors for English-speaking and non-English speaking students in a Master's Degree Business Administration Program were investigated. The ATGSB and other variables were found to correlate with success in the program. (Author/JKS)

The predictive ability of a mathematics test organized into a branching test for computer-interactive administration was investigated. Twenty-five blocks of five items were used in the branching. Each testee took 25 items, with each subsequent block being determined by prior performance. Results supported the branching technique. (JKS)

The reliability and validity of the Judging Criteria Instrument as a measure of creativity in physics is examined. The instrument was found to be both reliable and valid. (JKS)

The concept of interval measurement in the fields of psychology and education is explored. Measurement and examples of its relationship to subsequent statistical analysis are discussed. (JKS)

G analysis is a type of discriminant analysis. A generalization of the G analysis to cases where variables are on an interval or ordinal scale is described. The placement system used in the description is distance-based, but other systems may also be applied. (Author/JKS)

Statistical tests for determining whether one pole of an attitude scale is characteristic of a population are examined. A less restrictive procedure and the situations in which it might be used are discussed. (Author/JKS)

Two techniques for writing achievement test items to accompany instructional materials are contrasted: writing items from statements of instructional objectives, and writing items from semi-automated rules for transforming instructional statements. Both systems resulted in about the same number of faulty items. (Author/JKS)

Most discriminant function analyses provide a table of hits and misses with respect to classifications. Cohen's kappa coefficient is shown to be applicable to these tables and useful for setting confidence limits for the accuracy of the discriminant prediction. (Author/JKS)

A multiple true-false item is one where a testee has to identify statements as true or false within a cluster (of two or more) of such statements. Clusters are then scored as items. This study showed such a procedure to yield less reliable results than traditional true-false items. (JKS)

A method for estimating variance-covariance matrices with incomplete data for use in discriminant function analysis is presented. An example of its use in college algebra course placement is also presented. (JKS)

Seven formulas for scoring test items with two options (true-false or multiple choice with only two choices) were investigated. Several conditions, such as varying directions for guessing and whether testees had prior knowledge of the proportions of false items on the test were also investigated. (Author/JKS)

A modification of Hoyt's analysis of variance model for test analysis was proposed by Lu. A difficulty that may be encountered in using Lu's modification is examined, and a solution is proposed. (JKS)