The purpose of this book is to make it possible for measurement specialists to solve practical testing problems through the use of item response theory (IRT). The topics, organization, and presentation are those used in a 4-week seminar held each summer for the past several years. The material is organized to facilitate understanding; all related…

If a loss function is available specifying the social cost of an error of measurement in the score on a unidimensional test, an asymptotic method, based on item response theory, is developed for optimal test design for a specified target population of examinees. Since in the real world such loss functions are not available, it is more useful to…

While standard setting procedures are typically discussed in terms of deriving a reasonable cutting score for a given form of a test, the situation may be structured such that the standard has been mandated without regard to the test form itself. This situation may result either through legislative or policy actions and may be a fait accompli by…

A common problem in practical educational research is that of perfect scores which result when latent trait models are used. A simple procedure for managing the perfect and zero response problem encountered in converting test scores into measures is presented. It allows the test user to chose among two or three reasonable finite representations of…

A quantitative method, meta-analysis, is used to integrate research findings from a representative group of 38 coaching studies. Because different studies reported results on different scales, effect size was used to transform all results to a common metric. This meta-analysis showed that there are two distinct literatures on the effectiveness of…

This study used meta-analysis research techniques to determine if there are differences in General Aptitude Test Battery (GATB) validities and test scores between males and females. The sample consisted of 26,111 subjects from 122 Specific Aptitude Test Battery (SATB) validation or revalidation studies analyzed since 1972. Four approaches were…

This paper compares five growth measures using the High School and Beyond database. The measures are: (1) simple gain (posttest minus pretest); (2) difference between group means; (3) percentage of students who scored higher on the posttest than the pretest; (4) percentage of items missed on the pretest which are subsequently answered correctly on…

Evaluation of mean score differences between the Stanford-Binet Intelligence Scale (Fourth Edition) and the Wechsler Intelligence Scale for Children (Revised) (WISC-R) for 126 children with academic difficulties found the Stanford-Binet composite score was significantly higher than the WISC-R score at the lower end of the ability continuum but…

Two factors which act to depress cognitive test performance are referred to as anxiety and boredom. The responses to a 20-item adjectival checklist administered to high school seniors after completing a cognitive test battery were subjected to iterative principal axes factor analysis. The relationships between anxiety or boredom and test…

The importance of using the standard error of measurement (SEm) in determining reliability in test scores is emphasized. The SEm is compared to the hypothetical true score for standardized tests, and procedures for calculation of the SEm are explained. (JDD)

This study examines the nature of the highly variable black-white difference across diverse tests and indicates the major systematic source of this between-population variation, namely, Spearman's g. Eleven large-scale studies (discussed in the appendix), each one comprising anywhere from 6 to 13 diverse tests, show a significant and substantial…

The stability of intelligence test scores over time was examined for the Wechsler Intelligence Scale for Children-Revised (WISC-R). Subjects included 64 children aged 6-16, who had been administered the WISC-R during prior evaluations. These students had been referred because of academic difficulties. One-third of the sample had taken the test…

A solution to reduce the likelihood of one examinee copying another's answers on large scale tests that require all examinees to answer the same set of questions is to use multiple test forms that differ in terms of item ordering. This study was conducted to determine whether varying the sequence in which blocks of items were presented to…

This is a descriptive study of the activities, statistics, and decisions made regarding the re-testing of six inner-city elementary schools, based on results of a norm-referenced achievement test given in the Cincinnati Public School district in April, 1985. Three statistical operations were performed on test data from the Design for School…

Information on statistical detection of two different kinds of unethical activity associated with standardized testing is presented. The first is advance presentation of specific test items or excessive coaching using items very similar to those on a test. Cluster analysis is recommended for detection of aberrant vectors of percentages correct on…