Review of Test Theory and Methods

摘要

The research literature on test theory and methods for the period 1975 through early 1980 is critically reviewed. Research on classical test theory has concentrated on relatively unimportant developments in reliability theory, with some new developments and applications of generalizability theory appearing during this period. The reliability of change or gain scores has received some attention from the classical test theory perspective, as have the applications of classical reliability concepts in experimental design and the analysis of experimental data. A minor amount of research with classical models was in the area of test-score equating. Classical item analysis procedures, however, received little attention. A fair amount of research during the period was devoted to different item types of test item response modes as replacements for the ubiquitous multiple-choice item. Several types of true-false items were proposed, and formula scoring was studied by a number of researchers in an attempt to reduce guessing effects. The perennial topic of response option weighting received attention, with efforts oriented toward demonstrating effects on validity and reliability. Response modes studied included answer-until-correct, confidence weighting, and free-response. A number of alternatives to classical test theory were studied in an attempt to solve some of the problems for which classical test theory has proven to be inadequate.