The consequences in terms of misclassifications of students that would occur by replacing the constructed-response portion of the Advanced Placement (AP) examinations in economics with more multiple-choice items were studied. The 1991 AP examinations in micro- and macroeconomics were used. Computer simulation found that a small but statistically…

Uses item analysis techniques to compare two multiple-choice test formats for assessing second language students' competence in English structure in terms of mean scores, item difficulty, and item discrimination. Findings reveal that difficulty and discrimination aspects are item-specific rather than format-dependent and items with moderate…

Explains the use of instructional design with a mastery learning strategy (IDML) in teaching mathematics. The IDML model, a combination of other theories including Gagne's instructional events, has been used to improve mathematics instruction in Malaysian schools, and includes developing test items, the teachers' development of learning materials,…

The accuracy with which expert systems (ESs) score a new nonmultiple-choice free-response test item was investigated, using 734 high school students who were administered an advanced-placement computer science examination. ESs produced scores for 82 percent to 95 percent of the responses and displayed high agreement with a human reader on the…

Describes a possible exam or homework problem for a physical chemistry course. Investigates the temperature and concentration dependence of homogeneous chemical equilibria. Provides acceptable solutions. Offers a computer program able to provide partial credit on questions requiring four-seven sequenced items. Program available from author. (MVL)

Discusses the need for a science-technology-society (STS) approach in science classroom and describes the characteristics of exemplary STS programs. Lists five domains for STS goals and indicates assessment instruments for the domains. (YP)

Presented is a summary of the actions taken by American College Testing in response to calls for reform in standardized testing. The areas of reform discussed are test development, test content, skills levels, test criteria, and new testing initiatives, such as alternative-response-format questions and calculator use. (MDH)

The need for scale linking in the National Assessment of Educational Progress (NAEP) is discussed, and the specific procedures used to carry out the linking in the context of the major analyses of the 1990 NAEP mathematics assessment are described. Issues remaining to be addressed are outlined. (SLD)

A method for analyzing test-item responses is proposed to examine differential item functioning (DIF) in multiple-choice items within the latent class framework. Different models for detection of DIF are formulated, defining the subgroup as a latent variable. An efficient estimation method is described and illustrated. (SLD)

Gender differences in omitting items and guessing on multiple-choice tests were studied in Israel for 302 male and 302 female ninth graders and 150 male and 150 female university applicants. Females tended to omit more items and guess less often than did males. Implications for scoring are discussed. (SLD)

A meta-analysis was conducted of 107 classroom and laboratory studies concerning the effects of expecting a recall, recognition, essay, multiple-choice, or true-false test on students' subsequent achievement. Laboratory studies did not generalize well to classrooms. In classroom studies, subjects performed better when they knew which test type to…

Results from 40 high school students on the mathematics subtest of the Scholastic Aptitude Test (SAT) support the prediction that successful students would have more prior knowledge and would be better at defining problems, assembling strategies, and avoiding computational errors. Results are discussed in terms of a cognitive processing model.…

An Australian study examined the role of language background and gender in the assessment of native and non-native English-speaking university students enrolled in physics courses. Attention was given to differential performance and was analyzed in relation to test-item type and level of student communication skills. Implications for academic…

Whether there is a gender-by-item interaction that makes easy mathematics test items easier for females than for males and difficult items harder for females than for males was tested in two studies, involving 10,321 eighth graders and 5332 ninth graders and results from nine mathematics tests. The hypothesis was generally supported, as discussed.…

An oral testing method used in an intensive high school Spanish program is based on a model of communicative competence positing four forms of competence: strategic; sociolinguistic; discourse; and linguistic. All oral domains but linguistic are tested in the first and second years. In the first year, benchmark testing occurs. The rubric used has…