Disentangled sources of differential item functioning (DIF) in a multilanguage assessment for which multiple factors were expected to be causing DIF. Data for the Third International Mathematics and Science study for four countries and two languages (3,000 to 11,000 cases in each comparison group) reveal amounts and sources of DIF. (SLD)

The degree to which an undergraduate mathematics curriculum matched the item difficulty levels of representative mathematics problems based on that sequence was studied with 62 items and 423 undergraduates and 937 high school seniors. Results suggest congruence between curriculum sequence and item difficulty levels. (SLD)

This article presents guidelines for writing better test items to evaluate student learning better. Recommendations are provided for writing true-false items (test only one idea in each item), multiple-choice items (include the bulk of information in the stem), and matching items (only use homogeneous lists). Examples are provided. (Contains 7…

Studied whether differential item functioning (DIF) in translated verbal items could be reduced or eliminated by revising these items. Results for six sections of an Israeli college admission test translated from Hebrew to Russian show that revisions can reduce DIF considerably. Discusses costs of the revision process. (SLD)

Knowledge of Memory Aging Questionnaire with a "don't know" response added was administered to 202 undergraduates (half using the new format), 44 undergraduates (with instruction on memory and aging), and 49 older and 41 younger adults. The new format resulted in more correct answers. Instruction resulted in fewer "don't know" responses. Younger…

Developed an item response model for characterizing test-compromise that enables the estimation of item preview and score-gain distributions. In the approach, models parameters and posterior distributions are estimated by Markov Chain Monte Carlo procedures. Simulation study results suggest that when at least some test items are known to be…

Investigates whether there is any relationship between an examinee's ability and choice of wrong alternatives when the correct answer is not chosen. Analyzes some constructed-responses data. (Author/KHR)

A study examined the interactive effects of accounting students' language background (English as a first or second language) and test item format on test scores. Items on the same material were in four formats: multiple choice quantitative, multiple choice theoretical, open-ended quantitative, and open-ended essay questions. (Author/MSE)

Indices that detail aspects of student test responses include overall aberrancy; tendencies to miss relatively easy items; tendencies to correctly answer more difficult items; and a combination that indicates how the latter tendencies balance each other. Mathematics test results for 368 college students illustrate the indices. (SLD)

Statistical properties of person fit indices are reviewed as indicators of the extent to which a person's score pattern is in agreement with a measurement model. Distribution of a fit index and ability-free fit evaluation are discussed. The null distribution was simulated for a test of 20 items. (SLD)

Use of the Mantel-Haenszel procedure as a test for differential item functioning under the Rasch model of item-response theory is examined. Results of the procedure cannot be generalized to the class of items for which item-response functions are monotonic and local independence holds. (TJH)

The consequences of using item response theory (IRT) item bias detecting procedures with multidimensional IRT item data are examined. Limitations in procedures for detecting item bias are discussed. (SLD)

The effect of using statistical and cognitive item difficulty to determine item order on multiple-choice tests was examined, using 120 undergraduate students. Students performed better when items were ordered by increasing cognitive difficulty rather than decreasing difficulty. The statistical ordering of difficulty had little effect on…

Reports on a study examining the pattern of question selection by English as a Second Language (ESL) students on a writing competence test and the relationship to performance. Results suggest that ESL students heavily favored first and second questions in each set of four questions, and shorter questions in sets. (CB)

A judgmental method for determining item bias was applied to test data from 2,000 Native American and 2,000 Anglo-American students for a statewide proficiency test. Results indicated some shortcomings of the judgmental method but supported the use of cross-validation in empirically identifying potential bias. (SLD)