Six procedures for combining sets of item response theory (IRT) item parameter estimates from different samples were evaluated using real and simulated response data. Results support use of covariance matrix-weighted averaging and a procedure using sample-size-weighted averaging of estimated item characteristic curves at the center of the ability…

Presents a study conducted to provide a comparative analysis of five item analysis indices using both IRT and non-IRT indices to describe the characteristics of flagged items and to investigate the appropriateness of logistic regression as an item analysis technique for further studies. The performance of five item analysis indices was examined.…

Three related Canadian studies assessed the content validity of 59 clinical problems designed as part of a test of medical decision-making skills. Focus was on the key features, i.e., the critical or essential steps in identification and management of the clinical problem. Results support content validity of the key features. (MSE)

The Rasch model was applied to selection of items for an Arabic version of the Otis-Lennon Mental Ability Test using a sample of 599 male and female Kuwaiti secondary school and university students. Results indicated that the test is suitable for the range of ability intended to be measured. (SLD)

The role of assessment in science education is explored, focusing on summative assessment in British public certificate examinations. Examples of test items are presented to illustrate difficulties in making valid and reliable assessments, and issues with implications for formative assessment are discussed. (SLD)

A total of 102 university English majors were administered 4 different language tests to form a General Language Proficiency measure against which the C-test was evaluated. Results confirmed its reliability and validity and also provided data on text difficulty/appropriateness, word structure, content, and different scoring methods. (13…

Describes how multiple-choice items can be designed and used as an effective diagnostic tool by avoiding their pitfalls and by taking advantage of their potential benefits. The following issues are discussed: correct' versus best answers; construction of diagnostic multiple-choice items; the problem of guessing; the use of justifications of…

Examined 6 widely used grade-8 standardized tests for content, required processes, and level to determine their alignment with the 5-8 NCTM "Curriculum and Evaluation Standards." Concluded that these tests inadequately covered the 5-8 standards. A follow-up study examined items from newly developed and foreign tests to demonstrate the…

The exact and closed-interval area measures for detecting differential item functioning are compared for actual data from 1,000 African-American and 1,000 white college students taking a vocabulary test with items intentionally constructed to favor 1 set of examinees. No real differences in detection of biased items were found. (SLD)

Using 38 beginning-level university students of French, this study confirmed that student listening comprehension of a foreign language video would be facilitated by the use of an advance organizer consisting of several short sentences, written in French, that summarized chronologically the events in the video. Sample test items and answers are…

Studied effects of students' use of calculators with 2 experimental forms of a university mathematics test taken by 765 and 725 college students, respectively. Calculator effects are not found for overall scores but are seen for some individual items. Analysis at the item level makes the actual impact apparent. (SLD)

For 42 Italian students studying English as a foreign language, there was no significant difference in subjects' success in recognizing words, regardless of whether words' final consonants were fully realized. Exercises that focus attention on sounds at the ends of words make learners practice a strategy that neither they nor first-language…

This paper illustrates how, in the item-response theory framework, collateral information about test items can augment or replace examinee responses when linking or equating new tests to established scales, using data from the Pre-Professional Skills Test for approximately 40,000 examinees. Collateral information can predict item operating…

The Affect Balance Scale (ABS) was administered to 380 undergraduate students. Factor analysis resulted in five negatively phrased items loading with the same signs as five positively phrased items. These results, contrary to previous results with older subjects, suggest that the ABS may not be appropriate for use across age groups. (Author/SLD)

Two extensions of the Mantel Haenszel procedure that may be useful in assessing differential item functioning (DIF) are explored. Simulation results showed that, for both inferential procedures, the studied item should be included in the matching variable, as in the dichotomous case. (SLD)