Violations of item unidimensionality on language tests produced distorted estimates of person ability, and violations of person unidimensionality produced distorted estimates of item difficulty. The Bejar Method was sensitive to such distortions. (Author)

The comparability of scores on test forms that are constructed through examinee item choice is examined in an item response theory framework. The approach is illustrated with data from the College Board's Advanced Placement Test in Chemistry taken by over 18,000 examinees. (SLD)

Discusses subject searching and describes a study that evaluated the match between the named terms and the expected subject headings by examining the subject search terms generated by students. Highlights include nonsemantic properties in evaluating search terms, including concreteness, complexity, and syndeticity; and metacognitive naming skills…

Sixty elementary students with learning disabilities were interviewed about parallel domains of uncontested and contested knowledge on the topic of space, including questions of morality and questions of empirical law. Students clearly distinguished between uncontested and contested knowledge, suggesting that they are capable of working with…

Twenty-two subjects with autism were tested on three perspective-taking tasks. Subjects were successful on two tasks that required them to make relatively direct connections between eating and hunger or visual access and knowledge. They failed a task which used stimuli of a more transient nature and less predictable reactions of the protagonists.…

Reliabilities and information functions for percentile ranks and number-right scores were compared using item response theory, modeling standardized achievement tests. Results demonstrate that situations exist in which the percentage of items known by examinees can be accurately estimated, but the percentage of persons falling below a given score…

A study was conducted with 782 elementary school students to determine whether the Homan-Hewitt Readability Formula could identify the readability of a single-sentence test item. Results indicate that a relationship exists between students' reading grade levels and responses to test items written at higher readability levels. (SLD)

Reports on a project which developed a measure of levels of difficulty of reference questions handled by the California Interlibrary Reference Referral Network, and identified indicative measures that would reliably stand for the concept of difficulty. Correlations of predictive difficulty and actual difficulty are discussed. (nine references)…

The readability, interest level, and content of 18 current college textbooks on mainstreaming were assessed. Results revealed that readability and interest level were comparable among the texts. More variation was found in the contents of the texts. A table lists 16 topics and the degree to which each is covered in the textbooks. (DB)

Analysis of the Bracken Basic Concept Scale (BBCS) indicated that all subtests except two (Quantity and Time) showed statistically significant correlations between item difficulty and frequency of concept words. Administration of the BBCS to 36 first graders revealed that most concept words showed expected item discrimination indices. (Author/JDD)

The hypothesis that a task that increases in complexity (increasing its correlation with a central measure of intelligence) does so by increasing its dimensionality by tapping individual differences or another variable was supported by findings from 46 adults aged 20-70 years performing a mental counting task. (SLD)

An associative model of serial learning is described based on the assumption that the effective stimulus for a serial-list item is generated by adaptation-level coding of the item's ordinal position. How the model can generate predictions of aspects of serial-learning data is illustrated. (SLD)

One-parameter and three-parameter item response theory (IRT) model estimates were compared with estimates obtained from two modified one-parameter models that incorporated a constant nonzero guessing parameter. Using small-sample simulation data (50, 100, and 200 simulated examinees), modified 1-parameter models were most effective in estimating…

Four studies determined when first, second, and third graders recognize that they make inferences to understand text, and the effect of this recognition on their ability to revise text and monitor its informativeness. Younger children tended to attribute inferred information to the text, while older children clearly distinguished inferred and…

Suggests a "complexity effect" for overall message-production speed (in which undergraduate students exhibited superior performance relative to older adults) and for initial message-production-skills performance and rate of skill acquisition, with these differences not pronounced under complex-task conditions. Finds that older adults' learning…