Analyzes simulated performance ratings on a six-point scale by two independent raters to account for nonsystematic error in performance ratings. Results suggest that rater agreement or covariance is not always a dependable estimate of score reliability and that the practice of seeking additional raters for adjudication of discrepant ratings is not…

A study was done to identify the cultural or linguistic bias present in an English as a second language placement examination at the university level. Test items exhibiting bias in their respective skill domains were identified by comparing 34 native Spanish speakers and 77 native Chinese speakers. (SED)

Presents a study of how well a set of language proficiency self-ratings fit the predictions of a probabilistic measurement model known as the Rasch Model. Applies the principles of the model to scalar rather than binary item response data. Concludes that scalar analysis of this kind is feasible with self-rating data. (Author/SED)

An English-as-a-second-language reading comprehension test is analyzed, first using classical measurement procedures and then using the Rasch Model latent trait procedures. Results indicate that the Rasch Model analysis procedure produced numerous advantages for the test developer. Summarizes and describes 13 of these advantages. (SED)

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)

Examines the effectiveness of an automated language proficiency test assembly system at an air force base English Language Center. The study focuses on the equivalence of mean score difficulty, total score variance, and intercorrelation covariance across test norms and finds a high level of test-form equivalence and internal consistency. (nine…

This simulation study considered the effects on statistical measures of test dimensionality that result from systematic sampling variation in both a single- and a double-trait assessment model. Results suggest that there are distinct psychological and psychometric states of test dimensionality, and that psychometric unidimensionality may be…