Issues related to the construction of instruments for course and instructor evaluation by students are reviewed, and the construction process is examined in terms of domain specification, scaling, item generation, and statistical analysis. A set of guidelines applicable to this process is derived from research and current practices. (Author/JMD)

Examined effects of item characteristics on cloze text performance. Focus was on the relationship between cloze item characteristics and scoring methods. Item characteristics examined were: content/function words; parts of speech; word frequency; number of occurrences of a word in text; alternative answers; syntactic variation; amount of content;…

Reports on a study that tested methods to control or reduce spurious response in order to improve survey technique. Examines two methods of controlling spurious response: alerting respondents that some questionnaire items are false, and using the "don't know" response option. Finds that only the "don't know" option actually…

The nature of differences in performance on the third revision of the Stanford-Binet Intelligence Scale (Form L-M) was assessed for two independent samples of Black (n=86) and White (n=86) preschoolers matched for parental education, age, and sex. Large differences, favoring White children, were found. Test bias is discussed. (TJH)

Multidimensional scaling (MDS) and components solutions were compared using two balanced self-report item sets. Results suggest that when self-report items are balanced and negatively worded items are not reverse scored, MDS and components solutions can differ largely, due to a component associated with measures of acquiescence. (SLD)

Causes of differential item difficulty for blind students taking the braille edition of the Scholastic Aptitude Test's mathematical section were studied. Data for 261 blind students were compared with data for 8,015 non-handicapped students. Results show an association between selected item categories and differential item functioning. (TJH)

The modified caution index's use in identifying judges whose patterns of item judgment appear aberrant when compared with the pattern produced by the entire group (N=158) of judges was studied. Effects on test standards and passing rates of removing test standards of these judges were also assessed. (TJH)

Swedish Scholastic Aptitude Test results from Spring 1984 through Fall 1986 were analyzed for 14,768 males and 16,574 females with regard to gender differences at the item level. For four of the six subtests, differences were related to item subject content; this was untrue for the two quantitative subtests. (SLD)

Nonparametric item response models were used to investigate psychometric properties of the Beck Depression Inventory in 648 depressed outpatients and 1,182 nonpatient college students. Estimated values for some options did not align with a priori weights in the college sample, and some option characteristic curves were problematic in both samples.…

Defines computer assisted language testing (CALT), discusses the various processes involved, outlines the advantages and disadvantages, and examines psychometric aspects of computer testing. A table of factors distinguishes between test content and the mechanics of test taking. These factors constitute a table for developing a CALT checklist. (24…

Rasch calibration procedures were applied to item-response data for the 1,262 infants and toddlers comprising the standardization sample for the Mental Scale of the Bayley Scales of Infant Development. Analyses tend to confirm the psychometric integrity of the instrument. (Author)

Conflicting evidence regarding listening comprehension as a separate trait in language proficiency is reviewed. Two studies conducted in Japan are presented that use the multitrait-multimethod methodology to examine the construct validity of the listening trait. (74 references) (Author/LB)

Studies with 220 college students found that multiple-choice test items with 3 items are more difficult than those with 4 items, and items with the none-of-these option are more difficult than those without this option. Neither format manipulation affected item discrimination. Implications for test construction are discussed. (SLD)

A motivation questionnaire was used to investigate dominant factors and features of motivation among 90 Singapore elementary school teachers. Through item analysis, the following factors emerged as underlying motivations: esteem, self-actualization, security, autonomy, belongingness, achievement orientation, professional challenge, and task…

Examines whether a defect of the C-test can be avoided by constructing a C-test with five texts and 126 items. The test was tried with 146 Iranian English majors. On the basis of item analysis, a tailored C-test with 100 items was developed and tried with 60 other subjects. Results show no gains were made with the classical item analysis.…