The average discrimination coefficient of the test paper is 0.4934. Among the choice items, there are 9.61 percent, 66.54 percent, 12.43 percent and 11.42 percent of them which respectively belong to best discrimination ones, better discrimination ones, usual discrimination ones and ugly discrimination ones.

When a question bank has been formed, further studies are needed for its perfection step by step. Reliability, validity, degree of difficulty, distinguishability are four indexes to evaluate the quality of question bank. Degree of difficulty and distinguishability were examined in this study. The results indicated that there was no close relation between degreeof difficulty and distinguishability and distinguishability was higher when degree of difficulty was moderate. It is suggested that the index of degree...

When a question bank has been formed, further studies are needed for its perfection step by step. Reliability, validity, degree of difficulty, distinguishability are four indexes to evaluate the quality of question bank. Degree of difficulty and distinguishability were examined in this study. The results indicated that there was no close relation between degreeof difficulty and distinguishability and distinguishability was higher when degree of difficulty was moderate. It is suggested that the index of degree of difficulty of sports medicine exam be 0.5-0.85. According to the index of distinguishability, exam questions can be divided into four types. The study made specific analysis on some exam questions. Dis tinguishability of diffe ent type of exam questions was discussed and suggestions to improve distinguishability were made.

Objective: To construct the General Administrative Aptitude Test (GAAT) for the government functionary. Methods: The content and structure of the GAAT were established by the method of work analysis, and use of the GAAT in actual entrance examinations. Reliability and validity of the GAAT were examined by test-retest, group comparison, correlation and factor analyses. Results: The overall difficulty of the test was 0.538, the discrimination of all items was over 0.20. There were middle or lower coefficients...

Objective: To construct the General Administrative Aptitude Test (GAAT) for the government functionary. Methods: The content and structure of the GAAT were established by the method of work analysis, and use of the GAAT in actual entrance examinations. Reliability and validity of the GAAT were examined by test-retest, group comparison, correlation and factor analyses. Results: The overall difficulty of the test was 0.538, the discrimination of all items was over 0.20. There were middle or lower coefficients of correlation between the subscale scores (0.2143～0.4909), but the coefficients of correlation between the subscales and the overall scores were higher(0.5858～0.8326). There were two main factors revealed by factors analysis of the subscales, and four main factors by factor analysis of subtests. The homogeneity reliability of GAAT was 0.932; correlation of retesting after one month was 0.836. There were some discrepancies with respect to age, sex, and education level, but the scores of GAAT on the discrepancies were not very high (1-6). There was obvious correlation between the GAAT and CAITG. Comparison of the two government functionaries' groups on their manifest work and actual abilities showed that the first group that had an entrance examination had higher working quantity and quality, mastery in knowledge and cognizing ability. Conclusion: The structure of the GAAT is reasonable, and its reliability and validity meet the need of psychometrics.

To investigate the quality of college-wide tests for non-English majors, two samples of such test papers are analyzed with reference to modern testing theory and associated criteria, one taken by 1185 grade-one students and the other by 976 grade-two students. A statistical review of the score characteristics from the two tests is followed by discussion of aspects like item structure, inter-correlation between sections, indexes of difficulty and item discrimination , with a view to bringing strengths and weaknesses...

To investigate the quality of college-wide tests for non-English majors, two samples of such test papers are analyzed with reference to modern testing theory and associated criteria, one taken by 1185 grade-one students and the other by 976 grade-two students. A statistical review of the score characteristics from the two tests is followed by discussion of aspects like item structure, inter-correlation between sections, indexes of difficulty and item discrimination , with a view to bringing strengths and weaknesses of such achievement tests to the attention of English teachers and test developers, so that subsequent measures can be adopted to improve the test quality.