The research project entitled "General English Preficiency Texting: Phase II, 1991" is part of a project coordinated by the Academic Affairs Division and Chulalongkorn University Language Institute to develop standardized tests of general English proficiency and related testing systems. The objectives of the research were as follows: 1.to generate more test items than those in the Phase I research by using the same test table of specifications 2. to analyze and improve the constructed test items item by item as well as by form 3. to standardize the tests and enter them into the computerized test-item banking system. The population of the research numbering 2,334 were Chulalongkorn University students from first year to graduate level, and the general public with at least a Bachelor's degree who voluntarily enrolled for the tests. The subjects in the first administration of the four forms of the tests numbered 481 or 20.57% of the population. The subjects in the second administration taking the revised version of the tests numbered 87. The tests constructed, administered, analyzed, and revised are criterion-referenced tests of general English proficiency consisting of 480 objective-type and 12 subjective-type items. The classical model of test item analysis was used for item revision while the Rasch model was for item banking. The study led to (1) the development of 480 objective items testing the listening, reading and writing skills grouped in four parallel test forms with reliability indices ranging from acceptable to high (KR 20 = .854 - .910, Phi coefficients = .820 - .896), the efficiency of prediction E= 48.00 - 58.50; and 12 subjective items measuring the skills of writing with acceptable to high reliability indices (r[subscript xy = .554 -.840); (2) the 471 standardized objective test items; (3) 12 standardized subjective test items and (4) 414 objective test items which fir the Rasch model item analysis.
