A comparison of the quality between the single and the double multiple choice tests in measuring the different cognitive domain levels administered to different groups of learning achievement students
Abstract:
The purposes of this study were: 1) to compare the quality of reliability, validity, difficulty and diserimination of a single and double multiple choice test measuring the first-three levels (knowledge, comprehension and application) and the last three levels (analysis, synthesis and evaluation) of Congnitive Domain, 2) to compare the differences of scores between high, medium and low learning achievement groups basing on single and double multiple choice test, 3) to find the interaction between learning achievement levels and the multiple choice tests using different alternatives. The instrument included two forms of an achievement test in General Science Course (Sc.305) for Mathayomsuksa III constructed by the researcher. The first one was a single multiple-choice form; the second was a double multiple choice format converted from the single multiple-choice test measuring the first-three levels and the last three levels of Cognitive Domain. The subjects were 328 Mathayomsuksa III students from 5 secondary schools in Nakornnayok. The subjects were divided into two groups, the first group took the test with the single multiple choice, and the second the double multiple choice test. Each group was also divided into three small groups; high, medium and low learning achievement by the General Science Achievement Test (Sc 204) for Mathayomsuksa II students of Navamarachanusorn School. The test was used as a criterion for determining the concurrent validity of all test constructed. The Kuder-Richardson Formular-20 was used to estimate the reliability index of each test. The Pearsons Product Moment Correlation Coefficient was used to determine the validity of each test. The Biserial Correlation was used to determine the item discrimination. The reliability, the validity and the discrimination of all test were then transformed to the Fishers Z coefficient, and then compared. The Median Test was used to test the differences of the difficulty. The different scores of the single and the double multiple choice test in different groups of learning achievement students tested by two way analysis of variance and the Scheffe method were used to test the difference between the means in posttension comparison. The results of study were: 1. To measure the first three levels of Cognitive Domain, the reliability of the single multiple-choice test was significantly higher than the double multiple-choice test at a 0.05 level. The discrimination of the single multiple-choice test was significantly higher than the double multiple-choice test at a 0.01 level. The double multiple-choice test was significantly more difficult than the single multiple choice test at 0.05 level. There was no significant difference between the validity of the single multiple choice test and the double multiple choice test. 2. To measure the last three levels of Cognitive Domain, there was no significant difference between the reliability, validity, difficulty and discrimination power of the single multiple choice test and the double multiple choice test. 3. There was an interaction between learning achievement levels and the multiple choice that used different alternatives. 4. There was no difference between score from single and double multiple choice test for both high and low learning achievement. Medium learning achievement learners obtain higher score from single multiple choice test than that from the double multiple choice test significantly at .05 level.