Writing English Language Test Items as a Learning Device: a Principle of Habit Formation Rules

The present study aims to uncover how valid and reliable the multiple-choice test items as learning device are and their effect on permanent language habit formation. In this pretest-posttest one group pre-experiment study, we observed and analyzed the English test constructed by junior high school teachers. The Pretest on constructing the test items was administered to fifteen English teachers and some weaknesses were revealed. These can be seen from the total score of 855 and the mean score was 57.00.  Treatment was given by ensuring them the rules of writing the test based on a principle of habit formation. The competence of teachers in constructing appropriate test items increased to higher quality indicated by scores of 79.33. It means the treatment has significantly influenced the ability of English teachers in constructing appropriate tests. The tests were tried out to 40 students of Junior High School and the result showed that 63 test items were valid and 37 was invalid. The fifty-valid test was administered to the students of different Junior High School and the result showed that the mean score was 72.53. 21 students from 40 students passed the minimum rank and 19 students still needed improving. The highest score was 95 and the lowest one was 26. The medium was 78 and the mode was 53. After measuring the reliability level, the standard deviation ranged 0.50 – 1.00, meaning that the test was appropriate to be applied as measuring device. The valid test was then reconstructed to be raw material of software module.