Abstract:
To compare the reliabilities of rating scale questionnaires with six different response formats : 1. letter scale, 2. square scale, 3.descrip¬tive scale, 4. numerical scale with 5 for very high to 1 for very low, 5. numerical scale with 1 for very high to 5 for very low and 6. graphic scale. The sample included 624 Mathayom Suksa Five students from four public secondary schools in Bangkok. The Evaluation of Teaching Efficiency and The Survey of study Habit and Attitudes were two rating scales questionnaires which used in this research. The test - retest reliability of the scales of each format was estimated by the Pearson Product Moment Correlation Coefficient. A comparison of the reliability coefficients Was done by using the chi - square test and the Fisher's z test. Results indicated that no significant differences were found in the reliabilities among the six response formats of the Evaluation of Teaching Efficiency. For the Surway of study Habit and Attitude, the graphic rating scale had the highest reliability and was significantly different from each of response format at .05 level.