Abstract:
The purposes of this study were to (1) develop computer-based testing system with immediate feedback, (2) study the effectiveness of computer-based testing system with immediate feedback applied to learners with different ability levels and different types of feedback, (3) evaluate the quality of computer-based testing system with immediate feedback, and (4) improve computer-based testing system with immediate feedback. Participants were 728 twelfth grade students in mathematics-science program from 7 schools in Bangkok educational service area. The instruments were form of genetics problem solving ability test and form of self-evaluative test of genetics problem solving ability as seen the design and development phase of the developed system. Data analysis was conducted by using descriptive statistics (mean and standard deviation), and inferential statistics (Paired Sample T-Test, Two-way ANOVA and Two-way ANCOVA). Findings were as follows: 1.The developed system consisted of 3 factors (input factor, testing process and consequence process). (1) The input factors consisted of 1) pretest and post-test of genetics problem solving ability, 2) self-evaluative pretest and post-test of genetics problem solving ability, and 3) test of genetics problem solving ability with immediate feedback. (2) The testing process had 3 phases; i.e. 1) registration, 2) computer-based testing system with immediate feedback, and 3) Computing, Finally, (3) the consequence process consisted of 1) the result of pretest of genetics problem ability, 2) the result of post-test of genetics problem solving ability, 3) the result of computer-based testing with immediate feedback, 4) the result of pretest of self-evaluation of genetics problem solving ability, and 5) the result of post-test of self-evaluation of genetics problem solving ability. 2. After the improvement of computer-based testing system with immediate feedback, learners from all ability ranges (poor, moderate, and excellent) exhibited considerably higher mean of post-test of genetics problem solving ability compared to former system at statistical significance level of .05. (1) Learners with low ability level as known as poor group exhibited higher mean of genetics problem solving ability when they were provided with full directive feedback (FDF) compared to other types of feedback at statistical significance level of .05. (2) Learners with moderate ability level as known as moderate group exhibited higher mean of genetics problem solving ability when they were provided with partial directive feedback (PDF) and knowledge of results feedback (KORF) compared to full worked example feedback (FWF) at statistical significance level of .05. (3) Learners with high ability level as known as excellent group exhibited higher mean of genetics problem solving ability when they were provided with KORF compared to FWF at statistical significance level of .05. In term of test items, multiple-choice (MC) test item had higher probability of correct response than constructed-response (CR) test item in every number of attempts at statistical significance level of .05. 3. Quality of the developed testing system had 4 determinants (accuracy, feasibility, utility and property). The highest quality was determinant of accuracy (M=3.87, SD=1.062). The determinant of feasibility was the second (M=3.70, SD=0.997). The third was determinant of utility (M=3.69, SD=0.997) and the last one was determinant of property (M=3.65, SD=0.748)