การเปรียบเทียบคุณภาพของการกำหนดมาตรฐานระหว่างวิธีแองกอฟที่ได้รับการปรับปรุงกับวิธีบุ๊คมาร์ค / สุรีพร อนุศาสนนันท์ = A comparison of the quality of standard setting between the modified Angoff method and the Bookmark method / Sureeporn Anusasananun
The objectives of this study are 1) to compare the quality of setting standards assessed by the modified Angoff method with that of the Bookmark standard-setting method, in terms of validity, reliability, propriety, and feasibility, and 2) to investigate the effect of the number of judges and the number of occations on the standard setting of the two methods. The research sample is comprised of 12 mathematics teachers (also serving as judges) and 987 Matthayom Suksa 3 students, Academic Year B.E. 2549, in Chonburi. The research instruments include 4 multiplechoice mathematical achievement tests for the students’ first semester, totaling altogether 100 items. Their scores were taken into thorough analysis—namely a) degree of difficulty and b) power of discrimination—on the basis of the IRT model. The data on the two standard-setting approaches from these teachers were analyzed in relation to reliability under the generalizability theory and analyzed in validity under correlation. As for their propriety and feasibility, the analysis of their median as well as inter-quartile range was performed. The following are the research findings: 1. The achievement tests attain 7 levels of cut scores, i.e. excellent (A), very good (B+), good (B), rather good (C+), fair (C), poor (D+), and very good (D). Via the modified Angoff method, the scores range from 72.86, 64.31, 56.18, 43.37, 33.59, 21.34 to 12.98 respectively for the first occation, from 73.28, 63.74, 55.28, 45.11, 35.67, 22.94 to 12.51 respectively for the second occation, and from 73.49, 63.25, 52.82, 52.82, 41.46, 31, 19.07 to 11.72 respectively for the third occation. By the Bookmark method, the scores range from 80.42, 65.75, 47.5, 33.17, 24, 14.75 to 7.83 respectively for the first occation, from 80.58, 66.67, 47.5, 33.75, 24.58, 14.75 to 7.83 respectively for the second occation, and from 80.92, 64.08, 47.33, 34, 24.08, 14.83 to 7.92 respectively for the third occation. 2. The reliability derived from the Bookmark scheme is significantly higher at the .01 level for all the 7 cut score ranks than that from the modified Angoff technique. 3. The correlation coefficients between the students usual achievement scores and their standardized scores obtained from the Angoff approach are .661-.678 and are statistically significant at the .01 level. The correlation coefficients between their usual scores provided by the school and those determined by the Bookmark method are .533 and carry a .01 significance but no significance at all at the .05 level. 4. The mean of the judges’ opinion-ranking scores as to the propriety and feasibility of the Bookmark approach is significantly higher that that of the modified Angoff scheme. 5. The greater the number of judges and number of occations, the higher the reliability coefficients.