การศึกษาพัฒนาการความสามารถทางคณิตศาสตร์ของนักเรียนโดยใช้วิธีการปรับเทียบแนวตั้ง ตามวิธีทฤษฎีตอบสนองข้อสอบร่วมกับวิธีการปรับเทียบเชิงเส้นตรง / พัชรี จันทร์เพ็ง = A study of students' mathematical ability growth using vertical equating combining IRT and linear equating methods / Putcharee Junpeng
The purposes of this research were to investigate the efficiency of vertical equating combining IRT by three-parameter logistic model and linear equating methods with internal anchor tests, and to study of Matayomsuksa 1-3 students' mathematical ability growth. The students in each grade level was used to equate the students' mathematical ability scores across grade levels so that they were on the same scale scores. The students in each level were divied into three sub-groups accordingly to their abilities; high, fair, and low ability groups. The sample was 1,941 students obtained by stratified random sampling from Matayomsuksa 1-3 students of Office of Sakonnakhon Educational Service Area Zone I.The students in each grade level were divided into two groups; the equated and validated groups. There were altogether 543, 534, and 504 students respectively for equated groups, while there were 113, 120, and 127 students for the validated groups. They were used for data collecting with the Mathematical ability test, the numerical and operation, for the Matayomsuksa 1-3. Each test had 30 items. The reliabilities of test were 0.803, 0.912, as well as 0.857. There were 6 internal anchor items between the M.1 test and the M.2 test, 6 internal anchor items between the M.2 test and the M.3 test. The research data were analyzed by employing BILOG 3.04 for estimating students' abilities parameters, employing EQUATE 2.0 for expressing linear interpolation equations, employing SPSS for Windows version 11.0 for expressing linear regression equation by mean of simple least square method. The results were as follows : 1. The standard error of the equating study model from students of three different ability levels was 0.063 that was low. The discrepancy index of equating in range of satisfy was the acceptable criterion. (C=0.225) 2. The study of the growth rate on students' Mathematical abilities found the significance at .01 level of mean scores of three different levels. It could be concluded that the growth rate on Mathematical ability from M.1 students' to M.2 students' was 0.167, and M.2 student's to M.3 students' was 0.174. Moreover, M.1 students' to M.3 students' was 0.312, that was the highest. In the growth rates study on students' Mathematical abilities were divided into three groups of abilities namely: low, fair, and high groups. considering the growth rates were found that all groups had gain scores accordingly class levels. The student in the high ability group had the highest growth rate that followed by the fair ability group and the low ability group respectively.