การพัฒนาโมเดลมูลค่าเพิ่มพหุระดับเพื่อการวัดประสิทธิผลของโรงเรียน / เพ็ญภัคร พื้นผา = Development of multi-level value-added models for measuring school effectiveness / Penpak Pheunpha
The objectives of this study were: 1) To develop multi-level value-added model (the hypothetical model) for measuring school effectiveness after adjusting non-school variables including: student demographics, input, and school external contextual variables. 2) To examine non-school variables affecting GPA in multi-level value-added models. 3) To what extent the value added scores differ between schools using the hypothetical model. 4) To develop value-added models after controlling non-school variables, and adding school processional variables which we call the processional model. 5) To what extent value-added scores from processional model differed between schools. 6) To compare school ranks using value-added score and GPA the separately, and 7) To compare school ranks using value-added score and five indicators (value-added score, growth rate, students’ satisfaction, teachers’ satisfaction, and absence rate) independently. School effectiveness can be defined as the value-added score. The value added score measures student achievement at the end of a period of formal schooling that is below, near, or above what one would expect of schools. The controlling factors of the value-added score include: student demographics, input, and external contextual variables. Data was collected from secondary schools in Bangkok and Nonthaburi Province, Thailand. Our sample consisted of 1,852 ninth-grade students and 446 teachers from 49 schools; the class size ranged from 21 to 54 students. In the hypothetical model, the proportion of variance explained (R2) is 0.76. This indicated that 76% of the true between-school variance in grade point average is accounted by these six factors: mean prior attainment, student–teacher ratio, prior attainment, student expectations, student’s opportunity to study outside and gender. The total variance explained by the hypothetical model is 69.9%. The result is higher than the conditional model which is around 18.6%. Value-added scores and GPA have a moderately positive correlation (r = 0.49), similarly value-added scores have relatively high correlation with growth rate (r =0.63). Meanwhile, value-added scores have no relationship with prior attainment (r = 0.00), and have a little bit association with students’ satisfaction (r = 0.29). The proportion of variance explained by the processional model is 77.9%. These results are higher than the hypothetical model which is around 2%, teaching quality affected the GPA the most when compares to other professional variables. Intraclass correlation in student level reduces 75.3%. When comparing school ranks using value-added scores and GPA separately, the school ranks are significantly different. Also, when classifying school effectiveness using value-added scores and five-indicators separately, correlation between two approaches were moderate (0.50). Value-added score’s variance in each group (effective school and ineffective school) is equal. Therefore, educational stakeholders should consider many variables, such as student background and school context, before evaluating a school’s performance. The value-added measures are also tools for (1) school improvement, (2) making school officials accountable to policy makers and (3) reporting outcomes to parents and communities; used properly value-added scores are reliable, precise, and consistent indicators of overall school performance.