The Factors Influencing Mathematics Achievement of Matthayomsuksa 3 Students, uder the office of Roi Et Educational Area : A Multi-level Analysis

Abstract

This research was aimed at : 1) Study the factors of student levels and classroom
levels that influence mathematic achievement. 2) Developing and evaluating multi-level
structural equation models that influence mathematic achievements. The research
sample consisted of 1,129 Matthayomsuksa 3 students through-out the Roi-Et Province,
during the 2nd semester of the academic year 2011, selected by the two-stage random
sampling technique. The research instruments consisted of (1) a 30-item-mathematics
achievement test with difficulties (p) of items rangedbetween 0.45-0.73 discrimination
value of items (r) ranged between 0.62-0.89 with a total reliability (KR20) of 0.95 2) Scale
on factors affecting achievement with 5 parts with discriminationvalue of item (t-test)
ranged between 4. 13-13.59 with a total reliability (α ) of 0.90.The data analysis was
employed using descriptive statistics from a Multi-level Analysis.
The results were as follows:
1) The student level and classroom level when considering the effects of
variable The direct effect is significant at the 0.01.
2) structural equation model is consistent with empirical data. The
development of the statistical validity of the model, student (χ2 = 168.364, df = 98, p =
0.2137, index CFI = 0.991, TLI = 0.990, RMAEAW = 0.024, SRMRW = 0.026 χ2/df = 1.718).
and classroom-level model (χ2=48.695, df= 32, p=0.1867, index CFI = 0.998, TLI = 0.997,
RMAEAB = 0.026, SRMRB = 0.022, χ2/df = 1.521).
3) The creation and ensure the consistency of the model by model
assumptions associated with the empirical data showed that the model assumptions are
consistent with empirical data. By considering the statistical validity of the model (χ2 =
347.256, df = 238, p=0.1837, χ2/df = 1.459, CFI = 0.980, TLI = 0.997, RMAEA = 0.041,
SRMRw = 0.016, SRMRB =. 0.060) showed that the model results with the class. Attitudes
towards learning. The concept is statistically significant at the 0.01 level, with effect sizes
of 0.545, 0.201 and 0.416, respectively, and by the high school was teaching behaviors are
statistically significant at the 0.01 level, with effect sizes of 0.269 and variable. recapture
the atmosphere in the classroom. There are statistically significant at the 0.05 level, with
the effect size was 0. 068. The combination of classroom and school-level models that,
explain the variability of the factors that influence the academic achievement of students
in the three provinces, were 79.70% and 75.90%.