利用重複測度探討高中學生升學及在校學業表現之影響因素-某高中案例研究

An ARM-Based study on explanatory factors of high school students' academic performance and college admission:A case study of a high school in Taiwan.

陳建成(CHEN,JIAN-CHENG)

指導教授 ： 王鴻龍

國立臺北大學/商學院/統計學系/碩士(2010年)

摘要

我國高中教育乃是以升大學為目標的準備教育，故學生在高中階段的成績表現一直是各校矚目及探討的重點。本文的研究對象為臺北市某公立高中九十五學年入學之新生，共計912位。在家長逐年重視升學率的情形下，為了解該校學生的學習狀況及課程配置的成效，故蒐集其入學基測分數、入學管道類別、高一至高三各學期各科之段考分數、類組別、高三兩次北區模考各科級分數、大學學測各科級分數等資料，使用直方圖(Histogram)、圓形圖(Circle graphs)、折線圖(Line chart)等統計圖形了解資料結構，並透過平均數T檢定( T test)、變異數分析(Analysis of Variable，ANOVA)及重複測度分析(Analysis of Repeated Measures)等假設檢定來探討資料差異的顯著程度，以此訊息提供學生、教師及學校行政人員作為調整教學的參考依據。
因本文分析的反應變數皆為一連串有相關性的成績資料，在分析各群體間是否存在顯著差異性時，若將其成績平均後使用平均數T檢定，則無法檢定出資料間增減變化的差異性；若不取平均數而改使用變異數分析，則資料間的相關性會讓原本可能有顯著差異的資料，檢定結果變成沒有顯著差異。故本文在分析此類資料時主要使用重複測度分析。
本研究重要結論有：
1、雖然登記分發的學生在入學的基測分數上有顯著的高於申請入學的學生，但此兩種不同入學管道的學生在高一各科的學習成效上卻沒有顯著的差異性。故教師不需要太過擔心申請及登記的學生在學習成效可能有所差異的問題。
2、在升大學考試的表現上，自然組的學生在數學科和自然科的表現上具有顯著的優勢，而社會組的學生則只有在社會科的成績上高於自然組學生，但未達顯著。
3、在五科的整體表現上，不同類組的學生成績表現並無顯著的差異性。因此，家長不必太過擔心學生選讀社會組會在升大學考試上處於劣勢
4、在物理科和化學科的在校成績表現上，三類組學生在高二三的成績表現顯著的高於二類組學生。顯然，自然科較優秀的學生大部分都會選擇三類組就讀。在各群分析時，二三類組學生在高二三的成績表現並無顯著的差異性，但在高一時的成績則有部分高分群別呈現出顯著的差異。因此，學校應鼓勵高一自然科成績不錯但非頂尖的學生選讀二類組班。
透過上述分析的結果，期望對於學生、家長及教師在加強學習效益上有所幫助，並期許該校未來在升大學的成績表現上可以有所提升。

關鍵字

重複測度

並列摘要

Given that high school education in Taiwan has been heavily oriented toward preparation for college admission, high schools lay great emphasis on their students’ academic performance. This paper aims to investigate this phenomenon and thus takes as subjects 912 students, who in 2006 were admitted to a high school in Taipei. To better understand how much the students have learned and the efficacy of the courses offered, various data are gathered, including the students’ grades on the Basic Competence Test, the channels of their admission to the high school, their grades on midterms and finals in the space of three high school years, Humanities/Sciences Class placement, their percentile grades on two mock college entrance exams, and their percentile grades on the Scholastic Ability Test (SAT). Histograms, circle graphs, and line charts are utilized to analyze the data, while T test, Analysis of Variable (ANOVA), and Analysis of Repeated Measures (ARM) are applied to explore the disparity in the data. The results could serve as reference points for teachers and school personnel to modify the curriculum and instruction.
The variables this paper intends to analyze are mostly grades that are correlated, and thus T test could not distinguish the distinct disparity among data variations because the grades are averaged. If ANOVA is used instead, the correlation between data would make the data, which might be of significant difference, have no significant difference in test results. Therefore, the methodology here focuses on ARM to analyze the data.
The conclusions reached are as follows:
1.Although the students admitted to the high school through registration have significantly higher grades than those through application on the Basic Competence Test, there is no significant difference between the two groups’ first-year academic performance. Hence, teachers need not be over-concerned that those grades might be indicative of learning outcomes.
2.On the GSAT, students in Sciences Classes enjoy a salient advantage on both Math and Science exams, while students in Humanities Classes excel on the Social Science exam. while students in Humanities Classes have a better academic achievement than those in Sciences Classes on the Social Science exam only, yet the result is not significant.
3.Concerning the overall academic performance on the SAT, little distinction is seen among students placed in three different groups. Thus, parents need not be over-concerned that students in Humanities Classes will be in an inferior position on the SAT.
4.As far as the subjects of physics and chemistry are concerned, students placed in the second group of Sciences Classes (Biology/Medicine-related majors) have much better academic performance than do those in the first group of Sciences Classes (Engineering-related majors) in their junior and senior years at school. Apparently, students with higher proficiency in science mostly select the second group of Sciences Classes (Biology/Medicine-related majors) to study in. In the cross-group analysis of the junior and senior general academic performance, there is no significant difference between both groups in Sciences Classes. However, significant distinction is found in some higher-level groups in their freshman-year academic performance. Therefore, school administration should encourage freshmen with decent, but not exceptional academic performance in science to choose the first group of Sciences Classes (Engineering-related majors) to study in.
Based on the results listed above, it is hoped that students, parents and teachers would have a better idea as to how learning effectiveness can be enhanced, and that students from that particular high school will perform better on college entrance exams in the future.

並列關鍵字

Repeated measurements

參考文獻

李佳玲(2002) 大學入學考試中心學科能力測驗與高中在校成績關係之研究。國立台北師範學院國民教育研究所碩士論文。

李秉書(2003) 高一學生選組及其學業預測之研究－以自然組物理科為例。國立臺灣師範大學物理研究所碩士論文。

Keselman, H. J., Algina, James, Kowalchuk, Rhonda K. ＆ Wolfinger, Russell D. (1999). A comparison of recent approaches to the analysis of repeated measurements. British Journal of Mathematical and Statistical Psychology, vol. 52, p. 63-78.

Keselman, H. J., Algina, James , Kowalchuk, Rhonda K. (2001). The analysis of repeated measures designs : A review. British Journal of Mathematical and Statistical Psychology, vol. 54, p. 1-20.

Algina, James, Keselman, H. J. (1997). Testing Repeated Measures Hypotheses when Covariance Matrices are Heterogeneous: Revisiting the Robustness of the Welch-James Test. Multivariate Behavioral Research, vol. 32(3), p. 255-274.