1990年與1996年男女教育量差異與發展類型探索

Exploring Gender Differences in Education and Clustering the Pattern of Development in 1990 and 1996

男女性別教育量的差異，向來為人忽視。本研究以聯合國教科文組織統計（United Nation Educational， Scientific， and Cultural Organization，簡稱為UNESCO，2000）與中華民國教育統計（民89）的1990年及1996年資料，探索各國男女性教育在學率之差異為男女性教育在學率差異的國家分類情形。本研究得到結果如下：第一、1990年與1996年各國之男女性的初等教育在學率、中等教育在學率、高等教育在學率的幾尼係數，大致是女性高於男性，表示女性的教育機會低於男性；同時如果以全球的初等、中等及高等教育在學率進行幾尼係數計算，發現初等教育在學率的均等率最高，其次為中等教育在學率，第三為高等教育在學率。第二、在男女性教育指標差異量的國家分類上，不同的男女性教育指標差異量的分類，有不同的國家數；就如1996年的二個指標的分類、高度、中度及低度男女性教育量差異的國家數；就如1996年的二個指標的分類，高度、中度及低度男女性教育量差異的國家數各有31、38及9個。第三、以區別分析檢定1990及1996年分類準確度在96%以上。第四、以斯皮爾曼相關檢定各種分類，其相關係數在.601至.808之間，達0.1顯著水準。第五、1990及1996年之高度、中度、低度男女性教育在學率差異的國家分類後，如以變異數分析檢定三群之間，都有差異，達到0.1顯著水準，經事後比較發現，各群之間大都有顯著差異。最後，檢定台灣的男女性教育在學率的差異，她大致被列為男女性教育量差異均等的國家群組。

Gender disparity is always ignored. The main purpose of this study uses the raw data which was collected from the United Nation Educational， Scientific， and Cultural Organization（UNESCO，2000）and Education Statistics of the ROC（2000）about the ratio of males and females in education in 1990/1996 to understand gender disparity worldwide， and uses the ratio of female primary， secondary， and higher education to cluster that pattern of development. The research results are as follows：First， using the ratio of male and female primary education to calculate the Gini index， the results are found that all the females’ Gini indices are higher than the male’s in the ratio of primary， secondary， and higher education， in 1991 and 1996， respectively；that is， the educational opportunity of males is higher than female’s in all levels of education. Second， because the indicators are different， the countries form three groups in the Cluster Analysis， that is high ratio of educational difference countries， middle ratio of educational difference countries， and low ratio of educational difference countries are different. For instance， using two indicators in 1996 can form three groups， that is， high， middle， and low ratio of educational difference countries are 31，38 and 9， respectively. Fourth， in order to test the consistency of these clusters， Discriminant Analysis is used to reclassify the 4 clustered countries. Above 96% of countries are correctly classified by the indicators. Fifth， the Spearman rank correlation is used to test the rank of the clustered countries by using 4 clustered countries. The correlations are between .602 to .808， and there are significant （p＜.01）， too. Finally， according to the discriminant scores of the ratio of education difference， in the ROC， Taiwan， three clusters are classified as a middle ratio of educational difference country.