影響國中基本學力測驗分數的因素之分析

Estimating the Grade Production Function in a Basic Competence Test

本研究之主要目的在於估計影響國中基本學力測驗分數之因素，由於影響分數的主要原因有三大類，包括學校因素、同儕群體因素及個人和家庭背景特徵因素等三大類，透過全台309個學校的問卷調查並獲得8，864個有效樣本，利用Probit兩階段最小平方法（Probit two stage least squares， P2SLS）與兩階段最小平方法（two stage least squares， 2SLS）來估計公私立學校選擇函數、教育的支出函數和教育生產函數。實證結果顯示，學校因素中的公私立學校選擇與城鄉差距都會對基本學力測驗的分數產生顯著的影響，公私立學校中，私立學校的成績比公立學校高，而城市學生的成績比鄉村地區的成績高。其次，學生的休閒與讀書時間、家長的教育程度和家庭所得等變數對基本學力測驗的分數亦皆有顯著的影響。最後，在同儕的效果方面，如家中有考上明星高中或讀國立大學、碩博士、專科學校的兄姊都會對其弟妹有正面的影響。然而，本研究最重要的發現是常態分班的平均分數高於能力分班的平均分數；再者，由研究結果也發現，實施能力分班的學校，好班學生的平均成績顯著高於其他班別學生的成績，此項結果隱含同儕群體效果在好班學生間有正向的影響，而在其他班別學生之間，產生負面的影響，而為確保每位學生接受基本教育的權利，並非只重視優秀學生，忽略了成績較差的學生，故教育部應確實執行常態分班的教育政策。

The purpose of this study is to estimate the impacts of numerous important variables including school effect， peer group effect， family and student background characteristics on student achievement measured by Basic Competence Test using sampling data drawn from the overall survey of 309 high schools in Taiwan. The sample sizes are 8，864. The endogenous decisions on the selection of public and private school， and educational expenditure are estimated by implementing Probit two stage least squares (P2SL) and Two stage least squares (2SLS) approaches. Later， such estimated variables are incorporated into the grade production function. The empirical results indicate that the selection of public-private school， the urban-rural difference， the peer effects， the different leisure time and study time of students， and parental educational and occupational background have significant impacts on student achievements. The other major finding of the estimation results shows that the performances from normal grouping is better than the ability grouping which indicates the Ministry of Education in Taiwan should enforce the policy of normal grouping to place equal importance on all students getting the basic education.