父母處罰之管教方式對高中、高職與五專學生成績的影響

Abstract

本文以子女做錯事時父母之反應做為研究主軸, 以台灣教育長期追蹤資料庫(TEPS) 第一波的學生與家長問卷做為研究對象, 探討父母處罰青少年子女對其成績之影響。首先可以觀察到, 不論是根據何種分類方式(例: 家庭收入、學校所在地或是父母教育程度), 處罰型的管教方式與非處罰型的管教方式比例約為一比三。在普通最小平方法(OLS) 的分析下, 發現處罰型管教方式對於學生成績有顯著負向影響。然而, 因普通最小平方法隱含內生性問題, 導致結果可能不具一致性,為取得一致性的結果, 本文使用兩階段最小平方法(2SLS) 進行估計, 文中使用手足之間的關係(是否彼此談心與是否彼此爭吵) 及父母因子女成績主動連絡學校做為工具變數(instrument variable), 三項變數的兩兩組合皆符合了工具變數所要求的兩項特性: 有效性(relevance) 與外生性(exogeneity), 進一步提供了一個更為可信的估計結果。其估計結果顯示, 處罰之教養方式不只具有顯著負向效果, 其係數絕對值亦較普通最小平方法來得大, 表示樣本之特定族群可能對於工具變數的反應極為劇烈, 以致結果有高估的可能性。希望後續的研究能進一步找出可能原因進一步解釋處罰與子女成績細部的關聯, 並提供有效方法減少父母使用處罰做為管教方式的可能以提升學生之學業表現。

By applying the data of the first wave from TEPS (Taiwan Education Panel Survey) and using the OLS method, we investigate the connection between the parents’ responds of children’s misbehaviors and grades. First, the data shows that the ratio between punishment parenting styles and non-punishment parenting styles is 1:3. Second, the results of OLS analysis show a significant negative effect of punishment on an adolescent’s grades. However, the OLS results may contain an endogenous problem and cause the conclusion to be less reliable and inconsistent. To solve this problem, instrument variables and 2SLS model were utilized. In our study, all instruments proved that they are provided with essential relevance and exogeneity; therefore, the 2SLS estimations are more reliable. Although utilizing 2SLS method led to a bigger coefficient of punishment variable, which suggested that some parts of the sample are very sensitive to instruments, the relation between punishment and school performance is still significantly negative. In conclusion, there is an inverse relationship between punishment practice and an adolescent’s education achievement. We expect future research can further investigate the relation between the two and figure out the effective methods to reduce the use of punishment in order to improve the school achievement of teenagers.