建置華語文診斷與補救教學系統

The research of building the diagnostic and remedial instruction system for Chinese learning

張庭綱(Ting-Kang Chang)

指導教授 ： 蕭顯勝

國立臺灣師範大學/科技學院/科技應用與人力資源發展學系/碩士(2012年)

摘要

在過往的華語文測驗中，往往只告在各項測驗上的成績，忽略了診斷後的回饋機制，讓學習者只知其目前的程度，卻無法了解個人在哪個概念上發生錯誤，即時給予補救。本研究根據華語文基礎能力等級的閱讀與詞彙認知概念建立知識結構，藉由貝氏網路具診斷與預測的特性，以及順序理論的試題路徑分析，發展華語文診斷與補救教學系統，比較適性化補救教學與紙本補救的方式，對華語文基礎能力等級的外籍人士之補救教學成效，以及閱讀與詞彙錯誤概念發生情形之影響與改善。
本研究以台北市某大學國語教學中心基礎能力等級65名學習華語文的外籍人士為研究對象，採準實驗研究法，先以能力診斷測驗得知受試者目前華語文的能力以及錯誤的節點，之後分別進行不同的補救教學模式，之後再進行補救教學成效測驗。根據兩次測驗之結果，分析適性化補救教學與紙本補救教學在不同的基本背景之學習成效與知識節點發生錯誤情形。本研究的主要發現如下：
1.使用適性化補救教學系統的受試者，其補救教學成效測驗結果明顯優於使用紙本補救教學的受試者。
2.使用適性化補救教學系統的受試者，補救教學成效測驗中知識節點發生錯誤概念的情形，明顯少於使用紙本補救教學的受試者。
3. 在經過適性化補救教學系統後，漢字文化圈與非漢字文化圈的受試者兩者的能力差距有縮小的情況。
4.對以非華語文為母語的外籍人士而言，在華語文基礎能力等級時，文法及詞彙的語意是最容易有錯誤概念發生之情況。

關鍵字

華語文 ； 貝氏網路 ； 適性化診斷 ； 補救教學

並列摘要

In the past, Chinese tests only informed to the students that they got the test scores, but ignored the feedback after the tests. The learners can not understand the wrong concepts that they occurred, and go to the remedial instruction. In this study, we built the knowledge structure according to the reading and vocabulary concepts of the new TOCFL Level 2, and built diagnostic and remedial instruction system for Chinese learning by using Bayesian network and Ordering theory. We compared the results of the remedial instruction, and the errors concepts of reading and vocabulary that the learners occurred between using ours system and using the paper-based remedial material.
In this study, the subjects are 65 students who are studying at the Mandarin Training Center in the university in Taipei City. At first, we gave them a diagnostic test. After the test, the system informed them the score and the ability of Chinese. Afterward the students of the experimental group used our system to do the remedial instruction, and the students of the control group used the paper-based remedial material to do the remedial instruction. Finally, all of the subjects tested for the effectiveness of the remedial instruction.
The purpose of this research was to find that the students based on different background, such as Chinese-character cultural circle or the time of learning Chinese with different remedial instruction methods would show variations on their learning effect and the errors concepts of reading and vocabulary. The study reached the following conclusion :
1. In the remedial instruction achievement, the students of the experimental group were significant better than the students of control group.
2. In the remedial instruction achievement, the error concepts that occurred by the students of the experimental group were significant less than the students of the control group
3. The capacity gap between the Chinese-character cultural circle and non-Chinese-character cultural circle is reduced by using our system to do the remedial instruction.
4. For all the students in the new TOCFL Level 2, the most difficult concepts are grammar and vocabulary.

並列關鍵字

Chinese as a Second Language ； Bayesian Network ； Adaptive Diagnosis ； Remedial Instruction

參考文獻

周雅釧、黃志勝、施淑娟、郭伯臣（2009）。結合線上診斷評量系統之適性補救教學研究。網際網路技術學刊，10（4），419-425。

彭妮絲（2009）。華語文閱讀教學研究─以閱讀理論與個案為基礎之探討。聯大學報，6（1），39-61。

黃湃祥、江新合、洪振方（2007）。知識結構網路表徵分析之研究─高一基礎力學為例。師大學報，52，79-104。

劉湘川（2008）。多點記分核平滑化無參數試題反應理論及其應用。測驗統計年刊，15（上），13-27。

蔡佳芸（2010）。「華語文數位教學設計」數位學習課程設計與發展。私立淡江大學教育科技學系碩士班碩士論文，未出版，台北。