背景：校本評核是教育改革的新方案，不少國家引入校本評核輔助公開考試，拓闊評核的基礎，提升學習效能，然而很多研究發現教師對評核變革的信念足以影響其實施。
目的：香港中學會考中國語文於2005-06學年實行校本評核，本研究以個案形式探討教師對校本評核的信念及其如何回應此變革。
研究對象：兩位高中中文教師
研究方法：採用半結構式訪談及檔分析蒐集資料
研究結果：研究發現乙老師支持校本評核，採取積極態度，安排各種評估活動，豐富學習與評核的內容；而甲老師卻持觀望態度，採取最省時省力的方法實施校本評核而不影響教與學的常態，反而致力幫助學生應付公開考試。
結論：研究揭示教師的信念影響其態度、決策與行動，為政策制定者帶來重要的啟示。
Background: School-based assessment is a new initiative of education reform. Many countries from different parts of the world have introduced SBA to supplement public examinations, broaden the assessment base and enhance learning effectiveness. However, many studies uncover teachers' beliefs, which affect the implementation of SBA.
Aims: SBA has been implemented in Chinese Language of Hong Kong Certificate Education Examination since the academic year of 2005-06. This case study aimed at exploring teachers' beliefs on SBA and how they responded to the change.
Sample: two Chinese Language teachers who teach senior forms
Method: Data were collected through semi-structured interviews and document analysis.
Results: The results showed that teacher B supported SBA and his attitude was positive. He was willing to arrange a variety of activities for assessment, enriching the contents of learning and assessment. However, teacher A with dubious attitude towards SBA, adopting time-saving and energy-saving strategies to implement SBA without hampering the norms of teaching and learning. Effort was focused on helping students prepare for the public examination.
Conclusion: This study highlighted the importance of teachers' beliefs, which exerted influence to their attitudes, decision making and action. The findings have important implications to policy makers.
[Copyright of New Horizons: The Journal of Education, Hong Kong Teachers' Associationl is the property of Hong Kong Teachers' Association at http://www.hkta1934.org.hk/]

This paper details the development and validation of the multiple-choice cloze test for the Hong Kong SAR Government's English language benchmarking initiative. The paper briefly describes the background to and rationale behind the benchmarking initiative. It then discusses how well the test worked and the suitability of a cloze test for use with teachers in a high-stakes examination context. The paper describes two cloze tests, set within the Reading Test component of the test battery, which were destined to be used as specimen benchmark test material and prepared for trailing, together with an established anchor test with known values. Efforts were made to include item types which tapped more than mere lexical / grammatical features, e.g., features of discourse and cohesion. The means which emerged on the two cloze tests would appear to indicate that the difficulty levels of the cloze tests trailed for use with teachers are appropriate. Item discrimination figures were also high, indicating that item types generally worked well and that the cloze test is able to differentiate teachers of differing abilities effectively.
本文詳述由香港特別行政區政府倡議的英語基準試中的多項選擇填空測驗(Multiple-choice clozetest) 的發展與成效。文中先簡介基準試的背景及其理論基礎，然後討論這個測驗運用上的適切性。本文論述的兩個填空測驗是屬於閱讀理解組內的測試類別，是預定作為基準試的測試材料樣本，同時更與一已建立信度的測驗一併使用。在測試項目的揀選方面，已盡量包括各項不同的類別，不僅側重詞匯及語法，還有其他如篇章結構等之特點。而這兩個填空測驗所得出的平均值正好標示上述測驗之程度是適合的，且各項目間的數據區分亦高，由此可見測試的項目大體上是運用成功的。填空測驗確能有效地辨別教師的不同能力。
[Copyright of New Horizons: The Journal of Education, Hong Kong Teachers' Associationl is the property of Hong Kong Teachers' Association at http://www.hkta1934.org.hk/]

This paper aims to explore rater behaviour in assessing oral presentations using verifiable quantitative measures (VQM) as an external validity check on ratings. Twelve raters from a range of backgrounds were recruited to rate 115 Secondary 3 student oral performances in 'individual presentations'. These performances were drawn from a sample of 10 schools participating in a pretest conducted for the commencement of the oral component of Hong Kong's Territory-wide System Assessment in 2006. About 20 students were drawn from each school in three ability categories: low, medium and high. Students were selected based on their internal examination results. Fifty-eight of the 115 student performances were transcribed and assessed on VQM for 'ideas and organisation', 'vocabulary and language patterns' and 'pronunciation and delivery'. VQM results were correlated (Spearman's and Pearson's 'r') against fair average scores derived from Rasch analysis of ratings. The resultant correlations ranged from 0.6 to 0.9. It was concluded that raters were estimating values for constructs highly similar to those measured in VQM.
本研究是以可驗證的量化量度方法，探究英文科說話評審員在說話評估「個人短講」中的評審表現。是次研究，邀請了12位具有不同學歷及經驗的人士擔任說話評審員，對115位中學三年級學生進行說話能力測試，評審學生的說話能力。這些學生是來自於參加2006 年「全港性系統評估」預試的學校。參加預試的學校共有10所，其中包括不同能力組別的學生。每所學校都是根據校內的英文科成績(高、中、低)，挑選20位學生參加說話能力測試。是次研究是從115位參加測試的學生中，抽取了58位學生進行研究，以文字紀錄了這些學生在個人短講中的內容，再以可驗證的量化量度方法(VQM)，評估學生在「內容和組織」、「詞彙和句式」和「發音和表達」三方面的表現。本研究亦利用史比爾曼「」及皮爾遜「r」計算VQM 數據和羅許平均值的相關系數。是次研究所獲得的相關系數為0.6 至0.9 。總而言之，說話評審員給予評級的建構值和VQM 所得的結果甚為相近。
[Copyright of Hong Kong Teachers' Centre Journal is the property of Hong Kong Teachers' Centre at http://www.edb.org.hk/hktc]