再探英語廣泛閱讀成效之後設分析研究

Abstract

近年來，許多廣泛閱讀研究的文獻中，記載了語言學習的正面成效，使得廣泛閱讀成為漸漸受到矚目的教學方法，尤其是在亞洲。雖然有大量的實證研究顯示廣泛閱讀的正面效果，但其應用在學校教學情境上仍受到質疑。因此，接續先前廣泛閱讀的後設分析研究 (如Krashen，2007；Nakanishi，2015)，本研究繼續針對泛讀之整體效益做一個系統性的後設分析，並專注在以英語為外國語的學習情境之中。此外，更針對五個調節變項做進一步的研究，包含語言學習成果、學習者之教育程度、廣泛閱讀之時間長度、廣泛閱讀之類型，以及泛讀後的活動類型。本研究在文獻搜尋中選出47篇於1990年到2015年之間發表的論文－包含研討會及期刊論文與博士論文－進行後設分析。 以下總結此研究的主要發現：第一，廣泛閱讀對於英語為外國語之學習者的語言發展具有中等的整體效應 (d = 0.49)；第二，整體閱讀能力為語言學習成果中成效最佳的項目，接近大效應值 (d = 0.77)，而閱讀速度 (d = 0.69) 及文法 (d = 0.52) 的成效則介於中至大效應值之間。研究顯示廣泛閱讀在字彙習得、拼字與閱讀理解能力此三個項目中的成效，只有介於小至中效應值之間 (約d = 0.40)，而在聽力理解能力及寫作方面，只有呈現小效應值 (d = 0.23)；第三，隨著學習者的教育程度提高，廣泛閱讀的重要性也一併增加。研究顯示大學生最能從廣泛閱讀中得利 (d = 0.57)；第四，本研究在泛讀的時間長度此調節變項中所呈現的結果，與先前的文獻不同：本研究顯示，泛讀實施一學期的成效 (d = 0.71) 大於泛讀實施兩學期的成效 (d = 0.40)；第五，課後的廣泛閱讀是三種廣泛閱讀類型中，成效最好的，並且展現中度效應 (d = 0.56)。泛讀同時實施於課中與課後的成效為小至中效應值 (d = 0.42)，成效最小的則是課中廣泛閱讀 (d = 0.40)；第六，同時側重語言形式與側重語意的泛讀後活動呈現了近高度效應 (d = 0.73)，而側重語言形式的泛讀後活動與側重語意的泛讀後活動，具有相同效應值 (d = 0.47)。效果最小的則是沒有泛讀後活動的這個項目 (d = 0.29)。 此後設分析研究證實了廣泛閱讀在改善英語為外國語之學習者的語言程度上，相當具有影響力，並進一步建議學校應將廣泛閱讀融入課程中，最後並呼籲研究者在未來的泛讀研究中，應執行延遲後測，以證實廣泛閱讀此種典型的隱性學習方法，與其它著重顯性教學的方式，具有更好的效果。

Extensive reading (ER) is an instructional approach that has increasingly received attention in the past few years, especially in Asia, in terms of its well-documented linguistic effects. While there is a large body of empirical research showing positive effects of ER, concerns over its applicability to school contexts still exist. Accordingly, following up on previous ER meta-analyses (e.g., Krashen, 2007; Nakanishi, 2015), the current study reports on a synthesis and meta-analysis on the overall effect of ER by drawing on research specifically done in EFL contexts. Also, five moderators are specified for further investigations, including linguistic outcomes, educational levels of learners, length of ER treatment, types of ER, and types of post-ER activities. The current study meta-analyzes 47 primary studies, including conference and journal articles and doctoral dissertations, all of which are published between 1990 and 2015 and are retrieved via a comprehensive literature search. The major findings are summarized below. First, ER contributes to a nearly medium overall effect (d = 0.49) on EFL learners' language development. Second, overall reading proficiency ranks the first, showing a near-large effect (d = 0.77). Reading speed (d = 0.69) and grammar (d = 0.52) reveal medium to large effects. ER is found to produce small to medium effects (approximately d = 0.40) on vocabulary acquisition, spelling, and reading comprehension. Small effects (d = 0.23) are shown in both listening comprehension and writing. Third, the importance of ER is enhanced as learners' educational levels advance, with university students benefiting the most from ER (d = 0.57). Forth, the current study presents contradictory results with previous findings in ER treatment duration, demonstrating that one-semester treatments (d = 0.71) show better effects than two-semester ones (d = 0.40). Fifth, out-of-class ER yields the largest and a medium effect (d = 0.56), followed by ER done both in-class and out-of-class (d = 0.42) and then in-class ER (d = 0.40). Sixth, a near-large effect (d = 0.73) is found in ER with both language-focused and meaning-focused activities. Language-focused activities and meaning-focused activities reveal identical effects (d = 0.47), followed by the smallest effect (d = 0.29) in no post-ER activities. This meta-analysis confirms the power of ER in improving EFL learners' language proficiency and further suggests that ER should be integrated into school curricula. Finally, administering delayed post-tests in future ER studies is urgently called for, due to the need to prove the claim that ER as a typical implicit learning approach retain the effects well, compared to other instructional approaches with a focus on explicit teaching.