| dc.description.abstract | 在第二語言寫作文獻裡，文法糾正回饋研究顯示集中回饋(即糾正某一特定錯誤類別)比非集中回饋(即糾正所有錯誤類別)還要有效(e.g. Ellis et al., 2008; Farrokhi & Sattarpour, 2011; Frear, 2010; Sheen et al., 2009)。而根據先前研究者的建議，糾正回饋為集中或非集中，取決於所糾正錯誤類別的數量為何(Ellis, 2009; Ferris, 2010)，而先前集中回饋研究則只針對某一種錯誤類別進行糾正(e.g. Bitchener, 2008; Rouhi & Samiei, 2010; Sheen, 2007)。然而，集中回饋效益可能取決於所糾正的某一種錯誤類別範圍多大或多小(Bitchener & Ferris, 2012)，而所謂「一種錯誤類別」的概念在集中回饋研究裡卻相當模糊。由於先前研究並未探討錯誤類別概念在集中回饋效益之影響為何，故有關錯誤類別概念之探討應該能對集中回饋文獻提供一些洞見，並幫助我們進一步瞭解集中回饋效益。
本論文旨在探討錯誤類別概念在集中回饋效益之影響為何。本論文進行兩組實驗來分別探討集中回饋在文法規則導向錯誤(即英文冠詞錯誤和主動詞一致性錯誤)以及詞彙導向錯誤(即動詞-名詞搭配錯誤)之效益為何。第一組探討集中回饋在文法規則導向錯誤的實驗顯示，當冠詞錯誤範圍縮小至不定冠詞(a)用以表示指示物為第一次提及，和定冠詞(the)用以表示前指照應的相關錯誤時，集中回饋是有效的。此外，受試者不因冠詞錯誤的糾正而減少他們在接下來寫作上冠詞的使用。而另一方面，當冠詞錯誤範圍擴大至包含不定冠詞(a)用以表示指示物為第一次提及，以及定冠詞(the)用以表示前指照應和第一次提及的相關錯誤時，集中回饋則是無效的。然而，察覺到冠詞錯誤類別的受試者，其冠詞使用的表現，比起沒察覺到該錯誤類別的受試者還要好。
集中回饋效益在主動詞一致性錯誤的結果大致上與冠詞錯誤的結果相似。當主動詞一致性錯誤範圍縮小時(只有分別糾正copula be或lexical verb錯誤)，集中回饋顯示有效。此外，受試者不因主動詞一致性錯誤的糾正而減少他們在接下來寫作上主動詞一致性結構的使用。另一方面，當主動詞一致性錯誤範圍擴大時(糾正所有copula be和lexical verb錯誤)，集中回饋則顯示無效。然而，察覺到主動詞一致性錯誤類別的受試者，其主動詞一致性結構使用的表現，比起沒察覺到該錯誤類別的受試者還要好。
第二組探討集中回饋在詞彙導向錯誤的實驗顯示，當糾正三到四個動詞-名詞搭配錯誤時，集中回饋是有效的。另一方面，只糾正一至兩個動詞-名詞搭配錯誤時，集中回饋則是無效的。然而，在只產生一個搭配錯誤並受到糾正的受試者裡，察覺到動詞-名詞搭配錯誤類別的受試者，其動詞-名詞搭配的使用，比起沒察覺到該錯誤類別的受試者還要好。
綜合以上結果，本論文提出下列三項主要發現：(一)、當冠詞錯誤和主動詞一致性錯誤範圍縮小，以及三到四個動詞-名詞搭配錯誤受到糾正時，集中回饋顯示有效。(二)、當冠詞錯誤和主動詞一致錯誤範圍擴大，以及只有一個動詞-名詞搭配錯誤受到糾正時，學習者對於所糾正錯誤類別的察覺對於回饋效益是有影響的。(三)、當冠詞錯誤和主動詞一致性錯誤範圍縮小時，學習者不會因受到集中回饋的糾正，而在接下來的寫作裡，避免使用該語言結構。本論文根據上述研究發現，對書面糾正回饋研究以及相關教學實踐應用等議題進行討論。
| zh_TW |
| dc.description.abstract | In the literature on second language writing, corrective feedback studies have shown focused feedback (which targets one specific error type for correction) to be more effective than unfocused feedback (which corrects all grammar error types) (e.g. Ellis et al., 2008; Farrokhi & Sattarpour, 2011; Frear, 2010; Sheen et al., 2009). The criterion for judging whether the feedback is focused or unfocused, according to what corrective feedback researchers have suggested, depends on the number of error types targeted for correction (Ellis, 2009; Ferris, 2010), and focused feedback has been operationalized to correct only “one error type” (e.g. Bitchener, 2008; Rouhi & Samiei, 2010; Sheen, 2007). The concept of “one error type”, however, remains unclear in focused feedback studies because the effectiveness or ineffectiveness of focused feedback practices might depend on how narrowly an error type is defined (Bitchener & Ferris, 2012). Since none of the focused feedback studies has investigated the notion of error type in feedback effectiveness, an investigation into the notion of error type may shed new light on focused feedback literature and advance our understanding of focused feedback effectiveness.
This dissertation accordingly, is an investigation of how the notion of error type influences the effectiveness of focused feedback. Two experiments were conducted to investigate focused feedback effects on rule-based errors (i.e. English article errors and subject-verb agreement errors) and lexically-based errors (i.e. verb-noun collocation errors) separately. The first experiment conducted to investigate the feedback effects on the rule-based errors showed that focused feedback was effective for article errors when article errors were narrowly limited to errors involving the indefinite article a for the first mention and definite article the for subsequent mentions. Additionally, learners whose article errors were corrected did not reduce the article usages in their subsequent writing. On the other hand, focused feedback was ineffective for article errors when article errors were broadly defined including errors involving the indefinite article for the first mention and the definite article for subsequent mentions as well as for the introductory use. However, learners who perceived the targeted error type of focused feedback on the broadly defined article errors made more improvements than those who did not perceive the targeted error type.
The results of focused feedback on the subject-verb agreement errors generally share similarity with results of focused feedback on the article errors. Focused feedback was effective for the narrowly defined subject-verb agreement errors (whether copula be or lexical verbs separately). Additionally, learners who received corrections on subject-verb agreement errors narrowly defined did not reduce the structure of subject-verb agreement in their subsequent writing. On the other hand, focused feedback was ineffective for the broadly defined subject-verb agreement errors (correcting both copula be and lexical verbs). Nevertheless, learners who perceived the targeted error type of focused feedback on the broadly defined subject-verb agreement errors made more improvements than those who did not.
The second experiment conducted to investigate the effects of focused feedback on the lexically-based errors showed that focused feedback was effective when three or four tokens of verb-noun collocation errors were found and corrected. On the other hand, focused feedback was ineffective when one or two tokens of verb-noun collocation errors were found and corrected. However, among learners who produced and were corrected on only one token of a verb-noun collocation error, those who perceived the targeted error type made significantly more improvement in the subsequent writing than those who did not.
Taking all the results together, the present dissertation research suggests: (1) Focused feedback is effective when English article and subject-verb agreement errors are narrowly defined, and three or four tokens of verb-noun collocation errors are found and corrected. (2) Learners’ perception of targeted error type is influential in feedback effectiveness when English article and subject-verb agreement errors are broadly defined, and only one token of verb-noun collocation error is found and corrected. (3) Focused feedback does not cause learners’ avoidance of targeted language uses when English article and subject-verb agreement errors are narrowly defined. Implications for the present dissertation research for the written corrective feedback research community and teaching practices are discussed. | en_US |