• 沒有找到結果。

EFL大學生寫作動機及寫作回饋的認知對寫作成績之影響

N/A
N/A
Protected

Academic year: 2021

Share "EFL大學生寫作動機及寫作回饋的認知對寫作成績之影響"

Copied!
86
0
0

加載中.... (立即查看全文)

全文

(1)國立臺灣師範大學英語學系 碩. 士. 論. 文. Master’s Thesis Department of English National Taiwan Normal University. EFL 大學生寫作動機及寫作回饋的認知 對寫作成績之影響. The Effects of Self-determination and Perceptions of Teacher Corrective Feedback on L2 Writing Performance. 指導教授:朱 錫 琴 Advisor: Dr. Hsi-Chin Chu 研 究 生:楊 筑 婷. 中 華 民 國 一零八年 一月 January 2019.

(2) 中文摘要. 本研究目的為二:(一)探討影響第二語言寫作成績的相關因素。(二)以路徑分 析法分析預測各變項和第二語言寫作成績的直接與間接的因果關係。 本研究旨在透過路徑分析,觀察第二語言寫作成績、動機以及第二語言學習者對於 教師寫作回饋的認知,了解這些變項之間的直接與間接的關係。本研究藉由文獻回顧, 先畫出對於這三個變項的假設路徑圖。之後,本研究收集 330 筆大學生問卷資料以及其 之英文寫作成績,來驗證此假設路徑圖。 路徑分析的結果顯示,「內在動機- 刺激」可以有效預測「第二語言學習者對於教 師寫作回饋在文法、語言問題上的認知」以及「第二語言學習者對於教師寫作回饋在架 構問題上的認知」。而「第二語言學習者對於教師寫作回饋在架構問題上的認知」可以 有效預測第二學習者的寫作成績。至於間接效果的部分,路徑分析的結果顯示,「第二 語言學習者對於教師寫作回饋在文法、語言問題上的認知」可以有效調節「內在動機- 刺 激」和「第二語言學習者對於教師寫作回饋在架構問題上的認知」 。除此之外, 「第二語 言學習者對於教師寫作回饋在架構問題上的認知」也可以有效調節「第二語言學習者對 於教師寫作回饋在文法、語言問題上的認知」和寫作成績。透過此模型,教師可以清楚 了解如何有效運用教師寫作回饋、動機來增強學生第二語言寫作的成績。. 關鍵字:路徑分析、內在動機、外在動機、第二語言寫作、教師寫作回饋. i .

(3) ABSTRACT. When compared to other language skills, most EFL learners consider writing the most challenging skill in second language learning (Maarof, Yamat & Li, 2011). To delve into the issue about how to teach L2 writing more efficiently, the current study aimed to examine the causal relationships among SDT-related variables, L2 learners’ perceptions of two different types of teacher WCF and writing performance. To this end, this study firstly drew a hypothesized path analysis model based on the rationales reviewed in the literature review. In addition, the valid data of 330 subjects adapted from Tsao (2018) were recruited to examine the model. The results revealed that intrinsic motivation had the largest predictive power on L2 learners’ perceptions of teacher WCF both about local and global issues and L2 learners’ perceptions of teacher WCF about global issues could significantly predict L2 learners’ writing scores. With regard to the mediating effects, intrinsic motivation toward stimulation could exert significant indirect effects on L2 learners’ perceptions of teacher WCF about global issues via L2 learners’ perceptions of teacher WCF about local issues. In addition, L2 learners’ perceptions of teacher WCF about local issues could also exert significant indirect effects on writing performance via L2 learners’ perceptions of teacher WCF about global issues. The importance of intrinsic motivation, L2 learners’ perceptions of teacher WCF regarding local and global issues in acquiring L2 writing were all highlighted in this study. In. ii .

(4) short, the results provided several insights for educators to teach L2 writing.. Keywords: path analysis; L2 writing; teacher WCF; intrinsic motivation; extrinsic motivation. iii.

(5) ACKNOWLEDGEMENT. I am deeply indebted to many people who helped me to complete this thesis. With their support and unconditional love, I have eventually accomplished my thesis. Thus, I would like to express my gratitude to those people. My earnest thanks first goes to my advisor, Dr. Hsi-Chin Chu. Whenever I encountered difficulties in writing my thesis, I would always turn to her for help. I can always be greatly enlightened by her to think the research questions from different perspectives, and learn to convey my thoughts in a more explicit way. In addition, Dr. Chu is always there to support me and give me advices when I was overwhelmed by my work and study. My earnest thanks also extend to Dr. Wen-Ta Tseng, who guided me to the knowledge of path analysis approach. With his superb and patient guidance, I gradually learned how to conduct this study and interpret the data of this study. From Dr, Tseng, I learned how to cope with the obstacles and think positively when encountering difficulties during the process of doing research. My wholehearted thanks go to my committee member Dr. Mei-Chen Wu, who gave me constructive advice and insights from a professional readers’ perspective. Those suggestions that I receive from her made me think deeper and more critically in writing the implication of the study.. iv.

(6) In addition, I also want to express my sincere appreciation to all the kind people giving me mental support and constructive advice when I was writing my thesis. They are my friends in TESOL program, including Cynthia, Ivy, Jessie, and William. I want to express my gratitude to them for their generous encouragement. Lastly, I owe a debt of gratitude to my family. They give me their unconditional love to support me through my study. Thank you all with my sincere heart.. v .

(7) TABLE OF CONTENTS 中文摘要 ................................................................................................................................... i ABSTRACT ............................................................................................................................. ii ACKNOWLEDGEMENT ....................................................................................................... iv LIST OF TABLES ................................................................................................................ vii LIST OF FIGURES ............................................................................................................... vii LIST OF APPENDICES ....................................................................................................... vii CHAPTER ONE ...................................................................................................................... 1 INTRODUCTION .................................................................................................................... 1 Background and Motivation of the Study ......................................................................... 1 Significance of the Study .................................................................................................. 7 CHAPTER TWO ...................................................................................................................... 9 LITERATURE REVIEW ......................................................................................................... 9 Introduction to SDT .......................................................................................................... 9 Intrinsic Motivation ........................................................................................................ 10 Extrinsic Motivation ....................................................................................................... 12 Empirical Studies on Motivation and Language Scores ................................................. 15 Teacher Written Corrective Feedback ............................................................................ 19 Empirical Studies about Written Corrective Feedback ................................................... 21 Learners’ Cognitive Processing Stages of Written Corrective Feedback ....................... 23 Individual differences and corrective feedback .............................................................. 25 L2 learners’ perceptions of two types of teacher WCF .................................................. 29 CHAPTER THREE ................................................................................................................ 31 METHODOLOGY ................................................................................................................. 31 Introduction .................................................................................................................... 31 Research Framework ...................................................................................................... 31 Participants ..................................................................................................................... 33 Instruments ..................................................................................................................... 33 Data Collection ............................................................................................................... 34 Data Analysis ................................................................................................................. 35 Path Analysis .................................................................................................................. 36 CHAPTER FOUR .................................................................................................................. 37 RESULTS ............................................................................................................................... 37 CHAPTER FIVE .................................................................................................................... 46 DISCUSSION ........................................................................................................................ 46 Pedagogical Implications ................................................................................................ 52 Limitations and Suggestions for Future Studies ............................................................. 55. vi.

(8) Conclusion ...................................................................................................................... 57 REFERENCE ......................................................................................................................... 58 APPENDICES ........................................................................................................................ 70 Appendix A .................................................................................................................... 70 Appendix B .................................................................................................................... 74 Appendix C .................................................................................................................... 78. LIST OF TABLES Table 1. Descriptive of all variables in the empirical path model .......................... 39 Table 2. Correlations between IMa, IMk, IMs, EMe, Emi, EMid, EMin, WCFLocal, WCFGlobal and WS .......................................................................................... 40 Table 3. The Goodness of Fit Indices of the empirical path model ........................ 41 Table 4. Standard Regression and Effect Size Estimates ....................................... 43 Table 5. Direct, Indirect, and Total Effects of the Outcome Model ....................... 45. LIST OF FIGURES Figure 1. A taxonomy of human motivation. ......................................................... 15 Figure 2. Hypothesized path model of this study. .................................................. 32 Figure 3. The empirical path analysis model. ........................................................ 43 . LIST OF APPENDICES Appendix. A The Self-regulated Learning Variables Questionnaires (English)….70 Appendix. B The Self-regulated Learning Variables Questionnaires (Chinese)…74 Appendix. C The Writing Proficiency Test ………………………………………78. vii.

(9) CHAPTER ONE. INTRODUCTION. Background and Motivation of the Study. When it comes to learning a second language, writing is one of the crucial abilities for learners to develop. Writing is not only a means of communication but also an important section in many English proficiency standardized tests. However, when compared to other language skills such as listening, speaking and reading, most EFL and ESL learners consider writing to be the most challenging skill (Maarof, Yamat & Li, 2011). Writing is a complicated process that requires learners to have accurate language use, adequate background knowledge about rhetorical organizations, and sufficient lexical knowledge (Kao & Reynolds, 2017; Tangpermpoon, 2008). What’s more, writing is also an arduous process because of its recursive nature (Nagin, 2003), which requires learners to repeat the similar process in different sequences (Tribble, 1996). Researchers suggest that learners can improve their overall writing qualities by implementing self-regulated learning (SRL) skills in their writing processes (Butler & Winne, 1995). SRL refers to that learners can “monitor, control, and regulate certain aspects of their own. cognition,. motivation,. and. behavior 1 . as. well. as. some. features. of. their.

(10) environments”(Pintrich, 2000). A number of studies about SRL have been conducted to prove that SRL can help learners deal with the difficulties during their writing processes and eventually improve their overall writing qualities (Bulter, 1995; Jun, 2012; Ruan, 2005). Jun (2012) delved into the writing processes of five adult Korean learners. Results showed that SRL skills did really help learners to enrich their lexical items and become more native-likeness in their writing. In addition, Ruan (2005) observed Chinese undergraduate learners’ learning journals from a SRL writing program and found out that SRL skills can lead to learners’ acquisition of metacognitive knowledge of L2 writing. Motivation is the key component of SRL (Schunk, 2005). In second language acquisition, Dörnyei (2001) pointed out that motivation is related to the orientation and the length of behavior, which means “motivation explains why people decide to do something, how hard they are going to pursue it and how long they are willing to sustain the activity” (2001, p. 8; italics in the original). The orientation of motivation concerns with the types of motivation, which specifically refers to the different reasons or goals behind actions (Ryan & Deci, 2000). On the basis of those different reasons and goals, Self-Determination Theory (SDT) specifies various types of motivation (Deci & Ryan, 1985; Deci & Ryan, 2000; Vallerand, 1997). The most two basic types of motivation are extrinsic motivation and intrinsic motivation. Extrinsic motivation refers to perform a behavior for the sake of a separable outcome (Deci & Ryan, 2000; Dörnyei, 2001). On the contrary, intrinsic motivation. 2 .

(11) refers to perform a behavior because of the inherent interests of the activities (Ryan & Deci, 2000). In addition, one of the basic assumptions of SDT is that motivation can be viewed as a continuum that ranges from self-determined (intrinsic) to controlled (extrinsic) types of motivation (Ryan & Deci, 2000). There are four types of motivation under the extrinsic motivation, including integration, identification, introjection and external regulation (Ryan & Deci, 2000). As for intrinsic motivation, Vallerand and other scholars (Noels, Pelletier, Clément & Vallerand, 2000; Vallerand, 1997) have pointed out that there are three subtypes of intrinsic motivation, which are IM toward know, IM towards accomplishments and IM towards stimulation. Moreover, many studies have been conducted to delve into the relationship between motivation and language scores in the field of second language acquisition (Bernaus, 1995; Binalet & Guerra, 2014; Chang & Lehman, 2002). A number of studies have proved that motivation has certain relationship with L2 learners’ listening scores (Harputlu & Ceylan, 2014; Larry, 2005; Liu, 2014; Rubin & Step, 2000; Zhang, 2015 ). Harputlu & Ceylan (2014) studied the motivation of 33 students from a Turkey university and their scores in a TOFEL’s listening section. Surprisingly, they found out that learners’ listening scores are highly related to their extrinsic motivation while having little relationship with intrinsic motivation. Besides, several studies also verified the relationship between motivation and L2 learners’ speaking competence (Chen, 2017; Khoiriyah, 2016; Tu & Zhou, 2015 ). Khoiriyah (2016) included 60. 3 .

(12) non-native major students from an Indonesia university as his participants. Those participants were further divided into extrovert and introvert students by a personality questionnaire. The participants’ attitude, motivation and speaking achievement were measured by a questionnaire and an oral test respectively (Khoiriyah, 2016). Significant correlation between motivation and speaking achievement was found both in introvert students and extrovert students. Similarly, researchers have also investigated the relationship between motivation and reading achievement in second language acquisition (Kondo-Brown, 2006; Dhanapala& Hirakawa, 2016; Mao, 2011; Ölmez, 2015). Dhanapala& Hirakawa (2016) assessed 406 undergraduate students’ motivation and reading comprehension. The results showed that both intrinsic and extrinsic motivation could predict reading comprehension in various ways. L2 learners’ intrinsic motivation was positively related (r=86) to their reading comprehension, while their extrinsic motivation was negatively related (r=-.37) to their reading comprehension. Whereas the similar results can also be found in other related studies, a study from Ölmez (2015) suggested different results. Ölmez (2015) studied the relationship between motivation and L2 reading achievement among 114 freshman students in a Turkey university. The results showed that there was no significant relationship between motivation and L2 reading achievement. Although a number of studies have been carried out to study the relationship between motivation and L2 learners’ listening, speaking and reading achievement, studies on the relationship between motivation and L2 writing scores are. 4 .

(13) relatively few. Some studies have assessed the relationship between motivation and L2 learners’ writing proficiency (Hashemian & Heidari, 2013; Zhang & Guo, 2012). However, they didn’t adopt the construct of SDT and didn’t further assess other variables between motivation and learners’ writing achievement. Thus, more studies are needed to fill these gaps. At the same time, learners’ perception of feedback is also another important component in SRL. Feedback refers to the advices that are given by oneself or other people, such as teachers, peers, parents, books and so on, concerning one’s performance (Hattie & Timperley, 2007). On the basis of the sources of feedback, feedback has two types- internal feedback and external feedback. Internal feedback comes from one’s self while external feedback comes from other people. In SLA, lots of studies have been conducted to delve into the effectiveness of different types of feedback on L2 learners’ writing competence. Among various types of feedback, teacher written corrective feedback (WCF) plays an important role in L2 writing (Agbayahoun, 2016; Hyland & Hyland, 2006). Teacher WCF is believed to be a crucial tool that can prompt L2 learners to revise their writing and eventually improve their writing skills (Agbayahoun, 2016). While most studies delved into types of teacher WCF that are more effective, they didn’t investigate how different types of teacher WCF relate to each other. Moreover, to maximize the effectiveness of feedback, it is necessary to take learners’ perceptions and preferences of feedback into consideration. Thus, this study classifies teacher. 5 .

(14) WCF into two types, which are feedback about local problems and feedback about global problems. The present study investigates L2 learners’ perceptions of those two types of feedback and observes the relationship between L2 learners’ perceptions of those two types of feedback. Taken together, it can be concluded that previous studies have yet to clarify the causal relationships between SDT-related variables, L2 learners’ perceptions of different types of teacher WCF and writing performance/achievement. The present study aims to fill this gap. The specific research questions of this study are as follows: 1. How do IM and EM impact differently on L2 learners’ perceptions of different types of written corrective feedback? 2. How do L2 learners’ perceptions of different types of written corrective feedback mediate the causal relationships between self-determination and L2 writing performance?. 6 .

(15) Significance of the Study. The significance of this study primarily focuses on the following two aspects. Firstly, this research mainly investigates the causal relationship among SDT-related variables, learners’ perceptions of different types of teacher WCF and writing achievement. In addition, this study will reveal specific causal relationships among these three variables. Accordingly, this research will present EFL teachers specific information about what types of motivation that learners’ have in L2 writing, how those various types of motivation impact differently on L2 learners’ perceptions of two types of teacher WCF and act through feedback to influence learners’ writing achievement. Then, EFL teachers will specifically know exactly how various types of motivation and L2 learners’ perceptions of teacher WCF can facilitate their writing achievement. Through an update of the current study, EFL teachers can establish course objectives, design activities and create the learning environments that can facilitate learners’ motivation to process teachers’ WCF and eventually enhance their writing achievement. In addition to the significance in EFL teaching context, this research also provides a new insight into investigating L2 learners’ writing achievement. This study clarifies the effects of L2 learners’ perceptions of different types of feedback on their writing achievement. At the same time, this study also clarifies the effects of different types of motivation on L2 learners’ perceptions of two types of teacher WCF. Besides, different types of motivation could also 7 .

(16) shift L2 learners’ perceptions of feedback. Thus, this study would provide a possible causal route among those three variables. In short, previous studies that delve into the causal relationship among SDT-related variables, L2 learners’ perceptions of teacher WCF and writing achievement are few. Thus, this study aims to fill this gap and provide a new insight into this research field.. 8 .

(17) CHAPTER TWO. LITERATURE REVIEW. Introduction to SDT. Self-regulated learning refers to “an active, constructive process whereby learners set goals for their learning and then attempt to monitor, regulate, and control their cognition, motivation, and behavior, guided and constrained by their goals and the contextual features in the environment” (Pintrich, 2000, p. 453). Thus, self-regulated learners actively engage in their learning process and then monitor, regulate and control their learning process. There are several crucial elements in SRL, such as learners’ attitudes, L2 self, motivation and so on. Motivation is the foremost element in SRL (Schunk, 2005). Motivation plays a key role in learners’ academic performance in classrooms (Corno & Rohrkemper, 1985; Pintrich & De Groot, 1990). In second language acquisition, Dörnyei (2001) pointed out that motivation would influence the orientation and the length of behavior, which means “motivation explains why people decide to do something, how hard they are going to pursue it and how long they are willing to sustain the activity” (2001, p. 8; italics in the original). Specifically, the orientation of motivation refers to various reasons and goals behind the behaviors (Dörnyei, 2001). In SDT, internalization refers to the process of assimilating in a value or. 9 .

(18) regulation (Deci & Ryan, 2000). Depending on how those values or regulations internalized, SDT specifies the classification of motivation (Deci & Ryan, 1985; Deci & Ryan, 2000; Dörnyei, 2001; Vallerand, 1997). In educational field, SDT concerns with how learners cultivate "an interest in learning, a valuing of education, and a confidence in their own capacities and attributes" (Deci, Vallerand, Pelletier & Ryan, 1991, p 325). There are several crucial components in SDT, including three psychological needs, teachers’ orientations, amotivation, intrinsic motivation, and extrinsic motivation (Tsao, 2012). Due to the focus of the present study, the topics about intrinsic motivation and extrinsic motivation will be further elucidated below.. Intrinsic Motivation. Intrinsic motivation is the most self-determined form of motivation in SDT (Noels, Clément & Pelletier, 2001). Intrinsic motivation refers to engage in an activity because of its inherent interests and satisfactions (Deci & Ryan, 2000). Operant theory (Skinner, 1953) argued that reward is an essential element in all behaviors. It is believed that the reward is in an intrinsically motivated activity itself. On the contrary, learning theory (Hull, 1943) believed that intrinsically motivated activities could satisfy people’s innate psychological needs (Deci & Ryan, 2000). Psychological needs, such as feelings of competence, autonomy, and relatedness, can be satisfied through intrinsically motivated activities (Deci, 1975; Deci. 10.

(19) & Ryan, 2000). Then, many researchers have delved into what conditions can facilitate and undermine people’s intrinsic motivation. SDT puts its focus on environmental and social factors that can enhance or diminish intrinsic motivation (Deci & Ryan, 2000). Deci and Ryan (1985) proposed Cognitive Evaluation Theory (CET), a subtheroy of SDT, to explain how social environmental factors can influence people’s intrinsic motivation. Social environment factors, such as communication, feedback and rewards, can somehow affect people’s intrinsic motivation (Deci & Ryan (1985). For example, factors that are beneficial to a sense of competence can facilitate intrinsic motivation because people’s some psychological needs can be satisfied by the activities. In addition, CET further stated that intrinsic motivation could be strengthened through the satisfaction of both feelings of autonomy and competence (Deci & Ryan, 2000). On the basis of SDT and CET, previous studies have confirmed that optimal challenges (Deci & Ryan, 1985, p. 123), positive performance feedback (e.g., Deci, 1971; Harackiewicz, 1979), autonomy-supportive environment (Deci & Ryan, 1985, 2000) can enhance intrinsic motivation. In contrast, tangible reward (Deci, Koestner & Ryan, 1998), threats (Deci & Cascio, 1972), deadlines (Amabile, DeJong, & Lepper, 1976), and competition pressure (Reeve & Deci, 1996) are believed to decrease intrinsic motivation. In short, both SDT and CET argued that environmental factors could either enhance or diminish intrinsic motivation through sustaining or obstructing the needs for autonomy and competence (Jones, Llacer-Arrastia &. 11.

(20) Newbill, 2009). As for the types of intrinsic motivation, Vallerand and his colleagues (Noels, Pelletier, Clément & Vallerand, 2000; Vallerand, 1997) have pointed out that there are three types of intrinsic motivation, including IM-Knowledge, IM-Accomplishment and IM-Stimulation. IM-Knowledge refers to do an activity for the pleasure of learning new knowledge. IM-Accomplishment refers to do an activity for the pleasure of achieving a goal or mastering a thing. IM-Stimulation refers to do an activity for the pleasure stimulated by doing an activity, such as aesthetic pleasure. The common basis of these three types of intrinsic motivation is the pleasure during the self-initiated and challenging activity.. Extrinsic Motivation. Extrinsic motivation refers to perform a behavior for the sake of a separable outcome instead of the inherent interests in the behavior (Deci & Ryan, 2000; Dörnyei, 2001). In other words, extrinsically motivated activities are done to attain instrumental goals, such as receiving extrinsic rewards, or escaping punishments. With the satisfaction of the psychological needs (i.e., needs for autonomy, competence, and relatedness), an individual could change from being more extrinsically motivated to being more intrinsically motivated (Deci & Ryan, 2000). Specifically, on the basis of the extent to which it is autonomous, extrinsic motivation varies greatly (Deci & Ryan, 2000). Organismic Integration Theory (OIT), a subtheory of SDT, specifies the different types of motivation, as shown in Figure 1. 12.

(21) (Deci & Ryan, 2000). The classification of those four types of extrinsic motivation can be thought of as a continuum, which is the active process of an individual’s internalization and integration (Deci & Ryan, 2000). Internalization refers to “the process of taking in a value or regulation, and integration is the process by which individuals more fully transform the regulation into their own” (Deci & Ryan’s, 2000, p. 60). If learners’ are more self-determined in doing an activity, it means that their extrinsically motivated behaviors are more internalized and integrated (Deci & Ryan, 2000). The detailed explanation of those four types of extrinsic motivation is elucidated in the following paragraphs. External regulation, right next to amotivation, is the least self-determined and autonomous form of extrinsic motivation. Those behaviors are done for the satisfaction of an external requirement or for getting an external reward. The second type of extrinsic motivation is introjected regulation. People perform such actions under some types of pressure so as to reduce guilt or achieve pride (Deci & Ryan, 2000; Noels, Clément & Pelletier, 2001). Although people internalize the pressure, it is not totally autonomous because people don’t act on their true choice. Identified regulation is a more self-determined and autonomous form of extrinsic motivation. People recognize the significance of an action and then have accepted the regulatory as his or her own. Lastly, the most self-determined form of extrinsic motivation is integrated regulation. In this stage, identified regulations would not only have been completely assimilated to the self, but also have been assimilated. 13.

(22) with people’s own values and needs (Deci & Ryan, 2000; Noels, Clément & Pelletier, 2001).. 14.

(23) Figure 1. A taxonomy of human motivation. Adapted from "Intrinsic and extrinsic motivations: Classic definitions and new directions," by R. M. Ryan and E. L. Deci, 2000, Contemporary Educational Psychology, 25, p. 61. Copyright 2000 by Academic Press.. Empirical Studies on Motivation and Language Scores. A growing number of studies have focused on the effects of various situational variables on intrinsic and extrinsic motivation in language learning. Noels, Clément, and Pelletier (2011) studied 59 French Canadian students of English and assessed the relationships between two types of motivation and the integrative orientation among those students. The results showed that the integrative orientation highly related to intrinsic motivation. In the sample of the study, those who learn English because of their desire to interact with English community had higher intrinsic motivation. In addition, integrative orientation was also positively related to other variables such as perceived autonomy, perceived competence and English achievement. Similar results were reported by Wang (2008). Wang (2008) analyzed two samples of non-English major students in a China university. Wang’s study assessed the participants’ intrinsic motivation, English achievement and two types of extrinsic motivation, including autonomous motivation and controlled motivation. Autonomous motivation “means that the individual has internalized the value and regulation of English learning (Wang, 2008, p.641)”. As for controlled motivation, it means that the individuals do a thing because of external rewards, such as passing the examination,. 15.

(24) graduation, receiving praise from teachers and so on. The results showed that an autonomous extrinsic motivation positively related to intrinsic motivation and English achievement. However, controlled extrinsic motivation would decrease autonomy and thus would negatively related to intrinsic motivation and English achievement. In short, while intrinsic motivation and autonomous learning are two important factors in language learning, extrinsic motivation also plays an important role in language learning because learners are still extrinsically motivated in certain learning contexts (Bidin, Jusoff, Aziz & Taniza, 2009). Undoubtedly, intrinsic motivation and extrinsic motivation would have impact on language learning and would affect the success or failure in second language acquisition (Noels, Clément & Pelletier 2011; Wang, 2008). To delve into the relationships among motivation and language learning achievement, the following paragraph will review a number of empirical studies on this issue in terms of learners’ English listening, speaking, reading and writing achievements in EFL learning context. The issue about the relationship between motivation and L2 learners’ listening scores have been studied for years (Harputlu & Ceylan, 2014; Larry, 2005; Liu, 2014; Rubin & Step, 2000; Zhang, 2015). Larry (2005) studied the relationships between three types of motivation (amotivation, extrinsic motivation and intrinsic motivation) and L2 learners’ listening scores. Fifty-seven adolescents who speak different first language took a French instruction course. The significant negative relationship (r = -.34) was found between those learners’ listening. 16.

(25) test scores and amotivation. This might because learners didn’t see the relationship between their actions and listening scores. Thus, they might form passive attitude toward the listening test. Moreover, passive attitude is neither a feature of autonomous learners (Littlewood 1996) nor a feature of self-regulated learners (Pintrich 1999; Wenden 1998, 2002). In addition, Larry’s (2005) study also provided another relatively surprising results that there was low or no significant relationship between extrinsic, intrinsic motivation and learners’ listening test scores. This might because there was a gap between their daily listening practice and the listening test in the study. Thus, even though learners have the desire to do well in the listening test, they might fail to do so. Similarly, there were a number of studies about the relationship between motivation and L2 reading achievement (Dhanapala& Hirakawa, 2016; Kondo-Brown, 2006; Mao, 2011; Ölmez, 2015; Salikin, Bin-Tahir, Kusumaningputri & Yuliandari, 2017). Ölmez (2015) collected the data from 114 college students in Turkey to assess the relationships between their motivation and their reading achievement. The result showed that there was no significant relationship between motivation and learners’ reading scores. However, this result is different from other researchers’ assertion that there was a close relationship between motivation and language learning achievement (Dörnyei, 1994, 1998, 2005; Oxford & Shearin, 1994). Ölmez (2015) attributed the unusual result to the nature of motivation test that his study used, which only assessed learners’ overall motivation in L2 reading. Ölmez’s. 17.

(26) (2015) study failed to take other variables, such as learners’ interaction with the reading materials and learners’ comprehension of the reading materials, into consideration and it might be the main reason that the result from Ölmez’s (2015) study is different from other studies. Besides, there are a number of studies investigating the relationship between motivation and L2 learners’ speaking achievement (Chen, 2017; Khoiriyah, 2016; Tu & Zhou, 2015). Chen (2017) studied 111 college students in China and delved into the relationship between their motivation and their pragmatic speech competence. Pragmatic competence in the study refers to “the ability to produce appropriate language and at the same time to interpret correctly what the speaker says according to the speech context” (Chen, 2017, p.1308). Thus, pragmatic competence mainly related to L2 learners’ speaking ability. There was significant relationship (r = .502*) between instrumental motivation and pragmatic competence. The result showed that instrumental motivation did enhance the development of pragmatic competence in EFL learning context. In addition, the significant relationship (r = .593**) was also found between integrative motivation and pragmatic competence, which indicated that learners could continuously take part in communication activity without external rewards. Compared with L2 learners’ listening, speaking and reading achievement, there are relatively few studies on the relationship between motivation and writing scores (Hashemian. 18.

(27) & Heidari, 2013 ; Zhang & Guo, 2012). Zhang & Guo (2012) investigated 43 freshman English major students and 23 sophomore English major students in a China university. Surprisingly, those two groups, freshman group and sophomore group, showed different results. In freshmen group, there was significant positive relationship (r = .334) between English writing proficiency and extrinsic motivation but not with intrinsic motivation. In sophomore group, there was no significant relationship between English writing proficiency and English writing motivation. It might because sophomore learners had already improved their writing skills a lot in the past or might because they were worried about another approaching important test at that time. Those two reasons might decrease their motivation in English writing. Although there are several studies on the relationship between motivation and L2 learners’ writing achievement, those studies didn’t specify what types of motivation that are related to L2 learners’ writing achievement. In addition, those studies didn’t assess the relationship between motivation and L2 learners’ perceptions of teacher CF.. Teacher Written Corrective Feedback. Feedback is regarded as a crucial activity in most EFL language teaching classrooms (Bitchener, 2008; Kepner, 1991). Feedback is also regarded as an indispensable requirement that should be offered for learners during the language learning process (Good & Brophy, 2000). Corrective feedback (CF), one type of feedback, is widely used in EFL language. 19.

(28) learning classrooms. Chaudron (1977) defined CF as “any reaction of the teacher which clearly transforms, disapprovingly refers to, or demands improvement of the learner utterance”(p.31). CF has always been regarded as an important tool for learners to develop their L2 writing skill in EFL learning context (Hyland & Hyland, 2006). The role of CF in the development of L2 writing can not only serve as an instruction to scaffold learners’ learning but also enhance learners’ motivation by letting them know how their production is (Hyland & Hyland, 2006; Sheen, 2011). Learners can receive CF in classrooms either from their teachers or from their peers. The focus of this study is teacher WCF. Teacher WCF can be defined as “any comments, questions or error correction written on students’ assignments” (Mark, 2009, p.34). There are various forms of teacher WCF such as error corrections, questions, phrases, criticisms and so on (Mark, 2009). Teacher WCF on L2 learners’ writing can be roughly divided into two major types. The first type of teacher WCF would address problems mainly on rhetorical-level (Berge, 1999), or global (Lundstrom & Baker, 2009). Specifically, this type of feedback would focus on the organization and content of learners’ writing (Sheen, 2011). The second type of teacher WCF would focus on problems on sentence-level (Paul er al., 2007) or local (Lundstrom & Baker, 2009). The second type of feedback refers to feedback on lexico-grammatical errors, such as spelling, punctuation, grammar and so on (Nassaji, 2017; Sheen, 2011). From learners’ perspective, they want teachers to provide CF not only on their grammatical errors but also on their writing structure. 20.

(29) issues (Hedgcock & Lefkowitz 1994, 1996). From teachers’ perspective, it is believed that giving learners WCF about their grammatical errors or about their writing structure issues is necessary for learners to develop their L2 writing skill (Hyland, 2003). Due to the importance of those two types of teacher WCF, this study also put focus on those two types of teacher WCF. In this study, feedback about global issues refers to the first type of feedback and feedback about local issues refers to the second type of feedback. In all, teachers’ objectives of providing CF are to enhance learners’ motivation (Good & Brophy, 2000), improve learners’ writing competence through decreasing errors in their writing (Alhosani, 2008) and making the organization of their writing more complete. However, whether WCF is effective or not is still a controversial question for researchers to debate. The following sections discuss the effectiveness of CF on L2 language learning, provide theoretical perspectives on CF, and then delve into the cognitive and affective factors that might affect the effectiveness of feedback.. Empirical Studies about Written Corrective Feedback. Whether WCF is effective or not has been a heatedly debated question for decades. Truscott (1996) argued that WCF about grammar correction should be abandoned because it has a number of harmful effects and it is unhelpful for learners. Moreover, students might also simplify their writing in order to avoid making errors (Kepner, 1991; Sheppard, 1992;. 21.

(30) Truscott; 1996, 2004). A number of empirical studies also reported that WCF might have negative effects on L2 learners’ writing performance (Kepner, 1991; Salteh & Sadeghi, 2012; Semke, 1984). For example, Semke (1984) pointed out that corrections couldn’t enhance learners’ writing accuracy and writing fluency. Instead, corrections would have negative impact on learners’ attitude (Semke, 1984). On the contrary, Ferris (1998) thought Truscott’s argument that “grammar correction has no place in writing courses and should be abandoned” (1996, p. 328) was too strong and premature because of the following reasons. Firstly, most L2 learners recognized the importance of receiving teachers’ feedback about grammar correction (Cohen, 1987; Ferris, 1995). Secondly, most research suggested that L2 learners could improve their writing accuracy over time with the help of teachers’ CF (Bitchener, Young & Cameron, 2005; Ferris & Roberts, 2001; Hartshorn & Evans, 2015). In addition to the empirical studies, several meta-analyses have been conducted to examine the effectiveness of CF in L2 leaners’ language learning (Kang & Han, 2015; Li, 2010; Mackey & Goo, 2007; Russell & Spada, 2006). Russell and Spada (2006) meta-analyzed 15 studies about oral and written CF. Their finding supported the arguments that WCF is effective on written performance and that WCF can also support L2 grammar learning. Similarly, Kang and Han (2015) meta-analyzed 21 primary studies and found out that WCF can lead to grammatical accuracy in L2 writing. Taken together, those evidence showed that WCF is a topic worth studying. Nevertheless, there is still a dearth of research examining L2 learners’. 22.

(31) perceptions of different types of feedback when assessing the effectiveness of various types of feedback. Thus, a focus in this study is to examine how L2 learners’ perceptions of different types of feedback would influence their writing performance.. Learners’ Cognitive Processing Stages of Written Corrective Feedback. There are a number of theories that are proposed to either support or oppose the effectiveness of WCF on L2 learners’ writing competence. To delve into the effectiveness of WCF, the following questions should be considered firstly. How do L2 learners process the WCF? Why some L2 learners can benefit from WCF while others can’t? Gass (1997) summarized several theories about cognitive processing and proposed five-stage framework to explain the cognitive processing of oral CF. However, the same stages are needed in the cognitive processing of WCF (Nassaji, 2017). “Notice input” is the first stage of the Gass (1997) framework. After learners receive WCF as input, they need to notice the gap between their errors and the WCF that they get (Sheen, 2011). To make this stage happen, Tomlin and Villa(1994) proposed three levels of attention that might be involved in this noticing process, which are alertness, orientation and detection. Alertness means that learners should have motivation to learn from the WCF when they receive the WCF (Tomlin & Villa, 1994). Orientation means that learners should put their focus on a certain type of information rather than other information (Tomlin & Villa, 1994). Detection refers to the process that learners. 23.

(32) attend to certain information for the further processing (Tomlin & Villa, 1994). Similarly, Schmidt’s (1990, 2001, 2012) Noticing Hypothesis also argued that noticing is an indispensable element for individuals to learn things because they only acquire the things that they attend to in the input. The second stage of the Gass (1997) framework is “comprehended input”. There are several factors that would affect the extent of learners’ understanding of input, such as the types of feedback, learners’ proficiency level, contextual factors and so on. The third stage of the Gass (1997) framework is “intake”, which means that learners should relate the feedback that they receive to their own knowledge. The fourth stage of the Gass (1997) framework is “integration”. In this stage, learners would form a hypothesis in their long-term memory about the feedback that they receive. With more and more related evidence received, the original hypothesis might either be confirmed or rejected. “Output” is the last stage of the Gass (1997) framework. This is a rather active stage for learners to test their original hypothesis. Learners might probably fail at some stages of the cognitive processing of WCF, so some of them might fail to benefit from the WCF that they receive. Nevertheless, Gass (1997) framework doesn’t consider other individual differences (ID), such as cognitive and affective factors, which might affect learners to reach each stage. Thus, the following section discusses how those factors would influence learners’ cognitive processing stages of the CF that they get.. 24.

(33) Individual differences and corrective feedback. Individual differences (ID) are assumed to have positive or negative influence on learners’ language learning process and their achievements in language learning (Dörnyei, 2014; Gardner & MacIntyre, 1992, 1993). Similarly, ID factors would also influence learners’ processing stages of WCF (Bitchener, 2017; Sheen, 2011). Specifically, cognitive factors such as language learning aptitude, as well as affective factors such as language anxiety, attitudes, and motivation, would also somehow influence learners’ processing stages of CF and eventually indirectly influence the effectiveness of CF and learners’ L2 achievements (Bitchener, 2017; Sheen, 2011). It is believed that language aptitude is a key cognitive factor that would influence language learning (Ellis, 2004). Language aptitude has three major abilities, including auditory ability, rote learning ability and language analytic ability (Carroll, 1973, 1981). High language analytic ability can help learners to be much more successful in attending to the WCF that they get (Bitchener, 2017). It is not only because learners with high language analytic ability are more likely to notice the gap between their errors and the feedback they get but also because they can relate the CF they get to their own knowledge (Bitchener, 2017; Robinson, 2001). Similarly, high language analytic ability can facilitate learners to reach the fist and third stages in Gass’s (1997) five-stage framework, which are “notice input” and “intake” respectively.. 25.

(34) Like cognitive factors, affective factors, including anxiety, attitude and motivation, are also important factors that should be considered. Whereas lots of studies have assessed the relationship between affective factors and L2 achievements (Faqeih, 2015; Rassaei, 2015; Sheen, 2008), the issue how affective factors would influence learners’ cognitive processing stages of CF and their final L2 achievement remain unexplored. MacIntyre (1991, p.5) defined language anxiety as “the apprehension experienced when a situation requires the use of a second language with which the individual is not fully proficient”. Studies about the relationship between language anxiety and L2 achievement can roughly be classified into three different results. Some studies argued that language anxiety has a debilitating effect on L2 achievement because it would increase learners’ affective filters (Cheng, Horwitz, & Schallert, 1999; Horwitz, 2000). Other studies pointed out that language anxiety plays a facilitative role in L2 learning because learners might feel the tasks are challenging (Kleinmann, 1978). Still other studies argued that language anxiety has no effect on L2 achievement because L2 achievement is exclusively related to cognitive factors (Sparks & Ganschow, 1991). Lots of studies put their focus on the relationship between language anxiety and learners’ outcome instead of the learning process. Krashen (1985) is a pioneering in the field of linking anxiety to CF. Krashen (1985) disapproved of using CF in the class because it would increase learners’ affective filters and indirectly has negative impact on L2 achievement. In addition, a number of studies also showed that various types of CF would. 26.

(35) have different effects on anxiety and resulted in different L2 achievement (Jang, 2010; Sheen, 2008). Another important affective factor is attitude. A number of studies have shown that students have positive attitude toward teachers’ CF on their grammatical errors (Eninarlar, 1993; Schulz, 1996). Schulz (1996, 2001) pointed out that L2 learners hold positive attitude toward teachers’ error correction and thought teachers’ error correction was crucial in learning a second language. However, there might be a gap between what learners want and what can exactly assist them in second langue acquisition (Truscott, 1996). Moreover, few studies have investigated how learners’ perceptions toward CF could mediate learners’ learning outcome. In order to address the gap in how cognitive and affective factors influence learners’ processing stages of oral and WCF, Sheen (2011) investigated how three ID variables – language aptitude, anxiety and attitude – mediate the effect of oral and WCF. Sheen used an aptitude test that was originally designed by ISTVÁN OTTÓ. The test provided an artificial language and their English translations as clues for the participants to do the fourteen questions of multiple choice. Each multiple choice question contained an English sentence and the participants needed to choose the correct translation among the four choices provided. In addition to the language aptitude test, the participants were also required to do a questionnaire, which aimed to measure their anxiety and attitudes toward CF. There were a control group and four treatment groups in Sheen’s (2011) study. Those four treatment. 27.

(36) groups received four different types of CF, including oral recast group, oral metalinguistic correction group, written direct correction group and written metalinguistic correction group. Three major results could be concluded from Sheen’s (2011) study. Firstly, language analytic ability can assist leaners to benefit from the CF in all treatment groups except in the oral recast group. The reason that the oral recast group learners didn’t benefit from their language analytic ability in processing CF might because they were not aware of what should be analyzed after receiving CF. Secondly, the results also revealed that learner anxiety did somehow keep learners from benefiting from oral feedback. Nevertheless, learner anxiety didn’t keep learners from benefiting from WCF. It might because learners needed to bear the risk of embarrassing in front of the entire class, but this kind of risk didn’t exist in the WCF (Sheen, 2011). Thirdly, the results also suggested that learners with positive attitude towards CF could benefit much more from explicit CF than from implicit CF. The reason that explicit CF is much more effective than implicit CF for learners with positive attitude might because it is also important for learners to understand what is being corrected and be aware that they are being corrected. Although Sheen’s (2011) study delved into the mediating influence of one cognitive factor and two affective factors on the effect of four types of CF had on their learning of English articles, he didn’t consider another important affective factor- motivation. Thus, this study will investigate how different types of motivation would influence L2 learners’. 28.

(37) perceptions of teacher WCF and indirectly influence their writing performance.. L2 learners’ perceptions of two types of teacher WCF. Substantial research has been conducted to examine the effectiveness of teacher WCF on L2 writing. However, there are relatively few studies assessing L2 learners’ perceptions of teacher WCF (Diab, 2005) and their beliefs about what types of teacher WCF that they think are more constructive and helpful. Thus, this study aims to delve into L2 learners’ perceptions of two types of teacher WCF. One type of teacher WCF in this study puts its focus on global issues in L2 learners’ writing, such as content and organization. The other type of teacher WCF in this study puts its focus on local issues in L2 learners’ writing, such as spelling, punctuation, grammar and so on. Previous studies that delved into L2 learners’ perceptions of those two types of teacher WCF reported mix findings. Some studies pointed out that most L2 learners recognized the importance of teacher WCF about local issues in their writing (Enginarlar, 1993; Hajian, Farahani & Shirazi, 2014; Leki, 1991; Radecki & Swales, 1988; Saito, 1994; Simpson, 2006). A study from Radecki and Swales (1998) showed that L2 learners expected teachers to correct all their language errors or they wouldn’t believe in their teachers’ professional ability. Similarly, Leki’s (1991) study also pointed out that L2 learners expected teachers to correct all their language errors in their writing because they regarded error-free writing as a crucial. 29.

(38) criterion for the so-called good writing (Leki, 1991). Moreover, Enginarlar (1993) investigated 47 EFL learners’ attitudes toward teachers’ feedback on their writing and the results revealed that they viewed teachers’ feedback on the linguistic errors in their writing as an effective tool for them to improve their writing. Unlike those studies, Diab’s (2005) study showed that most L2 learners recognized the importance of feedback about their global issues in their writing, such as content and organization, rather than only on surface-level errors, such as spelling, and grammar. In addition to emphasizing either on local issues or global issues, still other studies suggested that L2 learners equate the importance of those two types of feedback (Straub, 1997). Straub (1997) studied 172 college students’ perceptions of those two types of feedback on their writing. The results indicated that students thought that feedback on global issues is as important as feedback on local issues. Although previous studies have investigated L2 learners’ perceptions of those two types of feedback, they didn’t further investigate what might be the relationship between learners’ perceptions of those two types of feedback and their writing motivation and the relationship between learners’ perceptions of those two types of feedback and their writing performance. Thus, this study aims to study the causal relationship among those three variables. At the same time, this study will also delve into how their perceptions of those two types of teacher WCF influence each other.. 30.

(39) CHAPTER THREE. METHODOLOGY. Introduction. This study investigates how different types of intrinsic and extrinsic motivation differently influence L2 learners’ perceptions of different types of teacher WCF. In addition, this study further delves into how do L2 learners’ perceptions of different types of WCF mediate the causal relationships between extrinsic and intrinsic motivation and L2 writing achievement.. Research Framework. On the basis of the rationales in the literature review and the research questions, Figure 2 is a hypothesized path model of this study. L2 learners’ writing performance (named WS) is hypothesized to be directly predicted by L2 learners’ perceptions of teacher CF about local issues (named WCFLocal in Figure 2) and L2 learners’ perceptions of teachers’ CF about global issues (named WCFGlobal in Figure 2). Three types of intrinsic motivation (named IMa, IMk and IMs in Figure2) and four types of extrinsic motivation (named EMe, Emi, EMid and EMin in Figure 2) are also hypothesized to exert an indirect effect on WS through WCFLocal. 31.

(40) and WCFGlobal. In addition, IMa, IMk, IMs, EMe, Emi, EMid, EMi are hypothesized to directly influence WCFLocal and WCFGlobal. Finally, it is hypothesized that WCFLocal influences WCFGlobal.. Figure 2. Hypothesized path analysis model of this study. Note: IMa= intrinsic motivation toward accomplishment. IMk= intrinsic motivation toward knowledge. IMs = intrinsic motivation toward stimulation. EMe= external regulation. Emi= introjected regulation. EMid= identified regulation. EMin= integrated regulation. WCFLocal = L2 learners’ perceptions toward teacher written corrective feedback about local issues. WCFGlobal = L2 learners’ perceptions toward teacher written corrective feedback about global issues. WS= writing scores.. 32.

(41) Participants. This study employed the data from 330 undergraduate students studying in four North Taiwan universities. All the participants had studied English at least six years because they were required to learn English since Grade 7. They also somehow developed their writing skills when they prepared for the college entrance exam in high school. After enrolling in universities, some of the participants even took some English courses that were either directly or indirectly related to English writing. Overall, all the participants were supposed to have adequate knowledge of English paragraph writing, which is a skill that this study aims to analyze.. Instruments. The questionnaire (A/B) adopted from Tsao (2018) was used to collect the data from the participants. A 6-point Likert scale (1= strongly disagree; 2= disagree ; 3= somewhat disagree; 4= somewhat agree; 5= agree; 6= strongly agree) was employed in the questionnaire. The questionnaire has three parts to assess L2 learners’ personal background information, their motivation and their perceptions of teacher WCF. The first part is to get the information about the participants’ personal background knowledge. The second part is to assess L2 learners’ writing motivation, including extrinsic and intrinsic motivation.. 33.

(42) According to SDT, there are four subtypes under extrinsic motivation, including external regulation, introjected regulation, identified regulation and integrated regulation, (Ryan & Deci, 2000). Similarly, there are three subtypes under intrinsic motivation, which are IM toward accomplishments, IM towards knowledge and IM towards stimulation (Noels, Pelletier, Clément & Vallerand, 2000; Vallerand, 1997). The third part is to assess L2 learners’ perceptions of two types of teacher WCF. One type of teacher WCF is about L2 learners’ local problems in writing, such as spelling, punctuation, and grammar. The second type of teacher WCF is about their global issues in writing, such as the contents and the organization. Besides, the writing scores that this study used was also adopted from Tsao (2018). To assess the participants’ present English writing competence, they were asked to take an English paragraph writing examination. The raters of this English paragraph test were two experienced college lecturers. To ensure the level of agreement between those two raters, a scoring rubric was used and the Cronbach alpha coefficient was also calculated. After data collection is completed, IBM SPSS and Amos version 22 are used to analyze the data and draw the path analysis model.. Data Collection. The data adopted from Tsao (2018) was used in this study and the data collection steps are as follows. The data was collected from a questionnaire (Appendix A/B) and an English. 34.

(43) paragraph writing test (Appendix C). The questionnaire followed by the English paragraph writing test was administered to the participants during their regular class that lasts for 100 minutes for two consecutive sessions by their teachers. Comprehensive instructions for completing the questionnaire and the English paragraph writing test were given to the participants. For instance, they were aware that there was no so-called right or wrong answer. All they need to do was to read every item on the questionnaire carefully and response to each item on the basis of their personal experiences and perceptions of learning English writing. Besides, they were also informed that the scores in their English paragraph writing test would not be included in their final writing course scores for the semester.. Data Analysis. This study will firstly pilot the scales adopted from Tsao (2018) to check their reliability and validity. Reliability will be checked by Cronbach’s alpha and validity will be analyzed with factor analysis. Then, path analysis will be further implemented to examine the causal relationships between seven types of motivation, L2 learners’ perceptions of two types of teacher WCF and L2 learners’ writing performance.. 35.

(44) Path Analysis. Path analysis is an extension of multiple regression that can be used to assess direct and indirect causal relationships among variables in a complex hypothesized model or to examine which hypothesized models best fits the data (Schumacker & Lomax, 2010; Streiner, 2005). The relationships among variables are in the ways of the direct, indirect and total effects (Dattalo, 2013; Raykov & Marcoulides, 2006). This study firstly drew a hypothesized path model (Figure2) based on the rationales in the literature review. Then, path analysis will be implemented through IBM SPSS and Amos version 22 to examine the causal relationships between the endogenous variables and the exogenous variables. As shown in Figure 2, the exogenous variables consisted of IMa, IMk, IMs, EMe, Emi, EMid and EMin, and the endogenous variables included WCFLocal, WCFGlobal and WS. The multiple regression approach to path analysis requires this study to perform as many analyses as the number of endogenous variables in the hypothesized path model (Meyers, Gamst & Guarino, 2016). Thus, this study will perform the following three regression analyses. 1) WS will be predicted from WCFLocal and WCFGlobal. 2) WCFLocal will be predicted from IMa, IMk, IMs, EMe, Emi, EMid and EMin. 3) WCFGlobal will be predicted from IMa, IMk, IMs, EMe, Emi, EMid and EMin and WCFLocal.. 36.

(45) CHAPTER FOUR. RESULTS. This chapter first calculated Cronbach’s alpha of the nine scales in the questionnaire (Appendix A/B) of this study to ensure their internal consistency. Secondly, the inter-rater reliability was measured to ensure the rating agreement between two raters on writing scores of the English paragraph writing test (Appendix C). Next, this chapter displayed the correlation matrix model of intrinsic/extrinsic motivation, learners’ perceptions of two types of teacher corrective feedback and writing scores. After that, this chapter further showed the goodness of fits indices of the empirical path analysis model of this study. Finally, this chapter revealed the direct effects and mediating effects among the empirical path analysis model. Firstly, Cronbach’s alpha was used to examine the internal consistency among the nine scales in the questionnaire of this study. As shown in Table 1, the Cronbach’s alpha coefficients for the nine scales were all above .7, which indicated that the questionnaire implemented in this study was reliable (Cortina, 1993). Besides, inter-rater reliability was implemented to ensure the rating agreement between two raters on writing scores among the 330 subjects. The results showed that a high inter-rater reliability was achieved (Kappa= .95).. 37.

(46) Moreover, the extent to which the subjects responded to the statements of IMa, IMk, IMs, EMe, Emi, EMid, EMin, WCFLocal, WCFGlobal and WS was also shown in Table 1. Regarding the subjects’ extrinsic writing motivation, most of them were motivated to learn English writing by Emin (M= 19.72, SD= 3.02). As for the subjects’ intrinsic writing motivation, most of them were motivated to learn English writing by IMa (M = 16.08, SD = 3.79). The subjects were more willing to spend time processing WCFLocal (M = 12.91, SD = 2.79) than WCFGlobal (M = 8.83, SD =1.92), which is a finding similar to several previous studies (Radecki & Swales, 1988; Saito, 1994; Simpson, 2006). Concerning the subjects’ writing performance (M= 3.64, SD= .62), their ability to write was at an intermediate level.. 38.

(47) Table 1. Descriptive of all variables in the empirical path model. Variables IMa IMk IMs EMe EMi EMid EMin WCFLocal WCFGlobal WS. N 330 330 330 330 330 330 330 330 330 330. M 16.08 13.25 12.51 12.25 16.44 14.54 19.72 12.91 8.83 3.64. SD 3.79 2.73 2.63 2.92 3.84 2.51 3.02 2.79 1.92 0.62. Cronbach’s alpha .91 .85 .79 .74 .79 .84 .88 .90 .95 na. Note: IMa= intrinsic motivation toward accomplishment. IMk= intrinsic motivation toward knowledge. IMs = intrinsic motivation toward stimulation. EMe= external regulation. Emi= introjected regulation. EMid= identified regulation. EMin= integrated regulation. WCFLocal = L2 learners’ perceptions toward teacher written corrective feedback about local issues. WCFGlobal = L2 learners’ perceptions toward teacher written corrective feedback about global issues. WS= writing scores.. Table 2 clarifies the relationships between the ten variables that were included in the empirical path analysis model. As can be seen in Table 2, almost most correlation coefficients reached statistical significance. The relationship that showed the strongest correlation was between WCFLocal and WCFGlobal (r = .80, p < .001) and this r value is indicative of a large correlation (Cohen, 1988). In addition, other relationships that showed a moderate correlation could be found among the intrinsic motivation group, which were the relationships between IMk and IMa (r = .64, p < .001), IMs and IMa (r = . 66, p < .001) and IMs and IMk (r = .59, p < .001 ).. 39.

(48) However, still other variable pairs only showed a low level of correlation, including the relationships between WCFLocal and IMa (r = .44, p < .001), WCFLocal and IMk (r = .41, p < .001 ), WCFLocal and IMs (r = .49, p < .001 ), WCFLocal and EMid (r = .33, p < .001 ), WCFLocal and EMin (r = .26, p < .001 ), WCFGlobal and IMa (r = .40, p < .001), WCFGlobal and IMk (r = .43, p < .001 ), WCFGlobal and IMs (r = .49, p < .001 ), WCFGlobal and EMid (r = .37, p < .001 ), and WCFGlobal and EMin (r = .28, p < .001 ). Moreover, the relationships both between WS and WCFLocal, and WS and WCFGlobal showed a small level correlation, which were r = .21 (p < .001) and r = .29 (p < .001) respectively.. Table 2. Correlations between IMa, IMk, IMs, EMe, Emi, EMid, EMin, WCFLocal, WCFGlobal and WS. Note: IMa= intrinsic motivation toward accomplishment. IMk= intrinsic motivation toward knowledge. IMs = intrinsic motivation toward stimulation. EMe= external regulation. Emi= introjected regulation. EMid= identified regulation. EMin= integrated regulation. WCFLocal =. 40.

(49) L2 learners’ perceptions toward teacher written corrective feedback about local issues. WCFGlobal = L2 learners’ perceptions toward teacher written corrective feedback about global issues. WS= writing scores.. Moreover, IBM SPSS and Amos version 22 were used to evaluate the empirical path analysis model of this study. The chi-square assessing model fit, with a value of 9.598 (7, N= 330), p = .213, was not statistically significant. Therefore, this model appeared to be a good fit to the data. As shown in Table3, the obtained Root Mean Square Error of Approximation (RMSEA) value was 0.032 with a 90% confidence interval of .000 to .077. In addition, the normed fix index (NFI), and the comparative fit index (CFI) yielded values of 0.994 and 0.998 respectively. Besides, Tucker-Lewis index (TLI) was 0.989, incremental fit index (IFI) was 0.998, and relative fit index (RFI) was 0.96. According to the criterion of acceptable fit from Hooper, Coughlan, & Mullen (2008) and Byrne (1994) , all of these values far-exceeded their own cut-off threshold (.90). Therefore, it was indicated that the empirical path analysis model was an excellent fit to the data. Table 3. The Goodness of Fit Indices of the empirical path model Model fit indices Acceptable fit The model. χ2/df. CFI. TLI. <3. > .90. 1.371. .998. IFI. RFI. NFI. RMSEA. > .90. > .90. > .90. >.90. < .80. .989. .998. .994. .032. 41. .960.

(50) Figure 3 and Table 4 revealed the standardized causal path coefficients. Some of the paths were not statistically significant and some of the paths reached statistical significance The standardized regression weights of all types of extrinsic motivation (EMe, Emi, EMid, EMin) toward WCFLocal and WCFGlobal were not statistically significant. The coefficients of the path both from IMa and IMk to WCFGlobal were not statistically significant. In addition, the standardized regression weights of the path from WCFLocal to WS was also not significant, either. Considering the causal relationships examined in the empirical path analysis model, WCFGlobal is a significant predictor of WS with standardized regression weights of .35 (p < .001 , f2 = .14). At the same time, WCFLocal contributed significantly to WCFGlobal (β = .72, p < .001, f2 = 1.08). Moreover, WCFLocal could be significantly predicted by all three types of intrinsic motivation (IMa, IMk, IMs). The coefficients were β =.15 (p < .05, f2 =.02 ) for the path from IMa to WCFLocal, β = .11 (p < .05, f2 = .01) for the path from IMk to WCFLocal and β =.28 (p < .001, f2 = .09) for the path from IMs to WCFLocal. That is, IMs is a more crucial role in influencing WCFLocal when compared with IMa and IMk. Similarly, IMs is also a significant predictor of WCFGlobal with standardized regression weights of .09 (p < .05, f2 = .01). In addition, WCFGlobal is a significant predictor of WS with standardized regression weights of .35 (p < .001, f2 = .14).. 42.

(51) Figure 3. The empirical path analysis model.. Note: IMa= intrinsic motivation toward accomplishment. IMk= intrinsic motivation toward knowledge. IMs = intrinsic motivation toward stimulation. EMe= external regulation. Emi= introjected regulation. EMid= identified regulation. EMin= integrated regulation. WCFLocal = L2 learners’ perceptions toward teacher written corrective feedback about local issues. WCFGlobal = L2 learners’ perceptions toward teacher written corrective feedback about global issues. WS= writing scores.. Table 4. Standard Regression and Effect Size Estimates. Paths IMa → WCFLocal IMk → WCFLocal IMs → WCFLocal IMs → WCFGlobal WCFLocal →WCFGlobal WCFGlobal → WS. Estimate .15 .11 .28 .09 .72 .35. 43. ES (f2=R²/1-R²) .02 .01 .09 .01 1.08 .14.

(52) Note: IMa= intrinsic motivation toward accomplishment. IMk= intrinsic motivation toward knowledge. IMs = intrinsic motivation toward stimulation. EMe= external regulation. Emi= introjected regulation. EMid= identified regulation. EMin= integrated regulation. WCFLocal = L2 learners’ perceptions toward teacher written corrective feedback about local issues. WCFGlobal = L2 learners’ perceptions toward teacher written corrective feedback about global issues. WS= writing scores.. In addition to computing direct effects, Table 5 further showed the mediating effects of both WCFLocal and WCFGlobal. That is, this study further examined the mediating effects of the two mediators by using the Sobel test (Sobel, 1982). In path analysis, a mediating effect refers to the transmission of the effect of an independent variable on a dependent variable through one or more variables, which are termed as mediator variables (Edwards & Lambert, 2007). With the help of the Sobel test (Sobel, 1982), the results revealed that WCFGlobal significantly mediated the causal relationship from WCFLocal to WS. Specifically, WCFLocal exerts significant indirect effects on WS (β = .72 x .35 = .25, p < .001) via mediation by WCFGlobal. In addition, WCFLocal significantly mediated the causal relationship from IMs to WCFGlobal. Specifically, IMs exerts significant indirect effects on WCFLocal (β = .28 x .72 = .2, p < .001). In sum, the results firstly revealed the significant direct effects in the path analysis model in Table 4. For example, all three types of intrinsic motivation were all significant. 44.

參考文獻

相關文件

鳳溪廖萬石堂中學 楊引子老師,楊瑜老師

另外,語文科高中的寫作活動也很多元化,題材亦很生活化,有助提高學生對創作 的興趣。 (高中語文寫作題目舉隅,見附件三 附件三 附件三。 附件三 。 。) 。 ) ) ).. 附件三

杏壇中學學生會擬發起「綠色生活 『無塑』校園」行動,此舉

互相交流是改善作品的好方 法。不同人對同一題材有不 同的處理,聽過別人的看法 (新輸入)後會豐富自己的想 法。.

討論結束,整理腦圖。首先嘗試將資料歸類,然 後可以開始收窄範圍,定出文章中心,再按照重

初中考核心篇章 額外字體分數 額外寫作大綱分數

% Urban university professors 93 Skilled printers 52 Mathematicians 91 Paper workers 42 Physicists 89 Skilled auto workers 41 Biologists 89 Skilled steelworkers 41 Chemists 86

大專界年度盛事「香江大專辯論賽」即將來臨,本屆比賽繼續秉承「身在校園,心繫社