即時線上討論對於文本理解與單字學習之影響

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. 即時線上討論對於文本理解與單字學習之影響. Enhancing Text Comprehension and Learning of Word Forms through Online Discussion: A Case Study of Focus on Form Intervention. 指導教授：劉宇挺博士 Advisor: Yeu-Ting Tony Liu 研究生：李怡馨 Yi-Hsin Ava Lee. 中華民國一○四年六月 June 2015.

(2) 摘要隨著科技產品如智慧型手機及平板電腦的普及，電腦輔助溝通(CMC)近來逐漸盛行。為了解是否可能將 CMC 融入第二語言教學課堂，須先研究其對第二外語學習所帶來的影響。據此，本研究採取 focus on form (FonF)教學的角度探究線上討論對於第二外語單字學習及文本理解的效果，並與個別習作互做比較。本研究包含 82 位來自臺北市一所女校的受試者，分為實驗組(EG)和對照組 (CG)，分別給予線上討論與個別習作的學習情境。在進行十次的學習之前，受試者先進行單字前測。每一次學習進行之後，受試者進行單字立即後測與該週文章的閱讀理解測驗，一週後進行單字延遲後測。每週五十分鐘的學習由七分鐘的個別閱讀開始，之後研究者進行問題詢問，目的在於引導目標字彙的使用。在線上學習情境中，受試者透過線上論壇與組員討論問題；而在個別習作情境中，受試者自行寫下答案後再與他人交換答案卷。在兩組別中，受試者各花十分鐘回應針對同一目標單字的所有問題。資料分析包括變異數分析(ANOVA)及歷時性資料變化分析。組內比較中，單字前後測上有顯著差異，顯示兩種學習情境皆有助於單字學習。針對單字後測平均分數做歷時性資料變化分析，則可見在總次數十次中，EG 的表現等於或高於 CG 高達五次。若將 EG 受試者在多工的學習情形下所需負擔的高度認知負荷量納入考慮，可知 EG 受試者在單字學習上的表現已超乎期待，也因而可推測線上討論應能對單字學習帶來正向影響。同時，針對閱讀理解測驗的平均分數做歷時性資料變化分析，也可得知就本研究中所探討的三個閱讀面向(總體理解、局部理解與推論)而言，EG 受試者表現優於 CG 受試者約七至九次，其中變異數分析組間比較的顯著差異也證明了線上討論對總體理解的正向學習效果。此學習效果可能來自於 CMC 營造了較為友善的環境，讓學習者能與其他組員分享看法，而引發意義協商。藉由互相交換訊息，學習者們能增進對文本的理解。. i.

(3) 本研究探討了將線上討論做為 FonF 學習方法的可能性，並認為線上討論有助於單字學習與文本理解。同時，本研究也嘗試將歷時性資料變化分析做為分析研究資料的一種方法。. 關鍵字：線上討論、單字學習、文本理解、動態機制理論. ii.

(4) Abstract With the heavy use of technological gadgets such as smartphones or tablet computers, computer-mediated communication (CMC) has been prevalent nowadays. To explore the possibility to involve CMC in L2 classrooms, it is necessary to investigate the effects it brings to L2 learning. This study adopts a FonF viewpoint to examine the effects of online discussion on L2 word form (WF) learning and text comprehension, compared with the effects of individual work. The participants were 82 tenth graders in a girls’ high school in Taipei, divided into the experimental group (EG) and the control group (CG). The two groups were respectively allocated with the learning conditions of online discussion and individual work. Prior to the ten-time intervention, a word form (WF) pretest was conducted. Each time after the intervention, the participants took a WF immediate posttest and a comprehension test based on the assigned reading, followed by a WF delayed posttest one week later. The weekly fifty-minute intervention began with a seven-minute silent reading. Afterwards, some text-based questions were asked for the purpose of eliciting the target words. In the online discussion condition, the participants discussed the given questions with their group members through an online forum, while in the individual condition, the participants wrote down their answers to the questions and then exchanged the answer papers. In both groups, the series of questions focusing on one target word took ten minutes. Data analysis included ANOVA and data variation analysis. ANOVA revealed significant differences in within-group comparisons of the collapsed mean scores of WF tests, indicating that both conditions facilitate the learning of word forms. The analysis of variations in diachronic mean scores of WF posttests shows EG performed iii.

(5) not worse than CG for five out of ten times. Considering the heavier cognitive load resulting from dual-tasking of EG’s treatment, EG participants’ performance in word form learning actually exceeded expectations, thus implying that online discussion was more effective in helping word form learning when compared with individual work. Also, the analysis of diachronic data development in the mean scores of comprehension tests reveals EG’s better performance than CG for seven to nine times in the three types of comprehension (i.e. global, local comprehension and inferencing), among which global comprehension is found to benefit most from online discussion condition, as having been proved by the significant difference of between-group comparisons of the collapsed mean scores. The possible explanation is that CMC might elicit more negotiation of meanings since it creates a more friendly environment for learners to comfortably share ideas with other group members. By giving responses and reading others’, learners can develop a better understanding of the texts. This study explores the possibility to involve online discussion as one option in FonF intervention, and suggests that online discussion might be able to facilitate learners’ word form learning and text comprehension. What’s more, this study provides an example to complete research results with analysis of diachronic data development.. Key words: online discussion, word form learning, text comprehension, Dynamic Systems Theory. iv.

(6) Acknowledgements This thesis would not have been completed without the help of many people, among whom Dr. Yeu-Ting Tony Liu is undoubtedly the one I am indebted to the most. I greatly thank him for his professional guidance through the process of my thesis writing. He also set a fine example of a diligent and conscientious researcher himself. Having had high expectations of me, he encouraged me to challenge myself. With his emotional support, I was able to overcome this greatest obstacle in my life, and have confidence that I am prepared for others yet to come. I am also grateful to my committee members, Dr. Yi-Chien Joyce Lee and Dr. Mei-Han Chen, both of whom professionally indicated some problems with the original version of this thesis, helping a lot to perfect this work. Also, this study wouldn’t have been possible without the cooperation of my participants at a girls’ high school in Taipei: The tenth-grade students of Class Shan and Class Sheng in the year of 2015. It is a pleasant to know that the students became fascinated with English reading because of the intervention. I wish to thank all my colleague friends. My special thanks go to Yi-Xuan, who patiently helped to solve my problems with the SPSS software; to Pei-Yin , who was by my side when I was weary of my thesis writing and teaching job. Her comforting words soothed my pain. Finally, I would like to thank my family for supporting me in this difficult task. My gratitude especially goes to my significant other, Steve Wu. He cares about my thesis as if it were his. He believed in my capacity to have this task accomplished at moments of my self-doubt, and patiently kept me company until the end.. v.

(7) Table of Contents Abstract ......................................................................................................................... iii Table of Contents .......................................................................................................... vi List of Figures ............................................................................................................... ix Chapter One: Introduction ............................................................................................. 1 1.1 Background and Motivation ............................................................................ 1 1.2 Rationale of the Study ...................................................................................... 2 1.3 The Purpose of the Study ................................................................................. 4 Chapter Two: Literature Review .................................................................................... 5 2.1 What Is “Focus on Form?” .............................................................................. 5 2.2 What Form to Focus? ....................................................................................... 6 2.3 How Can FonF Intervention Facilitate Learning of Language Forms? ........... 9 2.4 How to Record And Analyze Learners’ Attention to Target Forms? ............. 16 2.5 What Is the Research Purpose? ...................................................................... 17 Chapter Three: Methodology ....................................................................................... 21 3.1 Participants ..................................................................................................... 21 3.2 Instruments ..................................................................................................... 22 3.3 Materials ........................................................................................................ 24 3.4 Procedure ....................................................................................................... 26 3.5 Data Analysis ................................................................................................. 31 Chapter Four: Results and Discussion ......................................................................... 33 4.1 Introduction of Data Analysis Methods ......................................................... 33 4.2 Results: Effects of FonF Intervention on Learning of Word Forms .............. 36 4.3 Results: Effects of FonF Intervention on Text Comprehension ..................... 40 4.4 Discussion ...................................................................................................... 43 vi.

(8) Chapter Five: Conclusions ........................................................................................... 50 5.1 Conclusions of the Findings........................................................................... 50 5.2 Implications.................................................................................................... 52 5.3 Limitations ..................................................................................................... 55 5.4 Suggestions for Future Studies ...................................................................... 56 References .................................................................................................................... 58 Appendix A .................................................................................................................. 64 Appendix B .................................................................................................................. 67 Appendix C .................................................................................................................. 71. vii.

(9) List of Tables Table 1. Schedule of the Conduct of the Intervention and the Tests ............................ 27 Table 2. The Procedure of the Study ............................................................................ 31 Table 3. Descriptive Statistics of All the WF Tests Over the Ten Times ..................... 36 Table 4. Within-Group Comparisons of the Mean Scores of All the Pre- and Post-Intervention WF Tests .......................................................................................... 37 Table 5. Between-Group Comparisons of the Mean Scores of All the WF Posttests .. 38 Table 6. Descriptive Statistics of All the Comprehension Tests in Terms of the Three Different Types............................................................................................................. 41 Table 7. The Results of One-way ANOVA of All the Comprehension Tests in Terms of the Three Different Types........................................................................................ 42 Table 8. Results of Familiarity Check and WF Posttests of Four of the Target Words .... 46. viii.

(10) List of Figures Figure 1. The Process of the Teacher’s Elicitation of Target Word Forms from the EG Participants ................................................................................................................... 29 Figure 2. The Mean Scores of All the WF Tests of CG and EG .................................. 38 Figure 3. The Diachronic WF Immediate Posttest Mean Scores of EG and CG ......... 39 Figure 4. The Diachronic WF Delayed Posttests Mean Scores of EG and CG ........... 40 Figure 5. The Diachronic Mean Scores of EG’s and CG’s Global Comprehension Tests ............................................................................................................................. 42 Figure 6. The Diachronic Mean Scores of EG’s and CG’s Local Comprehension Tests ...................................................................................................................................... 43 Figure 7. The Diachronic Mean Scores of EG’s and CG’s Inferencing Tests.............. 43 Figure 8. Two Examples of Negotiation of Meanings Through Online Discussion in Two EG Groups ........................................................................................................... 49. ix.

(11) Chapter One Introduction Reading plays a vital role in expanding vocabulary and learning writing skills, hence fostering the deeply-rooted belief that developing reading skills should be prioritized when it comes to L2 learning. Crucial as it is, however, English reading discourages many EFL learners, who have limited language input outside of class. Many EFL learners have neither pleasant experience of English reading nor strong motivation for doing it. And then, their lack of abundant English reading hinders their language learning, thus forming a vicious circle. To help students with text comprehension and to take delights in reading, EFL teachers design a variety of classroom activities in English class, group discussion being one of the most commonly used techniques and proved to be beneficial in facilitating reading comprehension. For example, group discussion is likely to promote dialogic discourse (Kucan & Beck, 2003). Also, comprehension problems are likely to emerge during discussions; the further input from group members can be modified accordingly until the texts understanding is achieved (Van den Branden, 2000). With technological advances, discussion takes various forms, online discussion being one of them. As effects of discussion has been thoroughly researched, what effect online discussion as an innovative discussion forms can make on reading comprehension is intriguing and worth investigating. 1.1 Background and Motivation The introduction of computer-mediated communication (CMC) as a new form of discussion has been embraced by young learners. As “digital natives”, the younger generation shows little concern for sharing information online (Toetenel, 2014; Wang, Woo, Quek, Yang, & Liu, 2012). When learners enjoy the discussion environments.

(12) provided by CMC tools, there would be more interactions among learners online (Birch & Volkov, 2007; Kamhi-Stein, 2000; Tower, Latimer, & Hewitt, 2013). Some learners using CMC tools said that through online discussion, they could hear different voices and opinions from their peers, under which circumstances negotiation of meaning tends to occur (Carico, & Logan, 2004). Social network sites (SNSs) become prevalent among the younger generations as they provide platforms where CMC takes place (Insightxploer, 2012). About 50% of Taiwanese teens log on to SNSs every day (Pollster, 2011; Zhang, 2012). As SNSs become inseparable parts of students’ everyday lives in Taiwan, it is worthy of further investigation into their possible effects on students’ learning if they are skillfully applied. Some previous research has explored the possibility to apply CMC tools to English learning. It was confirmed by some that applying SNSs is effective in motivating L2 learning as students can contact peers through them (Baran, 2010; Promnitz-Hayashi, 2011). L2 learning through SNSs can increase learners’ confidence, help students foster positive attitudes (Kabilan, Ahmad, & Abidin, 2010) and facilitate learner autonomy (Promnitz-Hayashi, 2011). Learners were found to be more active in discussion through SNSs (Zhang, 2012), and many agreed that SNSs can create an environment for English learning (Kabilan et al., 2010). 1.2 Rationale of the Study Its effects on L2 learning reported by many previous studies, CMC, as an innovative discussion form, is often favored over traditional classroom discussion. This research is motivated by such preference, based on CMC’s capacity to create friendly discussion environments where learners’ anxiety is reduced and student-centered learning may occur, as specified below. Discussion through CMC alleviates learners’ anxiety over expressing themselves 2.

(13) (Carico & Logan, 2004; Tower et al., 2013), as CMC provides a friendly environment where learners share ideas with a new language identity and get free from others’ stares. Some introverted learners might thus benefit, finding it easier to “speak up” through CMC than in traditional classrooms. Another reason for learners’ lower anxiety in CMC is that it provides more time flexibility in giving responses than face-to-face discussion in traditional classroom. Learners need not to give real-time responses; instead, they are allowed to make more preparations until feeling ready for idea sharing. Online discussion through CMC tools is also preferred for it encourages student-centered discussion (Carico & Logan, 2004; Promnitz-Hayashi, 2011), thus providing optimal conditions for focus on form (FonF) intervention. As conversations in CMC were often steered by learners based on their interests and needs, learners’ self-paced and self-regulated language learning is facilitated (Lai & Gu, 2011). Such student centeredness in class, leading to teachers’ teaching being minimized, creates an ideal environment for FonF instruction. FonF emphasizes teachers’ guidance on directing learners’ attention, shifting it from meaning making to language forms at appropriate time, which teachers are more capable of doing when their roles change from controllers to facilitators in class. To sum up, student-centered learning environment created by CMC enables teachers to devote more energies to monitoring students’ language use and doing the job of attention shifting when they need not to be involved too much in discussions among students. As justified above, the reasons for preference of CMC as a discussion form are that it creates a discussion environment that is both speaker-friendly and suitable for FonF intervention. Learners feel more comfortable expressing themselves with a new language identity and more time flexibility in CMC, while teachers are able to focus more on monitoring language use instead of leading discussions themselves. 3.

(14) 1.3 The Purpose of the Study The study aims to investigate the effects that online discussion makes on the learning of language forms and learners’ comprehension of assigned texts through online discussion via SNSs. Online discussion is conducted with the purpose of eliciting learners’ focus on form. During the online discussion, learners are expected to be aware of their language forms that are being used while they are trying to communicate with other group members. Can learners’ focus on form facilitate their learning of the target forms? Can learners’ reading comprehension be enhanced during the process? Searching for answers to these research questions, the study attempts to confirm the effects online discussion brings to learning of forms and text comprehension, and explores the possibility of involving online discussion as one teaching technique option in FonF instruction as well.. 4.

(15) Chapter Two Literature Review To provide relevant background information for the current study, existing research on FonF will be briefly reviewed in this chapter. The review will be divided into five subsections that deal with the following five questions: 1) What is “focus on form (FonF)?” 2) What form to focus? 3) How can FonF instruction facilitate language acquisition? 4) How to record and analyze learners’ attention to target forms? 5) What is the research purpose? 2.1 What Is “Focus on Form?” First introduced by Long (1991), the term “focus on form (FonF)” refers to how attention is drawn to linguistic forms during the process of meaning making. Focus on form “overtly draws students’ attention to linguistic elements as they arise incidentally in lessons whose overriding focus is on meaning or communication” (pp.45-46). Learners primarily engage in meaning making and communication, and pay attention to linguistic features when necessary. FonF was introduced in response to problems with “focus on formS” (FonFs) and “focus on meaning” (FonM). These two poles of stances view how and whether learners should attend to linguistic forms differently. Doughty and Williams (1998) pointed out that FonFs confines learners’ attention merely to formal aspects of language without communicative contexts, whereas FonM excludes almost all such focus on formal aspects of language. FonF, however, stresses the need to unobtrusively draw learners’ attention to linguistic forms under rich communicative contexts. Through this way, FonF on the one hand attempts to address the criticisms about the discrete linguistic learning involved in FonFs instruction. On the other hand, FonF provides alternatives to FonM by solving two of its serious problems. One is 5.

(16) that linguistic accuracy in learners’ output is often neglected for lack of attention to form in FonM, and the other is that certain grammatical forms cannot be learned only through communication, which is highly emphasized in FonM (Long, & Robinson, 1998). 2.2 What Form to Focus? This section talks about what form would be an ideal candidate for FonF instruction; that is, what kind of linguistic forms can be learned more effectively in FonF instruction. For this purpose, features of such forms will be discussed, followed by the rationale regarding why lexical items are considered as an ideal candidate for the FonF intervention in this study. 2.2.1 Features of forms as candidates for FonF intervention The features that may determine whether a form is an ideal candidate for FonF intervention are as follows: 1) Salience of a form, or the degree that learners can notice the form in communication (Harley, 1993); 2) transparency of a form, or lack of formal or functional complexity; and 3) length of a form and nature of its rule. All of them will be further elaborated individually. Salience of a form Learners may find a form more salient when they have already had partial mastery of the form (Williams & Evans, 1998), under which circumstance learners’ attention can be more effectively directed to the form. Compared to externally-created salience, generated by teachers or FonF teaching methods, internally-created salience, as made possible through learners’ cognitive readiness for a target form, contributes more to learning of a certain form (Park, 2005). Park (2005) clearly indicated the vital role played by learners’ cognitive readiness in learning a target form. In her study, when the participants had produced the target form in the pretest, which was a sign of learners’ partial knowledge before 6.

(17) the intervention, they tended to produce the target form more in the posttest. This result was observed in both the experimental group and the control group. Learners’ readiness of a target form seemed to be able to predict the learning effects of the form. It is thus advised that target forms which match learner readiness be selected for FonF instruction (Han, Park, & Combs, 2008). To include more salient forms in FonF intervention, teachers need to take learners’ developmental readiness into consideration Transparency of a form A transparent form is lacking in formal or functional complexity. It might be part of Universal Grammar, and its contrast with learners’ L1 is so unapparent that little negative evidence, such as error corrections, is required (DeKeyser, 1998). To help explain transparency of forms, VanPatten (2005) introduced “communicative value,” which is determined by two factors: inherent semantic value and redundancy. More transparent linguistic forms are those with higher communicative value, or according to VanPatten, high in inherent semantic value but not redundant; that is, the information conveyed by the forms are essential for communication but cannot be retrieved elsewhere except from the form itself. For example, the use of progressive form of a verb denotes an action that is happening, but other than the suffix “-ing,” few linguistic items provide similar clues. In this case, the progressive form, with high inherent semantic value but low redundancy, has high communicative value. Learners are more likely to proactively notice transparent forms with high communicative value, which explains why such forms may be great candidates for FonF instruction. Transparent forms are likely to enable learners to notice the gap between the target forms and relevant formal knowledge registered in their interlanguage system, and thus to better able to attend to the form to overcome 7.

(18) communication breakdown. When learners are initiators of production of certain forms, their acceptance of the form instruction becomes higher than in a learning context where they are passively presented with the forms (Lightbown, 1998). Length of a form and nature of its rule The length of a form and nature of its rules are also worth considering for their importance in attracting learners’ attention and facilitating learning of the form. Park (2005) investigated learners’ acquisition of backshifting of reported speech and suggested that target forms should be shorter in length, especially when presented in enhanced input, so as to attract learners’ attention and avoid increasing cognitive load. Also, after analyzing the rules of backshifting of reported speech, she pointed out that backshifting of reported speech would not be a perfect target form to include in FonF instruction because of its numerous exceptions in actual use. Therefore, if FonF intervention is to maximize its effects on learning forms, teachers or researchers need to be careful when choosing target forms that are appropriate in length and simple in rules so that learners would be able to notice them and work out the forms themselves. 2.2.2 Rationale for using FonF intervention to familiarize learners with forms of unfamiliar vocabulary Forms of unfamiliar vocabulary can be involved in FonF intervention. Such possibility has been raised by Doughty and Williams (1998), who suggested that “form” includes “all the levels and components of the complex system that is language” (p. 212). It can be argued that lexical items are also one ideal candidate for FonF instruction, based on the review of features of ideal forms for FonF intervention in the previous section. The arguments are presented below. First, as lexical items are meaning units in communication, they are salient for learners, often easily noticed by them. Second, lexical items have high “communicative value” (VanPatten, 2005); 8.

(19) forms of lexical items are high in inherent semantic value but low in redundancy. Also, as single words or phrases, lexical items are often short enough to easily catch learners’ attention. When forms of words can be regarded as one candidate for FonF intervention, whether they can be learned effectively is largely determined by whether FonF intervention brings optimal learning environment. To explore this issue, the next section includes theoretical premises concerning how learners’ attention can be successfully directed to forms as well as discussions on methodological problems about implementation of FonF intervention. 2.3 How Can FonF Intervention Facilitate Learning of Language Forms? Learning effects of linguistic forms in FonF intervention are determined by 1) interactive tasks that are involved; 2) communication breakdown; and 3) “task-essentialness” of a target form (Loschky & Bley-Vroman, 1993, as cited in Doughty & Williams, 1998, p.209), which refers to the necessity of using a target form to complete a task. The three factors, elaborated in 3.1 Theoretical premises, are influential in learning effects of linguistic forms because they decide whether learners’ attention would be effectively directed to forms in interaction. As for 3.2, more discussions on optimal implementation environment for FonF intervention will be provided. 2.3.1 Theoretical premises Interactive tasks Motivated by Interaction Hypothesis (Long, 1983), FonF emphasizes the importance of interaction in learning. Interactions between learners are crucial in learning of linguistic forms, especially forms of vocabulary. Ellis, Tanaka, and Yamazaki (1994) examined learning effects on word meanings brought by different learning conditions, and found interactionally modified input condition was more 9.

(20) effective in helping comprehension and learning of L2 words than conditions of premodified and unmodified input. In their study, the learners voluntarily interacted with teachers to solve their comprehension problems of L2 words so as to successfully perform the listening tasks. This insightful study provided the foundation for lexis learning through communicative tasks, especially interactive ones. Interactive tasks were proved useful and highly recommended in word learning based on the following reasons (Nation and Newton, 1993): First, meaningful contexts in interactive tasks offer adequate clues for word guessing. Second, opportunities of repeated use of the new items are available in interactive tasks. Zimmerman (1997) has researched into the learning effects on L2 words brought by instruction that combined a communicative method with reading. She found that the participants receiving this instruction outperformed their counterparts in the control group, proving the benefits of communicative activities on learning. Also, the questionnaires showed that the learners viewed communicative activities as useful for vocabulary learning. It was commented in this study that communicative activities were beneficial for facilitating learners’ word learning, and that they are also considered useful by learners. Communication breakdown FonF researchers agree with the idea of learning L2 through interaction, but further indicate that learners’ attention should be occasionally and unobtrusively shifted to linguistic forms when comprehension problems arise and communication breaks down. The switch in attentional focus was clearly clarified in definition of FonF provided by Long and Robinson (1998): “Focus on form often consists of an occasional shift of attention to linguistic code features –by the teacher and/or one or more students –triggered by perceived problems with comprehension or production” (p.23). When communication breaks down, learners are likely to be aware of the 10.

(21) mismatches between the target language and their interlanguage, their attention thus shifting to forms. Communication breakdown has been asserted to be one of the best triggers for learners to shift their attention to form in FonF instruction (Long & Robinson, 1998). Difficulty in comprehension or communication breakdown is in fact more likely to result from failure to understand words than grammatical errors (Laufer, 2006), as words serve as basic units that convey meanings in communication (Oxford & Scarcella, 1994). Learners’ failure to recognize word forms would easily cause difficulty in comprehension and the meaning making later on, which would drive the learners to pay attention to the forms. It has to be noted that learners’ shifting of attention to form should not cause the interruption of the communication flow, however. Focus on form “occur[s] in conjunction with –but not interrupt[s]—communicative interaction” (Doughty & Varela, 1998, p.114). In order that communication flow is not interrupted, FonF tasks should guide learners to put major focus on communication. Learners should engage in meaning making and communication first and then pay attention to linguistic features when necessary. Task-essentialness of a target form in a task Whether learners’ attention could be directed to a target form is also determined by the degree of “task-essentialness” of the form in a communicative task. When a target form is presented in a way that makes it “task essential,” it plays a vital role in meaning making in a task and thus drives learners to simultaneously attend to form and meaning (Lightbown, 1998; Long & Robinson, 1998). Without using the “task essential form,” to successfully perform a task is difficult, if not impossible. Significance of “task-essentialness” was shown by William and Evans (1998). They found their FonF intervention produced different learning outcomes in learning 11.

(22) of two forms, participial adjectives and passive sentences, with the former learned better than the latter. One of the provided explanations for such discrepancy in learning effects was that the form of participial adjective was more task essential in the treatment, a dictogloss task. Learners needed to attend to the meanings of paired adjectives (e.g. boring and bored) so as to complete the task, but they did not necessarily have to do so with the other target form, i.e., passive sentences. In the distogloss task, the participants showed a tendency to use active voice instead of passive voice. On the contrary, they used participial adjectives more consciously, monitoring their own usage of the adjectives. It was concluded in this study that even with the same FonF intervention, learners’ learning outcomes can vary according to the degree of task-essentialness of target forms. A target form that is task essential enables learners to make connections between meaning and form. Besides the basic principles above, some methodological issues regarding FonF intervention should also be addressed. The following section touches upon some methodological problems with instruction design and the way target forms should be presented in a FonF activity. Several FonF studies, mostly centering on learning of word forms, would be reviewed. 2.3.2 Optimal implementation environment for FonF intervention For more understanding of optimal implementation environment for FonF intervention, issues about FonF instruction design and the way target forms should be presented in a FonF activity will be discussed in this section. Previous studies that are related to these issues will be reviewed, with some methodological problems raised. At the end of this section, how the present study will be designed in response to these problems will be roughly described. Instruction design First and foremost, FonF intervention of learning word forms should help 12.

(23) learners to connect word forms and meanings in a meaning-oriented activity. Saeidi, Zaferanieh and Shatery (2012) and Laufer (2006) both examined the effects of FonF instruction on vocabulary learning, but with different conclusions drawn because the treatments in their studies attracted learners’ attention differently. The treatment in Saeidi et al.’s study (2012) was a dictogloss activity, in which learners listened to an article that involved target forms and then worked in groups to reconstruct the text. As an interactive task, the dictogloss task can successfully direct learners’ attention to connections between word forms and meanings. FonF intervention lacking in meaning-oriented tasks may cause learners to neglect connections between word meanings and forms. Compared to the FonF activity in Saeidi et al.’s study (2012) as reviewed above, that in Laufer’s (2006) study seems not equally meaning-oriented. The task in Laufer’s (2006) study required the learners to answer comprehension questions, originally designed with an attempt to draw learners’ attention to meanings of the target words in the texts. However, some of the comprehension questions were actually straightforward enough for learners to answer simply by copying the words from the articles without careful interpretation of the word meanings. Plus, participants in this study were under the condition of working alone. Without interactions with others, their need to engage in negotiation of meaning was obviated; they might be able to complete the task of answering comprehension questions without paying attention to target word forms. Whether FonF helps learners to connect word meanings and forms through such intervention thus remains questionable. With “negotiation of meaning” as the core, FonF requires activities in instruction be interactive, and learners’ attention be directed to connect word forms and meanings so as to ensure learning effects of word forms. The next issue regarding FonF instruction design is that the duration of FonF intervention should be long enough so that the treatment can make significant effects 13.

(24) on learning of forms. Noted by Han et al. (2008), many FonF studies equated effectiveness of FonF intervention with production of target forms in immediate output-oriented posttests. But it was often found in such studies that their FonF intervention only involved one-shot treatment and lasted for a short period of time, thus deriving false conclusions that cast doubts about the effectiveness of FonF. With so little time spared for processing, the expected results of form learning are hard to achieve. Two examples of studies which help to explain the significance of long-term treatments are reviewed as follows. Swain and Lapkin (2001) examined language learning effects of a dictogloss task but found no significant differences between the pretest and the posttest. One of the explanations was that participants in this study had limited time on task (i.e. ten minutes), not to mention the fact that the participants received only one-shot treatment. The attempt to promote learners’ interlanguage, which is rather stable within a short period of time, obviously failed. On the other hand, Alcon (2007), aiming to investigate effects of teacher-initiated focus on form of vocabulary, collected data of audio-recorded classroom conversations, learners’ diaries, posttest and delayed posttest translations for a whole year. Her conclusion was thus based on rich data, suggesting that learners’ noticing of target forms was likely to lead to language usages in immediate and delayed posttests, a sign of language learning. Long-term intervention is more likely to make complex language problems solved (Long & Robinson, 1998) and learning effects take place. The present study, confined by the school schedule of the participants, will not be able to allow data collection as long as that in Alcon’s study. But still, it will be made sure that the period of treatment is realistically long enough for possible learning outcomes to occur. The way target forms should be presented in a FonF activity 14.

(25) The context in which target words are presented, that is, whether they are given in oral or written discourse, is influential (Han et al., 2008; Plonsky & Loewen, 2013). As found in Plonsky and Loewen’s study, not all target words were learned equally, suggesting that the participants’ conscious attention to word forms might be associated with but unable to guarantee successful word learning. One explanation for this, though having not yet been fully investigated, was that the participants might have difficulty in inferring meanings of target words from spoken discourse, as processing word meanings through mere listening might be cognitively demanding. The connections between word meanings and forms are difficult to make by receiving input simply by listening. The similar problem of presenting target words in oral discourse was posed by many other studies investigating word learning mainly through listening comprehension tasks as treatments (de la Fuente, 2002; Ellis et al., 1994; Shintani, 2012, 2013). In these studies, the target words were all embedded in oral discourse. The learners had difficulty learning word forms by only being exposed to the input in listening comprehension tasks; they might succeed in completing the listening tasks and show comprehension, but lack of sufficient time to process meaning and forms of unfamiliar words while listening might limit the learning effects. The present study is thus motivated to investigate learners’ learning of word forms through treatment with target words presented in written discourse, in which learners are provided with more time to process meaning of words. Some methodological issues of FonF intervention has been dealt with in this section. The present study attempts to respond to problems raised by previous studies, designing treatments that might further enhance learning effects of word forms in FonF intervention. FonF tasks in this study will be interactive and meaning-oriented so that learners pay major attention to communication. Target words will be made 15.

(26) “task-essential” through task design to drive learners to make the connections between word forms and meanings. Learners’ focus on form will take the form of written discourse so as to provide more time for word processing. Also, the FonF treatment in the present study will be administered for a period of time long enough for possible learning effects to take place. 2.4 How to Record And Analyze Learners’ Attention to Target Forms? Focus on form episodes (FFEs), or parts of language protocols that show learners’ monitoring their language production, provides insights into learners’ focus on form. Whether participants’ attention is directed to form being essential in FonF research, it is thus necessary to record such attention shift for further analysis of their focus on form. The present study, aiming to investigate the effects online discussion has on learning of word forms, will attempt to elicit FFEs of the participants during their discussions. 2.4.1 Definition of FFEs FFEs, or LREs (language-related episodes) as in some studies (e.g. Park, 2005; Swain & Lapkin, 2001), are defined as “any segment of the protocol in which the learners talk about language they are producing, question their language use, or otheror self-correct their language production” (Park, 2005, p.10). Language protocols are recorded for further analysis, during which the language-related segments can be identified based on the operational definitions: “An FFE was considered to begin when the word was topicalized in the discourse; the FFE ended when the topic changed to something else” (Plonsky & Loewen, 2013, p.10). Researchers analyze FFEs to identify occasions “where there was attention to linguistic form (i.e. grammar, vocabulary, spelling, discourse, or pronunciation)” (Ellis et al., 2001, p.294). In this study, whose research focus is on learning of word forms, online discussion will be implemented with an attempt to elicit FFEs that mainly center on lexical items. 16.

(27) 2.4.2 Reasons for eliciting FFEs from treatments in this study FFEs are recorded as evidence of learners’ attention to form during communication in FonF intervention. The treatment of this present study (i.e. online discussion) aims to elicit FFEs based on the following reasons. First, FFEs are mostly collected in communicative interactions between teachers and students or students themselves (e.g. Alcon, 2007; Ellis et al., 2001; Park, 2005; Plonsky & Loewen, 2013; Swain & Lapkin, 2001), which responds to one of the FonF theoretical premises, the heavy emphasis on communication. Second, either initiated by teachers or by students, FFEs are direct evidence for learners’ occasional shift of attention from meaning making to forms. FFEs work as a simple way to prove learners’ attention shift, which is hard to observe otherwise. Also, eliciting FFEs is a relatively natural method to use without interfering with the process of communication in FonF intervention. 2.4.3 Categories of FFEs Ellis et al. (2001) defined categories of FFEs, which will be used as guidelines on elicitation of language-related segments in online discussion in this study: Type A: Responding FFEs: In this type of episodes, participants respond to utterances produced by others because meanings are not conveyed clearly. Type B: Student-initiated FFEs: In this type of episodes, participants initiate the episodes due to some gap in meaning making. Such FFEs typically begin with questions raised by participants. Type C: Teacher-initiated FFEs: In this type of episodes, teachers initiated the episodes because they think certain linguistic forms might be problematic to participants. In the present study, FFEs will be elicited mainly through teachers’ guidance, which will serve as some “triggers” to enable learners to be aware of their linguistic gap and shift their attention to forms accordingly. 2.5 What Is the Research Purpose? 17.

(28) The main research purpose of this study is to investigate effects FonF intervention has on learning of unfamiliar word forms and reading comprehension. The FonF intervention is provided in the form of online discussion on reading materials, with the purpose of eliciting learners’ FFEs. Based on previous studies about effects of FonF intervention on learning of forms, the research questions of the current study are framed as follows: Research Question 1: Do FFEs in online discussion enhance learners’ learning of the forms of the target words in the assigned reading texts? Research Question 2: Do FFEs in online discussion enhance learners’ retention of the forms of the target words in the assigned reading texts? Research Question 3: Do FFEs in online discussion influence learners’ reading in terms of three types of comprehension of the assigned reading texts? The three research questions are elaborated as follows. 2.5.1 Research Question 1: Do FFEs in online discussion enhance learners’ learning of the forms of the target words in the assigned reading texts? Learning of word forms in FonF intervention might be beneficial and effective, thus deserving more attention. As vocabulary serves as meaning units of a language, failure to know a word is very likely to cause communicative breakdown (Laufer, 2006), which can be a good trigger for shifting learners’ attention to form in FonF intervention (Long & Robinson, 1998). This study therefore hopes to examine the effects learners’ FFEs in online discussion make on their learning of the word forms. 2.5.2 Research Question 2: Do FFEs in online discussion enhance learners’ retention of the forms of the target words in the assigned reading texts? Retention of linguistic items over a longer period of time serves as one of the indices of learning, but this issue needs further exploration in FonF research. Some FonF studies have investigated retention of words after FonF intervention (e.g. de la Fuente, 2002; Dobinson, 2001; Shintani, 2013). Dobinson, with his study results 18.

(29) showing that the shifting of learners’ attention to words facilitates vocabulary learning, found that the retention rates of vocabulary were high. Some of the retained words in the delayed tests were not even immediately recalled after intervention, which implied that learning word forms requires a rather long period of time. De la Fuente, on the other hand, suggested that output from learners in meaning negotiations facilitates retention of words, while negotiations without output from learners failed to do so. However, Shintani found word retention through the treatment with input-based tasks but without output. The needs to explore word retention through FonF instruction was stressed by these studies, which this current research follows and hopes to provide an insight into the issue of word retention in FonF intervention under the condition of longer period of time for the treatment to make effects. 2.5.3 Research Question 3: Do FFEs in online discussion influence learners’ reading in terms of three types of comprehension of the assigned reading texts? Few FonF studies have measured participants’ comprehension of input, but it actually plays an essential role in thorough understanding of learning effects brought by FonF intervention. It has been suggested that comprehension should be viewed as one dependent variable in FonF studies so that “balanced development” in input comprehension and learning of forms could be pursued (Han et al., 2008, p.603). However, the current situations are that although some FonF studies have applied listening comprehension tasks as treatments (e.g. de la Fuente, 2002; Ellis, Tanaka, & Yamazaki, 1994; Shintani, 2012, 2013), almost none of them involve measuring learners’ comprehension of input in their study design. The present study thus attempts to fill this research gap by including participants’ comprehension of input as a dependent variable. Another important reason why this present study views measuring learners’ comprehension of input as necessary is related to the vital role played by vocabulary 19.

(30) in text comprehension. As vocabulary has long been regarded as essential in learners’ reading comprehension (Hirsh & Nation, 1992; Hu & Nation, 2000; Laufer, 1989; Nation, 2001; Qian, 1999; Schmitt, Jiang, & Grabe, 2011), whether learners’ focus on word forms would be associated with their text comprehension is thus worth investigation. The last research question attempts to fill the research gap in relationship between FonF intervention and learners’ comprehension of input, and to confirm the essential role vocabulary plays in reading comprehension from the perspective of FonF. This research question will be dealt with through assessing different comprehension types, including global comprehension, local comprehension and inference. It is worth investigating whether the effects brought by FFEs will vary among different comprehension types.. 20.

(31) Chapter Three Methodology This research aims to investigate the effects that online discussion through SNSs as a FonF intervention makes on learning word forms and text comprehension. The study design will be elaborated in detail in this chapter, which is divided into five subsections. The first subsection gives background information about the participants; the second specifies the instruments were used in the study. Materials in the FonF intervention are introduced in the third subsection. The subsection four details the process of the study. Last but not least, how study data were analyzed is specified in subsection five. 3.1 Participants The 82 participants of this study were tenth graders in a private girls’ high school in Taipei City. They came from two intact classes, taught by the same English teacher. The researcher approached the English teacher three months earlier to invite her students as the participants. Making sure that all her students were willing to participate, the English teacher kindly offered ten periods of her class, throughout four months, for the intervention, and two additional ones before the intervention for preparation for the study. The two classes were randomly appointed as the experimental group (EG) and the control group (CG). While the CG participants did not form groups, EG participants were divided into groups of four to five. Considering not only the needs to maintain lively discussion but also the principle that a group should comprise no more than six people (Johnson & Johnson, 1989) to increase opportunities for every member to share their ideas, four or five was an appropriate number for group members in this study. Each group in EG created a “community” on Google Plus, in which the 21.

(32) researcher took part as one member. There was a webpage exclusively for every discussion group. 3.2 Instruments Google Plus In this study, Google Plus was used as a platform for CMC as some functions it provided could meet the needs of this study. Users of Google Plus could create a “community” and had a community page exclusively for their group members. They could post any messages or pictures on the platform, and so could they give responses to one another. Users would be informed of new messages through the notification function of Google Plus. This platform also provided a “Plus One” button that allowed its users to show they like or agree with a message. To have access to Google Plus, the EG participants needed to have a Google account, which most of them had had thanks to the popularity of Google. After each EG participant provided their Gmail address, the researcher created nine communities on Google Plus, each of which included four to five participants as group members1. Invitation letters were sent by Google Plus to provide access to the online communities. What the EG participant needed to do was to click “accept the invitation” and then they could see the webpage of their online communities. The participants were encouraged to create a new identity on the platform of Google Plus. They could use whatever names they liked other than their real names when they were sharing ideas. Vocabulary Size Test An online version of Vocabulary Size Test (Nation & Beglar, 2007) was applied to measure the participants’ approximate vocabulary size. The participants answered 140. 1. The EG participants had chosen the group members by themselves in advance, and handed in the name lists to the researcher for the grouping. 22.

(33) questions for the test. For every item, they saw an L2 word with an example sentence, and they needed to choose the L1 meaning for the word. To lessen guessing effects, an option “I don’t know” was available for the participants to choose if they had never seen the word. After answering 140 questions, the participants would see a webpage that showed their estimated vocabulary size, which was recorded by the researcher. Vocabulary familiarity check Vocabulary familiarity check aimed to help the process of target word selection. This checklist was adapted from Vocabulary Knowledge Scale (Wesche, & Paribakht, 1996) (Appendix A). On the checklist, the participants saw all the potential target words, of which they demonstrated their familiarity by choosing one of the four options: I. I don’t remember seeing the word before. II. I have seen this word before, but I don’t know the meaning. III. I have seen this word before, and I guess it means____ but I’m not quite sure. IV. I have seen this word before, and I know this word means____. If the participants chose option III or IV, they would be required to write down the meaning of the word in either L1 or L2. The chosen option would be valid if a correct meaning was given; otherwise, the option at the lower level would be counted. For example, if a participant chose option III for a word but gave a wrong meaning, her familiarity of this word in question would be rated as level II. Among the four levels, options II and III indicated some partial knowledge had been developed. According to the theoretical premises of FonF, target forms can be learned more effectively if there is learner readiness of them (Han, Park, & Combs, 2008; Park, 2005). Thus, the potential target words responded by options II and III most were selected as the target words. Pretest of word forms (WF) 23.

(34) The WF pretest involved all the target words from each assigned text, and the participants selected the correct word forms for their L1 meanings. One point was given to a correct response. This is one example question of the pretest: 吸引 (A) attract. (B) attack. (C) attact. (D) attrack. As can be seen in the example question, the correct response should be (A), and other distracters might be either English words with similar forms (e.g. (B)) or non-words that were made up based on English alphabetic principles (e.g. (C) and (D)) to prevent the participants from relying on test strategies to get the correct answers. Immediate and delayed posttests of word forms (WF) WF posttests aimed to assess the learning of word forms. With its format the same with those in the pretest, an immediate WF posttest was taken right after the treatment every week. One week after the immediate posttest of word forms, the participants took a delayed WF posttest, which was the same one as the immediate posttest, but the order of the questions was rearranged. All the weekly posttests included the target words covered in the weekly assigned texts as well as some filler items. One point was given to a correct response in WF posttests. Reading comprehension check The reading comprehension check, aiming to assess reading in terms of three comprehension types, involved nine questions, composed of three global comprehension questions, three local comprehension questions, and three inference questions, based on the assigned texts the participants just read during the intervention. The comprehension check was taken after weekly treatments and immediate WF posttests. One point was given to each correct response. 3.3 Materials Texts for reading 24.

(35) The Green Room, a graded reader by Robert Campbell, was used as the reading materials in the study, with one chapter assigned weekly during the intervention. Written with language learners as target readers, graded readers usually come in a range of difficulty levels for L2 learners with different language proficiency levels. One reason behind the choice of this book was that the complexity of vocabulary use and sentence structures in The Green Room was similar to that in the participants’ textbooks, suggesting its appropriateness in its difficulty level for the participants and its suitability as the reading materials for such controlled conditions as a study. Another reason was that the story of The Green Room was about the friendship among four teenagers and the process of their pursuit of dreams. The theme might be interesting and motivating for the participants. Target words Three target words were selected from each assigned chapter in The Green Room through the following steps. First of all, the researcher selected some words that might be unfamiliar to the participants in advance. As a high school English teacher who had some understanding of the approximate vocabulary size of tenth graders in Taiwan, the researcher was able to screen all words in the reader and select potential target words. Second, the potential target words that appeared in the frequency word list which was consistent with the participants’ current vocabulary size were eliminated. For example, if it was estimated by Vocabulary Size Test that the participants were equipped with approximate 2,000 vocabulary, words included in the 2,000 high frequency word list would be eliminated. And then, the rest of the potential target words were compiled into the list of the vocabulary familiarity check, through which the participants demonstrated their familiarity of the words so that the researcher could do the ultimate decision on target word selection. It was made sure that all the target words were similar in their difficulty level in terms of their length (i.e. number 25.

(36) of syllable) and frequency range. With three target words selected from every chapter, there were thirty target words in total included in the study. Text-based questions All the text-based questions aimed to elicit the use of the target words. There are mainly two types of them: display questions and referential questions. Display questions: This type of questions are mainly comprehension questions based on the contents of the assigned chapter, and questioners have already known the answers to them. In this study, display questions were asked for the purpose of directing learners’ attention to target word forms and helping them to link the target word forms to their meanings. Referential questions: The answers to this type of questions are unknown. In this study, such questions encouraged learners to express their own opinions about the reading or make some inferences and predictions. Asked for the purpose of providing more opportunities for the participants to revisit the target word forms, these questions were asked only after the participants had produced the target word forms in the EG intervention. Each question was asked one by one in both EG and CG interventions. For an example of how discussion questions were asked to elicit target words in a sample text, please refer to Appendix B. 3.4 Procedure 3.4.1 Preparation for the Study Some preparations were made two weeks before the study formally started. The participants took Vocabulary Size Test in a computer lab. The test was not timed, and it took about 30 minutes for the participants to finish 140 items. The results of the test were recorded by the researcher. Based on the participants’ vocabulary size, the target words were selected through the process that has been described above. The WF pretest was also taken during the preparation phase. The participants in both EG and CG spent 15 minutes on the WF pretest. 26.

(37) 3.4.2 The Intervention Table 1 shows the schedule of the intervention and the tests.. Table 1 Schedule of the Conduct of the Intervention and the Tests Time. assigned chapter. WF immediate posttests. WF delayed posttests. comprehension check. 1. Chapter 1. 1. x. 1. 2. Chapter 2. 2. 1. 2. 3. Chapter 3. 3. 2. 3. 4. Chapter 4. 4. 3. 4. 5. Chapter 5. 5. 4. 5. 6. Chapters 6, 7. 6. 5. 6. 7. Chapter 8. 7. 6. 7. 8. Chapters 8, 9. 8. 7. 8. 9. Chapters 10,11. 9. 8. 9. 10. Chapter 12. 10. 9. 10. 11. x. x. 10. x. As shown in Table 1, the intervention took ten class periods (about eleven weeks, with one class period, or 50 minutes, per week) to enable possible learning outcomes produced by the treatments to occur. The participants were assigned one chapter from The Green Room every week. Both EG and CG participants did silent reading individually for seven minutes and then received their treatments, which will be elaborated in the following. No reading was assigned in the last week, when the participants took the last delayed WF posttest of words they had learned in the previous week. Treatments of the experimental group (EG) The EG participants did online discussion, so their treatment was implemented in a computer lab. After reading the assigned chapter for seven minutes, the participants 27.

(38) spent 30 minutes having discussions online on questions related to the assigned reading. Discussion on one series of questions centering on the same target word took ten minutes. During the online discussion time, the EG participants were required to answer directly to the questions, respond to their group members at least once, and read all the others’ responses. Online discussion brings benefits that help to solve some problems in traditional discussions, one of which is that learners’ anxiety can be alleviated (Carico & Logan, 2004; Tower, Latimer, & Hewitt, 2013) through new language identities and more time flexibility to get prepared for idea sharing, as mentioned in Chapter One. To benefit the EG participants in this regard, their treatment of online discussion was designed as follows. First, the EG participants were encouraged to use names they like during the online discussion instead of being required to use their real names. Using another name is like creating a new identity, helping learners to feel safe and comfortable in idea sharing. Second, the EG participants could take their time reading questions and all the responses of others before giving responses on the platform. After reading a question or a response, they gave a “Plus One” to it. Teachers serve as a facilitator in FonF online discussion. Instead of leading the discussion, they can devote more time and energies to directing and monitoring learners’ use of language forms. In this study, the researcher did so through skillful use of different types of text-based questions to elicit the participants’ use of the target words. Figure 1 shows the process.. 28.

(39) Figure 1. The Process of the Teacher’s Elicitation of Target Word Forms from the EG Participants. For the purpose of guiding the participants’ attention to a target word form, the two different kinds of questions were posed according to the responses given by the participants. As illustrated in Figure 1, a display question was asked first. If it succeeded in eliciting the target word (c), a referential question that provided opportunities for more exposure to the target word forms were asked. As in case (b) when the target word was not elicited but the participants’ responses showed comprehension of the first question, another display question was asked to help the participants notice the word form they previously neglected and linked it with the meanings. The other case (a) happened when the participants did not use the target word form, and neither did their responses show their comprehension of the first display question. The display question was rephrased to help the participants elicit the target word form. All the questions were posted one by one. A discussion thread, led by one display question, took ten minutes, after which the next display question for another target word was posted to start a new topic thread. During the online discussion, the 29.

(40) participants could ask questions on the platform in L2 if they had problems producing language forms, but no interactions in other forms (e.g. oral) were allowed. More detailed instructions on the online discussion have been included in Appendix C. Besides giving text-based questions to elicit the target words, the researcher occasionally participated in the EG participants’ discussion, but only when necessary. While the EG participants were having online discussion, the researcher monitored their discussion and gave occasional responses to help the participants keep on the right track. To create a FonF environment, the researcher made sure that most of her responses were meaning-oriented and that only few of them were reminders of the word spelling. The approximate proportion of meaning- and form-focused responses was 70 to 30 percent. Treatments of the control group (CG) The treatment of CG was given in a traditional classroom. After doing the seven-minute silent reading on the same reading texts, the CG participants were given the same discussion questions as those the EG participants received one by one. The CG participants were asked to write down their answers to the questions in L2 individually. The treatment of individual work in CG took 30 minutes every week, with ten minutes spent on each series of questions dealing with one target word. No interactions in any forms should happen among students or between the teacher and the students. During the ten-minute session for each series of questions, however, the CG participants were asked to pass their responses to their peers and also read others’ responses. They put a star beside others’ responses as a sign of finishing reading. Appendix C involves more detailed instructions for CG treatments. 3.4.3 The Posttests After the 30-minute treatments, both CG and EG took a three-minute immediate WF posttest and a five-minute comprehension check. To avoid extra exposure to the 30.

(41) forms of the target words, the immediate WF posttests preceded the comprehension check. A three-minute delayed WF posttest was given one week later. The whole procedure of the study is summarized in Table 2:. Table 2 The Procedure of the Study Activity. Time (mins). participants’ vocabulary size measured. 30. target words selected. NA. WF pretest. 15. silent reading. 7. EG: online discussion on text-based questions. 30. CG: individual work on text-based questions. 30. WF immediate posttest. 3. comprehension check. 5. WF delayed posttest (one week later). 3. Phase. Preparation. Intervention (ten times) Posttest (each time after the intervention). 3.5 Data Analysis With the purpose of answering the research question, which is to investigate the effects online discussion has on learners’ reading comprehension and learning of word forms, the data was analyzed with SPSS software. 3.5.1 Research question 1: Do FFEs in online discussion enhance learners’ learning of the forms of the target words in the assigned reading texts? The first research question was dealt with by two-way ANOVA to see if significant differences existed among the WF pretests, the EG’s and the CG’s scores of the weekly immediate WF posttests. The significance differences would imply that FFEs could enhance learners’ learning of word forms. 3.5.2 Research question 2: Do FFEs in online discussion enhance learners’ retention 31.

(42) of the forms of the target words in the assigned reading texts? The second research question investigates the effects which FFEs bring to learners’ retention of word forms. Comparisons was made among the WF pretests, the EG’s and the CG’s scores in weekly WF delayed posttests. Two-way ANOVA was applied in the comparison. The significant differences might indicate that FFEs could enhance learners’ retention of word forms. 3.5.3 Research question 3: Do FFEs in online discussion influence learners’ reading in terms of three types of comprehension of the assigned reading texts? To answer this research question, one-way ANOVA was done to examine if there were significant differences existing between weekly comprehension check scores of the EG and the CG in terms of global comprehension, local comprehension, and inference. It might be inferred that FFEs could influence learners’ reading comprehension of the assigned reading. Besides ANOVA, variations in data were also analyzed in order to have a comprehensive view into the process of learners learning the target word forms and improving their text comprehension through the interventions, which will be further elaborated in the next chapter.. 32.

(43) Chapter Four Results and Discussion 4.1 Introduction of Data Analysis Methods This research investigates the effects of online discussion on L2 learning of word forms and text comprehension from the perspective of FonF. With the hope to thoroughly examine this learning process, this study applied analysis of variance (ANOVA) combined with analysis of variations in the data, with the elaboration given below. This study applies both methods to analyze the data because while working together, they offer a comprehensive view on the possible learning effects brought by the interventions. ANOVA serves as a straightforward quantitative method to analyze the data, proving the effectiveness of the FonF interventions through collapsed mean scores. On the other hand, analysis of data variation investigates the diachronic development of the data, thus enabling more intriguing research results to emerge. As analysis of data variation is an approach not adopted as often, how it originated theoretically and how it analyzes data are explained in this section. 4.1.1 Theoretical bases of the analysis of data variations The analysis of variations in the data is based on Dynamic Systems Theory (DST). DST suggested that all systems are composed of sub-systems, which interact with one another continuously. Such interactions among sub-systems and factors influence the larger systems, which accordingly evolve in a chaotic and unpredictable way (de Bot, Lowie, & Verspoor, 2005). That DST can be applied in language learning research is founded on the concept that language development itself can be regarded as a dynamic process (Lowie, & Verspoor, 2015). Factors of language learning, such as learning of pronunciation, 33.

(44) grammar, vocabulary, other sub-systems of languages, and learner factors, no matter detectable or not, keep affecting one another. Their relationships continuously change over a long period of time, and thus contribute to the complexity of language development. Some concepts of DST are applicable to the data analysis in this study as some theoretical bases of FonF learning is consistent with them. For one, the inclusion of word form learning and text comprehension based on “balanced development” of FonF learning (Han, Park, & Combs, 2008) actually represents an attempt to look into the relationship between these two sub-systems of FonF language development. For another, as elaborated previously, FonF intervention might need a rather long period of time to take effects (e.g. Swain & Lapkin, 2001; Alcon, 2007), which supports the claim of DST that evolution of a system (language development in this regard) is a longitudinal and dynamic process. Based on the justifications above, this study applies both ANOVA and analysis of variations, founded on DST, to analyze the data. While ANOVA provided a quantitative method and a clear insight into the results, analysis of data variations was on the other hand more likely to enable the researcher to better examine the process of the learners learning word forms and improving their text comprehension with the aid of FonF intervention. 4.1.2 Interpreting data variations based on DST Though most DST studies look at variations in individual cases (Lowie & Verspoor, 2015), this study chose to analyze variations in diachronic mean scores of the whole group considering its research focus on effects brought by online discussion. It is assumed in this study that interactions among individuals may facilitate learners’ learning of words and their text comprehension. As the EG participants were divided into different heterogeneous groups, each of which developed their own interpersonal 34.