台灣國中英文課室聽力教材比較之研究

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. 台灣國中英文課室聽力教材比較之研究. A Comparative Study of the Use of Scripted and Unscripted Listening Text in the English Classroom in Taiwan. 指導教授：劉. 宇. 挺. Advisor: Dr. Yeu-ting Liu 研究生：林. 欣. 秀. Hsin-hsiu Lin. 中華民國一百零六年六月 June, 2017.

(2) 中文摘要. 多數的台灣英文教師使用「有固定講稿」的聽力教材，而這種聽力教材似乎無法有效幫助學生在教室外理解真實的情境對話。此研究的目的是要探究「無固定講稿」的聽力教材是否能增強學生的聽力理解，並比較學生在這兩種教材聽力的表現。共有七十位就讀台北市國中八年級的學生參與此實驗。參與者被分為四組並在十個禮拜中接受不同的聽力訓練與測試。結果顯示無論是「有固定講稿」的聽力教材或是「無固定講稿」的聽力教材皆無法幫助學生達到顯著進步。而「無固定講稿」的聽力「測驗」卻有益於學習者的聽力，同時也提升了學生學習聽力的動機，也製造了好的考試準備效果。根據「遷移適宜歷程理論」(TAP)，當學習與使用的過程一致時，學習者的能力可發揮最大效用。此研究結果建議「無固定講稿」的聽力教材的使用，因為教材的性質與真實對話相近，因而能培養學生真正的聽力能力。而且，「無固定講稿」的聽力測驗能測出學習者的真實能力。支持使用「無固定講稿」的聽力教材之論點將於第五章中接續討論。. 關鍵字：「有固定講稿」的聽力教材、「無固定講稿」的聽力教材、「遷移適宜歷程理論」 i.

(3) Abstract In Taiwanese English classes, teachers mostly utilize ‘scripted’ listening materials in listening teaching, which did not fully equip learners with competence to comprehend the real-life speech outside the classroom. The purpose of the current study was to (1) investigate whether the ‘unscripted’ text is beneficial to learners and, (2)compare and contrast their performances in both ‘scripted’ and ‘unscripted’ listening tests. Seventy eighth graders in Taipei public junior high school were included as participants. Participants were categorized into four groups and treated with different combinations of listening materials during 10-week experiment. The results indicated that neither unscripted nor scripted texts conduced to significant improvement in learners’ listening comprehension. However, the ‘unscripted tests’ not only benefited learners’ listening abilities but also raised their motivation in English listening. The positive backwash was found in this type of listening test. Based on the hypothesis of Transfer Appropriate Processing (TAP), the congruency between learning and using conditions can benefit learners. In this study, the result suggested implementing the ‘unscripted’ text in classroom settings because the text is more like real-life speech and it consequently can equip learners with the real ability to comprehend native speakers. What is more, the ‘unscripted’ test is a ii.

(4) useful assessment to measure learners’ accurate listening ability. The reasons of calling for the use of ‘unscripted text’ were discussed in the pedagogical implication. Keywords: Scripted Text, Unscripted Text, Transfer Appropriate Processing (TAP). iii.

(5) Acknowledgements Writing a master’s thesis is not something that can be finished without supports, cooperation and assistances from many people. First and foremost, I would like to show my sincere gratitude and respect to my advisor Dr. Yu-ting Liu for his kindness and patience. With his professional teaching and valuable feedback, I was led to deeper thought about my study and the fantastic world of academic research. During the whole writing process, I learned how to decide a good topic, design a proper experiment, analyze the data and make the discussion more complete. On every step, he made a careful scrutiny into my study and guided me in great patience. I was inspired by Dr. Liu’s fine personality. I learned not only how to write a research but also how to conquer the obstacles with positive attitude and strong will. Thanks to Dr. Liu, I did not give up in the middle of research and have confidence in not surrendering to difficulties easily in the future. My gratitude also goes to my committee members, Dr. Angela Wu and Dr. Shiau-ping Tian. Their professional suggestions made my thesis more complete and enriched the content of my thesis. I would like to thank to my dear students in Liu-Gong junior high school as my participants, who devoted their valuable learning time to this research. Ten weeks were long time for these students in a semester. Without their cooperation, I could not iv.

(6) finish my experiment. Special thanks also go to Emily Kam, Winnie Chiu and my dear friend, Florence. They devoted valuable time and efforts to recording all the dialogues. The listening materials they made were the most crucial factors to shed the new light in my study. Next, I want to express my gratitude to all my friends in NTNU, who would share their precious suggestions with me and encourage me to complete the whole journey. We had taken many courses in the TESOL Master’s program and went through hard time together. Finally, I must acknowledge my family and friends, who always support me with their whole hearts. With their encouragement, I could finally achieve my goal in the end of this academic journey.. v.

(7) Table of Contents Chinese Abstract ……….…………………………………………………………... i Abstract ………………..…………………………………………………………… ii Acknowledgements …...……………………………………………………………. iv Table of Contents ……………………………………………………………………vi List of Tables and Figures …………………………………………………………..ix Chapter One: Introduction ………………………………………………………… 1 1.1 Background and Motivation ……………...……………………………….. 1 1.2 Rationale of the Study ………………………………………………………. 7 1.3 Purpose of the Study ………………………………………………………... 8 1.4 Significance of the Study …...………………………………………………. 9 Chapter Two: Literature Review ………………………………………………… 10 2.1 The Differences between Scripted and Unscripted text ……..…………….. 10 2.2 Existing Empirical Studies on the Performance of EFL Learners Using Scripted and Unscripted Text …………………………………………………. 12 2.3 Current Use of Unscripted Text in the EFL Classroom …….……………... 16 2.4 Transfer Appropriate Processing …………………………………………. 21 2.5 Research Questions ………...……………...…………………………….... 23 2.6 Summary …………………………………………………………………... 23 vi.

(8) Chapter Three: Methodology ………………….…………………………………. 25 3.1 Participants ……………………………………………………………….. 25 3.2 Materials ……………………………………………………….…………. 25 3.2.1 Listening Materials ……...………………………………………… 26 3.2.2 Testing Instruments ………………………………………………… 29 3.3 Design ……………………………………………………………………... 31 3.4 Procedure …………………………………………………………………. 32 Chapter Four: Results …………………………………………………………….. 34 Chapter Five: Discussion …………………………………………………………. 43 5.1 The Interpretation of the Results based on TAP ………………………..…. 43 5.2. How Unscripted Texts Influence on Learners’ Comprehension ………….. 45 5.2.1 The Benefit of Discourse Features ……………………………….... 45 5.2.2 The Time Effect on Listeners’ Comprehension ………………….…. 47 5.3 The Comparison between Two Types of Listening Texts ……………….…. 48 5.3.1 Learners Positive Perception toward the Unscripted Texts ……….. 48 5.3.2 Learners’ Negative Perception toward the Scripted Texts ……….... 50 5.4 The Unscripted Listening Tests …………………………………………… 52 5.4.1 The Value of the Unscripted Tests ……………………………….… 52 5.4.2 The Backwash of the Unscripted Tests …………………………….. 53 vii.

(9) 5.5 Comparison of the SU and UU Groups ……………………………...……. 53 5.6 Pedagogical Implication ………………..………………………………… 54 Chapter Six: Conclusion ………………………………………………………….. 58 6.1 Summary ………………………………………………………………...… 58 6.2 Limitations of the Present Study and Suggestions for Future Research …... 59 References ….……………………………………………………………………… 61 Appendix A.....………………………………………………………..……………. 64 Appendix B.....…………………………………………………………..…………. 67. viii.

(10) Lists of Tables and Figures Tables Table 1. The Sample of Scripted Listening Test ……………..………………….. 30 Table 2. The Sample of Unscripted Listening Test……………...……………….. 30 Table 3. The Design of Grouping ………………………………………………... 32 Table 4. The Summary of the Results of One-way ANOVA at the First Time Point ... 35 Table 5. Two Way Split Factor ANOVA ………………………………………..... 36 Table 6. One Way ANOVA for the UU Group ……..…………………….…...…. 37. Table 7. Paired t-test for the UU Group …………………………………………. 38. Table 8. One Way ANOVA for the SU Group ……………...…………………..... 39 Table 9. Paired t-test for the SU Group ………..……………………………….. 39 Table 10.. One Way ANOVA for the US Group …………………...……………..… 40. Table 11.. One Way ANOVA for the SS Group ……………………………………. 40. Figures Figure 1. Summary of Experimental Design over a Week ……….……………….... 33. ix.

(11) CHAPTER ONE INTRODUCTION 1.1 Background and motivation Listening is an important skill that EFL learners cannot ignore. With a good listening comprehension ability, learners can understand native speakers better. In foreign language classrooms, English teachers have taught listening skills for several years, and have conducted various kinds of listening tests to evaluate learners’ listening comprehension ability. Students are used to listening to dialogue that native speakers read while reading script in textbooks and taking listening tests, which are also constructed with script. Listening to text and answering questions via artificial conversation are the most frequently-used types of listening instruction in the foreign language classroom. To give readers a clear understanding of how an English teacher runs his or her English listening instruction, here is a concrete example in a classroom setting. Murphy, who teaches English in a junior high school in Taiwan, expends a lot of effort teaching listening in her English class. Knowing that an English listening test will be included in the Comprehensive Assessment – a test used to evaluate junior high school students’ academic abilities in Taiwan – Murphy spends nearly one hour out of four each week teaching listening skills. In her listening class, students are 1.

(12) asked to listen to dialogue and text read by native speakers and answer multiple-choice questions to evaluate their listening comprehension. To make the class more interesting, students even watch an animated version of the dialogue with script. The script content of the animation is the same as the reading content in each lesson. Now, let’s take a look at Murphy’s English class. Students are learning the sentence pattern “What time do you do something?” The following text is the script that the students listened to: Woman: “What time do you go to bed?” Man: “I go to bed at 10:30. How about you?” Woman: “I go to bed late. I go to bed at 11:30.” Question: What time does the man go to bed? Almost two-thirds of Murphy’s students were able to give the right answer to the question. With this kind of listening training, students seem to perform well in Murphy’s English class, until one day, George, an English native speaker, was invited to the class to carry out an activity. Students were excited to take turns to interview their guest. When it came to Peter’s turn, he was confused about what he heard. Peter: “What time do you go to bed?” George: “Well…this is an interesting question. Uh…sometimes I go to bed after 2.

(13) a shower…after a shower…so... (pause)…maybe at about 12:00. But if I need to wake up early the next day, I’ll end my day at 11:00….you know…I don’t want to oversleep.” Peter: “So you go to bed at 11:00?” George: “Only if I have to get up early the next day. But yes, yes, you’re right. At about 11:00 to 12:00.” After the interview, Murphy asked the question “What time does George go to bed?” To her surprise, most students were uncertain and unable to answer the question. The phenomenon posed some questions: “Why can’t students understand, even though George has attempted to elaborate upon his response using very simple words?” “Is there any difference between the record in the textbook and George’s speech?” “What’s wrong with the listening materials?” “Is it necessary to simplify the listening text?” In fact, Murphy is not alone. Many EFL teachers encounter similar problems. As you can see from the above example, Murphy trained students using lots of scripted listening texts, but scripted repeated listening activities do not seem to effectively assist foreign language learners to understand and converse with native speakers successfully. The truth is that learners still find it very difficult to understand native speakers in the real-world setting. The problem might be caused by the 3.

(14) listening texts that learners are exposed to in EFL classrooms. Real-life listening contains some variables that do not exist in textbooks. Take Peter and George’s conversation as an example; there are some fillers, pauses, and repetitions. Listeners seldom encounter these oral discourse features, which are often considered as “noise”, in EFL listening training materials when they listen to the textbook version. Murphy’s experience is not a scenario unfamiliar to many EFL teachers. The content of listening materials might be the key to unlocking the issue that EFL teachers face. To investigate whether spoken text in reality and textbook text used in the classroom are different, Flowerdew and Miller (1997) analyzed five English textbooks for academic purposes (EAP), and discovered that the spoken text in textbooks was largely different from real-life conversations or lectures. Although researchers and textbook authors have attempted to address this problem, making text closer to real-life content, the text in textbooks is still far from reality (Gilmore, 2004). How are they different? If spoken text is written, well-organized and simplified, it is called “scripted text.” Speakers artificially read aloud the text with a lower speech rate and obviously clearer enunciation. Most texts used in the classroom setting fall into this category. On the contrary, “unscripted text,” which is much like natural talking, tends to be less logically organized and contains more redundancies, less complex syntax and shorter idea units. There are more pauses, fillers, and false starts in this 4.

(15) type of spoken language, because speakers neither have scripts to follow nor rehearse in advance. In most language classrooms, teachers prefer to use “scripted text” rather than “unscripted text”, because the former is easier for students to understand and for editors to produce. Peacock (1997) compared learners’ preference toward “scripted text” and authentic materials, and found that learners showed more interest in scripted text because they had difficulty decoding the oral discourse features in natural conversations. In Kmiecik and Barkhuizen’s study, they also discovered that learners had a more positive attitude towards “unauthentic” material (2006). In addition to the learners’ attitude toward listening materials, textbook writers’ practice also exacerbates students’ reliance on scripted text. Writers generally edit text with script and focus more on linguistic rules than pragmatic usage. This practice, in turn, leads learners to focus more on rule-based knowledge instead of teaching or exposing learners to prosodic cues that are important to parse and comprehend oral target language input, such as pauses, fillers, etc. The prevalence of using scripted text does not necessarily mean that it is effective enough in improving learners’ listening comprehension ability. There is still the need to explore other alternatives, e.g., using unscripted text in a language learning class. According to Wagner (2013), most high-stake assessments used to 5.

(16) examine students’ listening ability in the foreign language classroom belong to the category of polished spoken text, which may pay little or even no attention to training in how to process and comprehend real-world conversation. This might lead to the result that occurred in Murphy’s anecdote described above; i.e., learners did not give a satisfactory performance under unscripted text conditions. What’s more, unscripted text might provide learners with support in fostering their listening ability. It is likely that, given time, the learners from Wagner and Toth’s (2014) study might receive equivalent benefit from unscripted text – a point to be empirically established in longitudinal studies. This is why Wagner and Toth urged researchers to gain further insight into the efficacy of unscripted text using a longer treatment design. In this vein, if we are to establish the need to replace the current dominant listening method – which is mainly based on scripted text – evidence from studies that compare the relative contributions of scripted and unscripted text is warranted. Following this line of research, researchers have obtained both positive and negative results for scripted and unscripted text. For instance, Read (2002) found that test-takers had a better performance with the “scripted” version, and considered “scripted text” much easier to comprehend. Richard (2006) also posited that scripted text should be utilized in language learning because “unscripted text” might be too difficult for learners to comprehend, especially for students at a lower proficiency 6.

(17) level. However, some studies highlighted flaws in using scripted text. For example, Vandergrift (2007) discovered that learners mostly listen to artificial and unnatural scripted text; nevertheless, without special training, they might have difficulty in understanding oral discourse features in unscripted text. What’s worse, getting used to listening to “scripted text” might hinder learners’ development in word-segmenting skills, which refers to segmenting oral input into meaningful units (Vandergrift, 2007). In other words, it is worthwhile to assess whether listening text should be simplified and those elements that impede learners’ comprehension should be eliminated (Vandergrift & Goh, 2012). 1.2 Rationale of the study Empirical studies have been undertaken to establish the theoretical effectiveness of using unscripted natural text to improve students’ listening comprehension (Herron & Seay, 1991; Gilmore, 2004 & 2011). Although these studies have shown a positive effect of using unscripted text in the classroom setting, there are some limitations and deficiencies that warrant further investigation. First, most previous studies were either cross-sectional in nature or included one or two teaching sessions (e.g., Herron & Seay, 1991; Gilmore, 2004; Wagner & Toth, 2014); this did not allow the researchers to objectively gauge learners’ improvement in listening skills, which takes time to develop. Furthermore, learners’ performance might fluctuate under various conditions. 7.

(18) Second, previous studies focused mostly on comparing students’ performance under “scripted” and “unscripted” listening conditions. Gilmore (2011) found that texts under scripted and unscripted text conditions are usually drawn from different genres (monologue vs. conversation), and thus vary in terms of difficulty. For this reason, learners’ performance under the two conditions cannot be compared (see also Read, 2002). Thus, research that addresses this methodological issue is warranted. 1.3 Purpose of the study The purpose of this study was to conduct regular tests to explore whether “unscripted text” was beneficial in terms of improving learners’ listening comprehension ability in the foreign language (FL) context, where scripted text is the mainstream and native input is usually very limited beyond the classroom setting. Some previous studies enrolled Asian EFL learners as participants. It would be interesting to assess how Taiwanese EFL learners perform on listening comprehension tests when presented with these two types of listening material. This study therefore investigated: . Whether scripted and unscripted listening materials vary in terms of their effects in improving Taiwanese junior high school students’ performance on formal listening comprehension tests.. . Whether scripted and unscripted listening materials vary in terms of their 8.

(19) effects in promoting Taiwanese junior high school students’ understanding of native speakers’ real-life conversation in the real-world setting. 1.4 Significance of the study In order to provide a clear picture of the value of “unscripted text,” more studies need to be performed to investigate the comparative effectiveness of “scripted” and “unscripted” text. The present study aspires to shed new light on the value of unscripted text. If “unscripted text” effectively benefits learners’ listening comprehension ability, it would possible to integrate unscripted text into the current formal assessment so that the formal assessment might demonstrate the true comprehension ability of learners. What’s more, the nature of “unscripted text” is much like natural speech or talking, which may make the listening material used in the classroom setting more authentic and natural.. 9.

(20) CHAPTER TWO LITERATURE REVIEW In most foreign language classroom settings, scripted text is commonly used as the listening teaching material to improve students’ listening ability and assess their listening comprehension (Wagner, 2013). The prevalent use of scripted text as the listening instructional material has led to the issue that assessment scores do not truly reflect students’ listening proficiency, and students are not capable of understanding what foreigners say in real-life speech. Numerous studies have discovered that the problems might be caused by the huge gap between the spoken language used in textbooks and natural speech (Flowerdew & Miller, 1997; Gilmore, 2004; Jung & Choi, 2008; Chan, 2013). This chapter describes the differences between scripted and unscripted text. The following section reviews the relevant literature of empirical studies that compared listeners’ performance under teaching using different types of text. Then, based on the Transfer Appropriate Processing input hypothesis, the last section focuses on the potential significance of unscripted text in promoting students’ listening competence in the foreign language classroom. 2.1 The differences between scripted and unscripted text It is widely accepted that spoken language is qualitatively different from written language (Tannen, 1982; Halliday, 1985). As Wendy and Lam (2002, p.250) pointed 10.

(21) out, “spoken language is not written language spoken aloud,” but has its own features and structures that vary from written language. This study mainly focused on spoken language. Even in spoken language, with various auditory features, there are different degrees of ‘orality.’ These different degrees of ‘orality’ will govern the extent to which the text is similar to spoken text. Tannen (1982) first coined the term, ‘a continuum of orality’ to describe the extent to which spoken and written text intersect. Specifically, ‘a continuum of orality’ is a spectrum, with one end representing text that is less planned with more oral features, and the other end representing text with more written and planned forms. In a similar vein, McCarthy and Carter (1995) analyzed the corpus collected at the University of Nottingham, and drew the conclusion that text can be described as being more ‘speakerly’ or ‘writerly’ than other text. In this study, more ‘speakerly’ text was defined as ‘unscripted text’, in which native speakers naturally talked about an assigned topic without simplifying the sentence patterns or decreasing the speaking speed. In contrast, more ‘writerly’ text was referred to as ‘scripted text’, where native speakers read aloud the simplified written text with clear pronunciation and intonation. The purpose of this study was to compare the influence of these two different types of text on EFL learners’ listening comprehension. The difference between scripted and unscripted text is that unscripted text is 11.

(22) similar to real-life speech, containing many auditory features such as speaker accent, hesitations, pauses, noise, distortion and a high speech rate. Tannen (1982) indicated that these elements of unscripted text might benefit learners’ listening comprehension. Griffiths (1991) called these elements ‘hesitation phenomena,’ including silent pauses, filled pauses, hesitations, false starts, and redundancies. These elements are considered to be of significance in facilitating listening comprehension; whether listeners are able to grasp these elements is regarded as critical in rendering the spoken message comprehensible. In the language classroom, most EFL learners are commonly exposed to more ‘writerly,’ dense and well-structured listening input, which is far from the loose, flowing, unscripted input people listen to in real life. From the materials development perspective, several studies have discovered that the features of unscripted text are virtually nonexistent in textbooks (Jung & Choi, 2008; Chan, 2013; Revesa & Brunfaut, 2013). Most listening materials that teachers use in the classroom setting are scripted text. It is important to explore whether using scripted or unscripted text in EFL classes has a beneficial effect on learners’ listening comprehension. 2.2 Existing empirical studies on the performance of EFL learners using scripted and unscripted text Most research and theory has seemed to only explore the difficulties that 12.

(23) unscripted text poses to learners or the effectiveness of use of authentic materials in the classroom setting. Few studies have investigated the potential effectiveness of unscripted text and/or focused on comparison of the effectiveness of scripted and unscripted listening materials. Described below are three empirical studies that compared the use of scripted and unscripted text. Gilmore (2011) conducted a classroom-based longitudinal investigation in an attempt to assess the potential of use of authentic materials to develop Japanese learners’ communicative competence in English. Participants were randomly assigned to a control group, which listened to textbook input only, and to a treatment group, which mainly received authentic input over a 10-month period. At the end of the experiment, the communicative competence of the participants was measured using eight different types of test. The results indicated that the treatment group outperformed the control group in five of the eight measures; there were no significant differences between the two groups in the other three measures. Although the study succeeded in proving that authentic materials and related tasks were more effective than textbook materials for the development of learners’ communicative skills, there remains a need to closely examine the method section. The measurements included 8 different kinds of test. Participants improved in 5 measures in terms of listening proficiency, receptive comprehension of phonological features and 13.

(24) vocabulary, oral interview and student role-play; however, more work is needed to examine the connection between the teaching in English courses and the measurements in post-course tests. During the 10-month period, the authentic listening input that the participants received was mostly taken from films, documentaries, reality shows, and Web-based resources, which differed from the material the participants studied in the post-course tests, which were well-scripted and carefully designed. Read (2002) investigated how EFL learners perform under different forms of input in listening assessments. Participants in the study were adult non-native speakers of English taking a 3-month intensive English course at Victoria University of Wellington. For the purpose of the study, two audiotape listening resources were created: one was a scripted monologue, and the other an unscripted discussion on the same topic by three speakers. The participants were randomly assigned to the two listening resource groups: monologue and discussion. The results showed that the monologue group significantly outperformed the discussion group. The monologue version of the test turned out to be less difficult than the discussion one, and the discussion version was harder for learners to comprehend. Though Read discovered that the scripted version was easier for learners to comprehend, he indicated that different types of listening input might lead to various results. In this study, learners 14.

(25) became used to listening to scripted written text of a relatively formal style, without informal speech being included, and they therefore considered the discussion version difficult to comprehend. Also, in the discussion version, the fact that the speakers took turns to speak might have added difficulty in comprehension. What’s more, there were two variables that might have assisted the monologue group to acquire higher scores: one was the practice effect, the other the relationship between the two forms of stimulus material and the test items. Participants had taken similar tests that also involved listening to a scripted talk, and the items used in the assessment had a closer relationship to the monologue script than to the discussion. Wagner and Toth (2014) compared the performance of L2 Spanish learners on a listening comprehension test using scripted versus unscripted text. For the unscripted version of the text, two native Spanish speakers were instructed to speak as naturally as possible while recording two role-plays that were used for the listening comprehension test. On the other hand, the scripted text was read by the same speakers, but with all the naturalistic speech features removed. The participants, after undergoing teaching using the two types of speech, took two listening comprehension tests and completed a questionnaire that enquired about the spoken text used in the tests. The results showed that group of test-takers who heard the scripted text scored significantly higher on the comprehension test than the group of test-takers who heard 15.

(26) the unscripted oral text. However, there was no significant difference in the mean score between the participants who undertook the two different listening tests. The reason for this might have been the insufficient duration of learning. Also, two tests were possibly not sufficient by which to accurately assess learners’ listening ability. The above-described studies (Gilmore, 2011; Read, 2002; Wagner & Toth, 2014) offered some valuable information that could be applied to further study. First, there remains a need to develop an efficient method that can reinforce the connection between learning and measurement. Second, researchers should be aware of the comparability of listening texts and the design of listening tests, and designers should avoid inserting exactly the same terms as in scripted texts. Finally, it is necessary to increase the number of measurements in order to assess learners’ true proficiency. It would be possible to acquire different results if the experiments were revised. 2.3 Current use of unscripted text in the EFL classroom The first section of this literature review mentioned the differences between scripted and unscripted text. One difference is that unscripted text contains more natural linguistic features, such as silent pauses, filled pauses, hesitations, false starts, and redundancies, which Tenan (1982) referred to as ‘hesitation phenomena.’ Hesitation phenomena affect the way in which listeners process spoken language (Corley, MacGregor & Donaldson, 2007). Numerous studies have shown their 16.

(27) positive effects on learners’ listening comprehension (Latifi, Youhanaee & Mohammadi, 2013; Clark, 2014; Rahmani, 2015). For example, Lam (2002) stated that pauses and fillers offered more processing time for L2 learners, and so the two features are beneficial to their listening comprehension. Similarly, Gilmore (2004) also suggested that dialogue should be integrated with more natural discourse features. This would render unscripted conversations more authentic, which might familiarize students with the oral features required in a real-life communicative situation. Nevertheless, the use of unscripted text in the language classroom is very limited. There are several reasons for which unscripted text is not typically included in textbooks or used by teachers in classrooms. According to Gilmore (2007), textbook writers tend to write books using their own intuition instead of examining spoken corpora. What’s more, writers focus more on linguistic rules than pragmatic usage. Conflict between linguistic researchers and language practitioners is another reason that has kept “authentic materials” from being integrated into textbooks (Gilmore, 2007). For instance, Richards (2006) discussed “the myth of authenticity” (p.16), indicating his opposition to the use of authentic text. Writers might have difficulty in selecting appropriate unscripted text and editing related tasks for learners. However, it is easier and more efficient for textbook writers to come up with tasks first and then write the scripted text. 17.

(28) In addition to the textbook issues noted above, learners’ preferences and attitudes also play important roles. Peacock (1997) stated that learners showed more interest in artificial materials. In Kmiecik and Barkhuizen’s study, they examined learners’ attitudes towards learning materials, and found that learners had a more positive attitude toward “unauthentic” materials (2006). The learners’ negative attitude towards authentic materials might be due to the greater speed or more difficult vocabulary embedded in unscripted text. Guarieno and Morley (2001) were concerned that authentic materials would frustrate lower-proficiency students and make them less motivated towards learning. Teachers do not want to scare students away, either, especially those at a lower level of ability, and so they are more inclined to use scripted text as their listening material. Although the above-described studies supported the use of scripted text as the listening material, there is some evidence of flaws of using scripted text. Vandergrift (2007) discovered that people without linguistic training would not “notice” the discourse features in unscripted text, and this is especially true for students who are taught using artificial and unnatural scripted text. Vandergrift (2007) also indicated that scripted text might hinder learners’ development in word-segmenting skills, which refers to segmenting oral input into meaningful units. In light of the above, Vandergrift posited that it is necessary to expose learners to characteristics of 18.

(29) unscripted spoken text, even though those elements are the exact difficulties that impede learners’ comprehension (Vandergrift & Goh, 2012). Herron and Seay (1991) investigated the effect of authentic, unscripted radio features on students’ listening skills in the foreign language classroom. In the experiment, the control group and the experimental group were taught using the same instrument. The only difference was that the experimental group received additional authentic input; i.e., the participants in the experimental group undertook an extra one-and-a-half classes consisting of listening to authentic materials, while the control group did not. The results confirmed that the learners’ listening comprehension improved with increased exposure to authentic speech. Another interesting finding was that adjusting levels of speech to meet students’ proficiency is not critical in improving students’ listening abilities. In summary, authentic input benefits learners’ listening comprehension. If authentic materials can provide learners with a more natural listening input, it is necessary for teachers to make good use of unscripted text in classes and to enhance students’ exposure to the auditory features in unscripted text where appropriate. Rashtchi and Afzali (2011) conducted a study to explore whether spoken grammatical features affected learners’ comprehension of real-life conversations. 45 EFL university students in Iran were divided into two groups. Both groups listened to the 19.

(30) same authentic materials. The experimental group took the listening comprehension test under instruction to address the spoken grammatical features, whereas the control group took the test without focused instruction regarding the spoken grammatical features. The results showed that the experimental group scored significantly higher than the control group, which indicated that drawing attention to the spoken grammatical features affects learners’ comprehension of real-life conversations. In brief, the purpose of learning a language is to be able to communicate with native speakers in the real-world setting. As using textbooks as listening materials does not effectively equip learners with good communication abilities, it would be worth attempting to train language learners using authentic materials instead of textbook materials (Gilmore, 2011). In addition, the studies above implied that extra exposure to authentic input is necessary and useful for improving learners’ listening comprehension in scripted listening comprehension tests (Herron and Seay, 1991; Rashtchi and Afzali, 2011); however, there is a lack of studies investigating whether learners’ performance is improved in unscripted listening comprehension tests following unscripted listening input in class. In addition, due to several reasons, teachers have become used to using scripted dialogue as the listening material in class; few teach listening using unscripted text. More studies are warranted to investigate students’ performance after being taught using unscripted text in listening classes. 20.

(31) 2.4 Transfer Appropriate Processing In this study, the hypothesis was postulated that the use of unscripted listening materials in the classroom setting would expose EFL learners to conditions that are closer to real-life situations. The study assumed that EFL learners might be more capable of understanding native speakers’ real-life conversations if they are also involved in unscripted listening input in classes; i.e., learners might perform best when the learning conditions are similar to real-life conditions. This assumption is in line with the Transfer Appropriate Processing input hypothesis. Transfer Appropriate Processing (TAP) is a concept of cognitive processing based on the idea that successful memory retrieval backs up the encoding process whenever an individual operates the two processes in a similar way. That is, the relationship between how information is initially encoded and how it is later retrieved is crucial to the way in which learning occurs and how it can be accurately assessed. In the field of SLA, communicative language teaching and form-focused instruction are consistent with TAP as a framework for matching classroom learning processes using situations outside the classroom. Specifically, if learners’ learning process involves content-based learning or engagement in natural communicative interaction, their learning might be more effective, because the learning situations are close to real-life situations. 21.

(32) The study of Martin-Chang and Levy (2006) proved the validity of TAP. The study compared the effects of context training and isolated word training on subsequent measures of word-reading fluency. The results indicated that the participants performed better following isolated word training. The findings from the experiment supported the tenet of TAP that reading performance receives maximal benefit after isolated word training as compared with context training, because the task performed in the experiment involved reading words in isolation. This hypothesis explains the mismatch between the conditions under which learning takes place and those in which it is then applied. Most of the time, learners are not capable of utilizing the knowledge they have acquired in class when they encounter new situations. For example, if learners do not retrieve language from memory in a communicative context, they will possibly experience difficulty in using language in a real conversation outside the classroom. The problems might be due to disparity between the learning process and the situations in which the learning is used. TAP suggests the importance of increasing the numbers of settings and processing types that learners will encounter in teaching material. The concept of matching the classroom learning process to the process that is utilized outside the classroom is key in the current study. Previous research has focused only on the essence of teaching itself or on learners’ performance in tests. 22.

(33) Few researchers have examined the connection between the teaching students receive in an English course and the tests they take for assessment. Thus, in order to strengthen the relationship between the listening input in the classroom setting and the conditions of the tests, the approach used in this study was to involve participants in both scripted and unscripted listening exercises; in addition, the tests that the participants took were presented in two versions, scripted and unscripted. It is interesting to examine whether listeners’ performance improves owing to the similarity between natural speech and unscripted text. 2.5 Research questions 1. Does EFL learners’ listening comprehension ability correlate with the listening input they receive in the classroom setting? 2. Are unscripted listening materials effective in improving EFL learners’ listening comprehension in terms of scripted and unscripted tests? 3. What are the relative contributions of the two types of listening input to EFL learners’ listening comprehension ability? 2.6 Summary The prevalent use of scripted text does not necessarily indicate the uselessness of unscripted text for language teaching. Although empirical studies have shown that learners subjected to unscripted text achieve statistically lower scores than those 23.

(34) taught using scripted text, this does not mean that unscripted text should be avoided in classes. In addition, a simplified version of the learning material may not always be optimal for learners in terms of increasing their language proficiency. Sometimes, it is more beneficial for learners to directly face the complexities of natural discourse in order that they might be better equipped to communicate in real-life situations outside the classroom. From the perspective of TAP, the similarity of the learning process and the situation in which learning is applied plays an important role in determining learners’ listening comprehension ability. In light of the points mentioned above, the current study attempted to compare students’ performance in “scripted” and “unscripted” listening comprehension tests after receiving unscripted listening input in the classroom setting and shed new light on the potential value of use of unscripted text for teaching purposes.. 24.

(35) CHAPTER THREE. METHODOLOGY 3.1 Participants The participants in the current study were 70 students (45 male and 25 female) of English as a foreign language at a public junior high school in Taipei city. Four classes were included in the experiment, which were chosen for several reasons: first, all participants were 8th graders, of an average age of fourteen years; second, the students were evenly assigned to each class through the entrance exam when they were in 7th grade, so the academic performance of each class was homogeneous; third, according to the teachers’ expert judgment, the characteristics of the four classes were quite similar; and fourth, the average scores of the classes in their mid-term examinations ranged from 62 to 65. 3.2 Materials All of the materials used were created specifically for this study. There were two versions of the listening material, one scripted, the other unscripted. Both versions of the listening text were based on the same content in the textbook. The textbook that the participants utilized in the English class was published by Hanlin Ltd. (Book 3). In order to avoid a practice effect, the content of Book 3 of the textbook series published by Kang-Hsuan Ltd was chosen as the listening material for this study. This 25.

(36) methodological decision was made because the two sets of textbooks are similar in the following aspects. First, they are similar in terms of content: the texts that the listening materials were based on both consisted of 100 to 150 words. Second, the two books both follow the same grammatical focus on the past tense, subordinate clause, and future tense. Third, both editions follow the curriculum guidelines published by the Taiwanese Ministry of Education in 2008, which require that all lessons should be based on the most frequent 1200–2000 word list. Finally, both textbooks deal with similar topics that were familiar to the participants in the study, including school life, personal relationships, and the environment.. 3.2.1 Listening materials Unscripted spoken text Unscripted listening texts used in this study were recorded by two English teachers. Before recording, speakers read the dialogue in each lesson (see example 1) and knew what kind of scenery they should create. In order to ensure that the speakers used words that were level-appropriate for the participants, words lists for each lesson were provided. Then, after quickly browsing through the content of a given lesson and the word list, the two speakers conversed with each other without referring to the textbook content. They were advised to speak as naturally as possible, using all sorts 26.

(37) of conversation strategies that they would use in a daily real-life conversation. An example of an unscripted conversation that the speakers created is shown in example 2.. Example 1 – Listening material from the Kang-Hsuan textbook <Book 3 – Lesson 1 > (On the street) Carl: Hey, Sam! How was your summer? Sam: It was great! Carl: What did you do? Sam: I joined a basketball camp and learned some useful skills. Carl: Cool! Show us next time. Sam: No problem. Oh, guess what! Jeremy Lin visited our camp yesterday and played a game with us. It was really exciting! Carl: Wow, lucky you.. Words <camp/ useful/ skill/ visit/ yesterday/ lucky you>. Example 2 – Unscripted spoken text <Unscripted Dialogue> A: Hi! Emily <1>, how <1> are you? B: <1> Hey. <1> B: I’m good. How are you, Winnie? A: I’m good (.) so (.) how was your summer? B: Uh (.) summer (.) it was -- it was fun, actually. A: So what did you do <2>? B: <2> It was fun. <2> Uh (.) so my parents have a tradition, every summer they would usually take me and my brother to the State so that we would join some camps. And this summer, they decided to pick a camp that -- that has something to do with basketball. A: Basketball=! That sounds interesting! So what did you do in the camp? B: Eh (.) I’m really a ‘camp’ person. 27.

(38) A: Did your brother go with you? B: He = refused to come with me in the beginning = A: = So you went there alone? B: So I flew there alone. A: OK. B: My brother was supposed to come afterward. A: OK. So you two were both in camp. <3> A basketball camp? B: <3> mhm <3> Yes. That’s strange. I mean, Yeah (.) I’m the girl. Why would my parents pick a (( )) camp. (.) <4> <@@> <5> I don’t like basketball. <5> I’m always playing tennis. And my brother likes basketball. But – but (.) I prefer tennis. So I have no idea why my parents pick a basketball camp. But one of the highlight was Jere- +Jeremy Lin actually visited our camp. A: <4> I = have no idea. <@@> <4> Maybe they think that you need something (.) like that sport (.) <5> Oh my gosh! He visited the camp?. Scripted spoken text The scripted spoken material was based on the unscripted material. After transcribing the unscripted conversation of each lesson, the researcher removed all the ‘noise’,. including. hesitations,. fillers,. pauses,. redundant. words. and/or. repeated/elaborated information, producing a ‘clean’ and concise conversation script to be used/read by the same speakers (see example 3). Two speakers were instructed to read aloud the script with clear pronunciation and intonation. To make the scripted spoken text ideal for the listeners, the speed of speaking was reduced by the speakers.. Example 3 – Scripted spoken text <Scripted Dialogue> A: Hi! Emily, how are you? B: I’m good. How are you, Winnie? 28.

(39) A: I’m good. How was your summer? B: It was fun! A: So what did you do? B: My parents have a tradition. Every summer, they would usually take me and my brother to the State so that we would join some camps. And this summer, they decided to pick a camp that has something to do with basketball. A: Basketball! That sounds interesting! B: I’m really a ‘camp’ person. A: Did your brother go with you? B: He refused to come with me in the beginning. A: So you went there alone? B: So I flew there alone. A: OK. B: My brother was supposed to come afterward. A: OK. So you two were both in camp. A basketball camp? B: Yes. That’s strange. I mean, I’m the girl. Why would my parents pick a basketball camp? I don’t like basketball. I’m always playing tennis. So I have no idea why my parents picked a basketball camp. But one of the highlights was Jeremy Lin actually visited our camp. B: Oh my gosh! He visited the camp?. 3.2.2 Testing instruments Each listening comprehension test contained 6 comprehension questions based on 2–3 dialogue excerpts. Each comprehension question included three response options.. All the listening dialogues for the tests were excerpted from the listening. materials recorded earlier in both versions, scripted and unscripted. The sample of a scripted listening test format was shown in Table 1. And the sample of an unscripted listening test format was shown in Table 2. Both listening tests shared the same multiple choice questions, including the test stems and items.. 29.

(40) Table 1. The sample of scripted listening test A. The listening script <Scripted Dialogue> A: So what did you do last summer? B: Sad experience. B: Sad experience. I failed my math test. Oh! I hate math. Math was just not my thing. I don’t like those numbers at all. Yeah. Numbers. Tables. And figures. Just not my things. So I failed it. And I had to find some tutors but I couldn’t find a good tutor so I went to a cram school that my aunt suggested. I had to spend all my summer in the cram school. A: That sounds absolutely dreadful. And it’s worse than staying at home and doing nothing.. Multiple Choice Questions 1. How was Carl’s summer? (A) It was wonderful! (B) It was terrible! (C) It was cold! 2. Why didn’t Carl have a good time during the summer? (A) He had a bad cold in summer. (B) He studied in the cram school in summer. (C) He broke his legs in summer.. Table 2. The sample of unscripted listening test B. The listening script <Unscripted Dialogue> A: So what did you do last summer? B: Summer. Eh (.) Sad experience. B: Sad experience. I failed my math test. Oh! And Gee (.) I = HATE = math. Math was just not my thing. I don’t like those numbers at all. Yeah. Numbers. Tables. And figures. Just not my things. So (.) I failed it. And (.) I had to find some tutors but I couldn’t find a good tutor so I went to a cram school that my aunt suggested. <sighing> That I had to spend all my summer in the cram school, in the classes – the math class. So = <2> very much so <2> uh-huh (.) very much so (.) A: = That sounds ABSolutely dreadful. <2> And it’s worse than staying at home and <2> doing nothing.. Multiple Choice Questions 1. How was Carl’s summer? 30.

(41) (A) It was wonderful! (B) It was terrible! (C) It was cold! 2. Why didn’t Carl have a good time during the summer? (A) He had a bad cold in summer. (B) He studied in the cram school in summer. (C) He broke his legs in summer.. 3.3 Design Transfer appropriate processing suggests that whether learning is effective depends on how strong the connection between learning and using conditions is. Based on the TAP, the experiment intended to create two learning conditions and two using conditions. The purpose of this experimental design was to explore how learners who had been involved in two learning conditions performed in two using conditions, hoping to acquire the results that learners could perform better when involved in highly similar learning condition. In this experiment, the learning condition was set in English classes where the teacher implemented instructions to participants with scripted and unscripted listening input. In terms of using condition, the participants were required to take scripted and unscripted listening tests. As shown in Table 3, the participants were separated into four groups, A, B, C, and D group. Among four groups, group A and group B received ‘unscripted’ listening treatment in English instructions. After being treated with the same ‘unscripted’ listening input, group A (UU group) took ‘unscripted’ listening tests while group B 31.

(42) (US group) took ‘scripted’ listening tests. On the other hand, group C and group D received ‘scripted’ listening treatment in English instructions. After being treated with the same ‘scripted’ listening input, group C (SU group) took ‘unscripted’ listening tests while group D (SS group) took ‘scripted’ listening tests. Table 3. The design of grouping Group. Treatment (instruction). Testing. Group A (UU group). unscripted input. unscripted test. Group B (US group). unscripted input. scripted test. Group C (SU group). scripted input. unscripted test. Group D (SS group). scripted input. scripted test. The design of UU group and SS group offered a similar using condition as their learning conditions to the participants. In contrast, US group and SU group gave participants a different using condition from their learning environments. A questionnaire asking about two types of listening materials was used at the end of the experiment. In order to explore deeply about the listening process during the whole experiment, oral interviews were conducted as well.. 3.4 Procedure A summary of one set of experiments over a week is shown in Figure 1. 32.

(43) There were four periods of English classes per week. Participants received instruction in the first three periods of English classes, with no intervals between periods. In the last period of English class for a week, students took the listening tests. Figure 1. Summary of the experimental design over a week. UU group US group SU group SS group Class period. 1. 2. 3. 4. Instruction Unscripted materials. Testing Scripted materials. The whole procedure had lasted for 10 weeks. After 10-week experiment, the participants were requested to complete a questionnaire. (see Appendix A) In addition, to probe into listeners’ perception of listening materials, four participants from each group were randomly chosen to have two interviews with the researcher.. 33.

(44) CHAPTER FOUR RESULTS The aim of the current study was to explore whether “unscripted text” was beneficial in terms of improving learners’ listening comprehension ability in the foreign language (FL) context. The statistical data of this study contained four mean scores of the four groups (UU, US, SU, SS)1. The mean pre-test score was the first score to be collected. The other three mean scores were taken from the post-tests completed in the subsequent 9 weeks. The regular interval between the three post-tests was 3 weeks. In order to make sure that the participants concentrated on the listening, the participants in each group were asked to take a quiz at the end of the lessons every week. One-way ANOVA was conducted to test whether there were differences among the four groups. As shown in Table 4, the Scheffe method was employed, and no significant differences were identified between the four groups, suggesting that the scores of each group were nearly the same, if not identical, prior to the experiment (F(3,57) = 3.323, p = .026 < .05).. 1. The UU and SU groups took unscripted listening tests following unscripted and scripted listening input, while the US and SS groups took scripted listening tests following unscripted and scripted listening input, respectively. 34.

(45) Table 4. Summary of the results of one-way ANOVA at the first time point. Source of Variation. Sum of Squares. Degree Mean of Square Freedom. Treatments. 13.426. 3. 4.475. Error. 76.771. 57. 1.347. Total. 598.000. 61. F. p. Scheffe’ Score. 3.323. .026*. NS. *p < .05, NS = not significant. Two-way split-factor ANOVA Two-way split-factor ANOVA was performed to explore the main effects of the un/scripted listening input and the time factor, and to examine whether there was an interaction effect between un/scripted listening input and the time factor on students’ listening comprehension ability. As shown in Table 5, the main effect of listening type yielded an F ratio of F(3, 171) = 2.602, df = 3, p > .05, indicating that the effect of listening type on listening comprehension was not significant. However, the main effect for the time factor yielded an F ratio of F(3, 171) = 4.368, df = 3, p < .05, indicating significant differences in listening comprehension between post-test 1, post-test 2 and post-test 3. The interaction effect was not significant, F(9, 171) = 1.494, df = 9, p > .05, suggesting that both listening types had no interaction with the time factor.. 35.

(46) Table 5. Two-way split-factor ANOVA. Source of Variation. Sum of Squares. Degree of Freedom. Mean Square. F. p. Between groups. 238.094. 57. 4.177. A (Type of listening material). 32.607. 3. 10.869. 2.602. .061. Within groups. 345.597. 228. B(Time Factor). 8.238. 3. 2.746. 4.368. .005*. A*B. 8.452. 9. 0.939 1.494. .154. Errors. 107.503. 171. Total. 394.894. 243. 0.629. * p < .05. One-way ANOVA and paired t-tests One-way repeated-measure ANOVA was conducted in each group to explore whether there were any significant differences between the post-tests. In cases in which differences were found, a paired t-test was conducted. As shown in Table 6, the results of one-way ANOVA indicated that there was a statistically significant effect of the time factor in the UU group, F(3,45) = 3.885, p = .015, suggesting that the group that received unscripted listening input made progress in the unscripted listening tests in the subsequent three post-tests. A paired t-test was conducted, and it was found that there was a significant difference between the scores of the pre-test (M = 2.44, SD = .911) and post-test 2 (M = 3.13, SD = .975); t = –2.300, p = .036. In addition, a significant difference was found between the scores of the pre-test (M = 2.44, SD = .911) and post-test 3 (M = 3.59, SD = 1.255); t 36.

(47) = –3.306, p = .005. In summary, the students’ scores on the second and third post-tests were significantly higher than the score on the pre-test, meaning that the participants who received unscripted listening input made progress in the unscripted listening tests, starting from post-test 2. These results are summarized in Table 7. Table 6. One-way ANOVA for the UU group. Source of Variation Between groups. Sum of. Degree of. Squares. Freedom. Mean Square. 11.918. 3. 3.973. 28.277. 15. 1.885. Errors. 46.020. 45. 1.023. Total. 86.215. 63. F. p. 3.885. .015*. Within groups. *p < .05. A growing trend was observed in the UU group, which was also found in the SU group. Specifically, as shown in Table 8, a significant difference was found for the main effect of the time factor in the SU group (F = 3.815, p = .016), meaning that the group that received scripted listening input made progress in the unscripted listening tests in the subsequent three post-tests. A paired t-test was conducted, and it was found that there was a significant difference between the scores of the pre-test (M = 2.50, SD = 1.033) and post-test 1 (M = 2.94, SD = 1.209); t = –2.406, p = .016. In addition, a significant difference was found between the scores of the pre-test (M = 2.50, SD = 1.033) and post-test 2 (M = 2.94, SD = .834); t = –2.573, p = .021. Finally,. 37.

(48) a significant difference was found between the scores of the pre-test (M = 2.50, SD = 1.033) and post-test 3 (M = 3.25, SD = 1.252); t = –3.770, p = .002. In summary, the students’ scores on the three post-tests were significantly higher than the score on the pre-test, which showed that the participants who received scripted listening input made progress in the unscripted listening tests starting from post-test 1. These results are summarized in Table 9. Table 7. Paired t-test for the UU group. M Pre-test. 2.44. SD 0.911. Post-test 1. 2.75. 1.265. Pre-test. 2.44. 0.911. Post-test 2. 3.13. 0.975. Pre-test. 2.44. 0.911. Post-test 3. 3.59. 1.255. Post-test 1. 2.75. 1.265. Post-test 2. 3.13. 0.975. Post-test 1. 2.75. 1.265. Post-test 3. 3.59. 1.255. Post-test 2. 3.13. 0.975. Post-test 3. 3.59. 1.255. *p < .05.. 38. t. p –.850. .409. –2.300. .036*. –3.306. .005*. –1.168. .261. –1.870. .081. –1.392. .184.

(49) Table 8. One-way ANOVA for the SU group. Source of Variation. Sum of Squares. Degree of Freedom. Mean Square. 4.563. 3. 1.521. 53.938. 15. 3.596. Errors. 17.938. 45. 0.399. Total. 76.439. 63. Between groups. F. 3.815. p. .016*. Within groups. *p < .05. Table 9. Paired t-test for the SU group. M Pre-test. 2.50. SD 1.033. Post-test 1. 2.94. 1.209. Pre-test. 2.50. 1.033. Post-test 2. 2.94. 0.834. Pre-test. 2.50. 1.033. Post-test 3. 3.25. 1.252. Post-test 1. 2.94. 1.209. Post-test 2. 2.94. 0.834. Post-test 1. 2.94. 1.209. Post-test 3. 3.25. 1.252. Post-test 2. 2.94. 0.834. Post-test 3. 3.25. 1.252. t –2.406. p .029*. –2.573. .021*. –3.770. .002*. 0.000. 1.000. –1.071. .301. –1.576. .136. *p < .05. In terms of the US group, no significant differences were found in the subsequent post-tests, F(3,45) = 0.899, p = .449 > .05, as shown in Table 10.. 39.

(50) Table 10. One-way ANOVA for the US group. Source of Variation. Between groups. Sum of Squares. Degree of Freedom. Mean Square. 3. 0.566. 1.699. F. p. 0.899. .449. Within groups 96.340. 15. 6.423. Errors. 28.363. 45. 0.630. Total. 126.402. 63. *p < .05. Finally, as shown in Table 11, no significant differences were found in the SS group in the subsequent post-tests, F(3,45) = 0.207, p = .891 > .05.. Table 11. One-way ANOVA for the SS group. Source of Variation Between groups. Sum of. Degree of. Mean. Squares. Freedom. Square. 0.375. 3. 0.125. 62.000. 15. 4.133. Errors. 27.125. 45. 0.603. Total. 89.500. F 0.207. p .891. Within groups. *p < .05. To briefly recapitulate the results reported above, the participants who took scripted listening tests did not benefit either from unscripted or scripted listening input. In contrast, both scripted and unscripted listening input enhanced the participants’ performance in the unscripted listening tests. The participants’ performance implied a. 40.

(51) potential influence of unscripted listening tests. In addition, the group who received scripted listening input (the SU group) made progress starting from post-test 1. The group who received unscripted listening input (the UU group) began to score more highly from post-test 2. In terms of improvement rate, scripted listening input began to facilitate an improved performance of the participants three weeks after the start of teaching. Results of the questionnaire After completing the 10-week experiment, the participants completed a 17-item questionnaire enquiring about the spoken texts used in the tests.( see Appendix B) How did the participants evaluate the two listening texts? In terms of the difficulty of the listening texts and the length of the dialogues, the participants considered both scripted and unscripted texts appropriate (Q3 、Q4). Speakers’ speaking speed in the unscripted texts was considered faster than that in the scripted texts (Q5). No participants thought that the speakers expressed the text unclearly in the scripted text, while a few considered that the unscripted text was unclear (Q7). How did the listening texts influence the participants? In terms of how interesting the content was to the listeners, more participants considered the unscripted text interesting (Q8) and increased their motivation for listening (Q9、Q10). In addition, more participants thought that the interaction between speakers in the 41.

(52) unscripted text was natural (Q13). More than half of the participants indicated that they did not lose patience when listening to both types of listening text. Did the participants perceive any disparity between the scripted and unscripted text? Most participants were capable of distinguishing differences between the two listening texts (Q14、Q 15). More than half participants preferred the unscripted text to the listening text contained in their textbooks (Q16). And more participants strongly agreed that the unscripted text was easier to comprehend than the text contained in their textbooks (Q17).. 42.

(53) CHAPTER FIVE DISCUSSION The purpose of this study was to conduct regular tests to explore whether “unscripted text” is beneficial in terms of improving learners’ listening comprehension ability in the foreign language (FL) context, where scripted text is the mainstream teaching tool and native input is usually very limited beyond the classroom setting. The following section discusses how Taiwanese EFL learners perform in listening comprehension tests when presented with these two types of listening material (scripted vs. unscripted text). 5.1 Interpretation of the results based on TAP Based on the hypothesis of transfer appropriate processing (TAP), the participants who were trained with scripted/unscripted listening input and took scripted/unscripted listening tests (SS/UU groups) were hypothesized to show an improved performance, while the participants who received scripted/unscripted listening input and took unscripted/scripted listening tests (SU/US groups) were not hypothesized to show significant progress in their performance. However, the results showed that only the groups who took unscripted listening tests (UU/SU groups) had made progress during the 10-week experiment, whereas the other groups who took scripted listening tests (SS/US groups) exhibited a constant level of performance. 43.

(54) Why didn’t these two specially-designed listening materials have significantly different effects on listeners’ performance? Initially, the original focus of TAP was that learners might acquire effective learning if language instructors could narrow the gap between the encoding process in learning and the retrieval process while using. If the classroom environment only offers language learners teaching in a simplified language context, learners might have difficulty using language in complex communicative situations outside the classroom. To solve the problem of a mismatch between learning and using conditions, TAP suggests increasing the number of settings that are similar to real-life language conditions. Instead of implementing repetitive drills or imitating memorized dialogue, instructors have been advised to utilize form-focused instructions such as communicative language teaching and content-based instruction in language classrooms. To be more specific, comprehension training will improve comprehension ability, while speaking practice is necessary to cultivate speaking ability. In this study, learners practiced listening with scripted and unscripted listening materials. No matter which type of listening input the learners were exposed to, listeners were offered opportunities to practice both encoding and retrieving information in the listening process. In other words, as long as students were trained using English input, progress was made in their listening comprehension ability. The 44.