情境學習任務的涉入程度對單字學習之影響

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. 情境學習任務的涉入程度對單字學習之影響. The Influence of Involvement Load in Contextualized Learning Tasks on Incidental Vocabulary Acquisition. 指導教授：朱錫琴 Advisor: Dr. Hsi-Chin Chu. 研究生：顧玲慎 Ling-Shen Ku. 中華民國一百零七年九月 September, 2018. i.

(2) 中文摘要這份研究主要在探討三種不同的學習任務，以及學習者的閱讀能力，是否會影響台灣國中生的偶發性單字習得以及單字保留。根據涉入程度假說，當學習者在執行涉入程度較高的學習任務時，學習者可以在單字學習上有更好的效果，因此本研究採用三種不同情境學習任務：閱讀理解、閱讀填空，以及圖片寫作，來驗證涉入假說的預測性，另一方面也探討閱讀能力是否影響學習任務所引發的效果。七十二位台灣中部的國三生參與這份研究。在實驗開始之前，研究者對他們施行了閱讀能力測驗以及前測。前測主要在確認受試者對於 18 個目標單字並不具備有先備知識；根據閱讀能力測驗的結果，所有的受試者被分布在三個學習任務中。受試者分成三階段來學習十八個單字，每一階段在學習完六個單字之後，會有一個立即後側，而五天之後，會再分三次執行延遲後測。立即後測的量化分析顯示，閱讀能力對偶發性的單字習得有顯著的效果，而學習任務的涉入程度並未帶來顯著的影響，閱讀能力及學習任務之間也並未產生顯著的交互影響。雖然閱讀能力對單字習得具有極大的影響力，但是學習任務也有一定程度的影響。描述統計顯示，在執行圖片寫作任務時，受試者產生較多的單字學習，特別是閱讀能力較弱的族群，他們透過圖片寫作所學習的單字量，和閱讀能力較強的族群在進行閱讀理解及閱讀填空時所學得的單字，不相上下。研究者在延遲後測二因子變數分析上所得到的結果和前測所產生的結果相似：閱讀能力有顯著的效果，但是，學習任務的涉入程度並未有顯著效果，閱讀能力和學習任務之間也未呈現顯著的交互作用。描述統計顯示，圖片寫作所帶來的單字學習優勢在延遲後測中消失了，特別是閱讀能力較弱的族群，他們原先在圖片寫作的單字學習，和閱讀能力較強的學生在閱讀理解及閱讀填空上的表現相當，但是在延遲後測中，他們的單字保留表現，和閱讀能力較弱的人在其他兩組的表現雷 i.

(3) 同。這些對於情境學習任務的研究發現，在教育上具有它的意義。首先，不管是接收式或產出型的學習任務都有助於偶發性的單字學習，雖然圖片輔助的句子寫作有利於單字的立即學習，但是這樣的學習任務所帶來的影響力並未持續。此外，在研究學習任務的涉入程度對單字學習所帶來的影響力時，還有其他因素需要納入考量：學習任務的所需時間、單字的提取次數、學習者的閱讀能力及學習者的先備知識。. 關鍵字：涉入程度、學習任務、偶發性單字學習. ii.

(4) ABSTRACT The aim of this study is to explore the efficacy of three distinct contextualized learning tasks, the reading-comprehension task, the gap fill-in task, and the picture-writing task, and reading proficiency on the incidental vocabulary acquisition and retention of junior high school students in Taiwan. The three learning tasks are formulated based on the Involvement Load Hypothesis, which assumes that the task with higher involvement load will facilitate more vocabulary learning. Based on this assumption, the incidental vocabulary acquisition from the picture-writing task is assumed to outperform than the gap fill-in task. The reading-comprehension task will elicit the least word learning among the three learning tasks. Seventy-two EFL ninth graders from three classes in a junior high school in central Taiwan participated in this study for two weeks. Prior to the experiment, a reading proficiency test to check students’ reading ability and a pretest to check students’ knowledge of 18 target words were administered. Based on the results of reading proficiency test, the participants were assigned to three task groups, each studied 18 words in three sessions via one of the three learning tasks respectively with the effect of session order and embedded text counterbalanced. In each session, 6 words were studied followed by an immediate posttest. Five days later, a delayed posttest was carried out. The results from ANOVA analysis on immediate posttest indicated that while reading proficiency reflected a significant effect, learning tasks did not, nor did interactive effect of learning task and reading proficiency. However, reading proficiency revealed a large effect size and learning task did reveal a moderate effect size. Descriptive statistics showed that PW Group, especially the subgroup of iii.

(5) lower-proficiency readers, yielded a markedly higher mean score than the other two groups, who scored similarly. ANOVA on the delayed posttest showed the same pattern of significant test as on immediate posttest: significant effect of reading proficiency, insignificant effect of task and insignificant interactive effect between learning task and reading proficiency, and the effect size of task was minimal and reading proficiency was small. When mean scores were examined, it was found that the superiority of PW Group waned, especially for the lower-proficiency subgroup, which showed in immediate posttest an equivalent mean to those in the higher-proficiency level in the other two task groups but decreased to a level similar to those of lower proficiency readers in the other two tasks. The findings from this study had a pedagogical implication that contextualized learning tasks, be it receptive or productive, may all be conductive to promote incidental vocabulary acquisition. Although sentence production with story-picture cues, seemed to instigate more word acquisition at the completion of task, with one-round of exposure-production, the retention for whatever type of task may not hold. It is therefore suggested that further studies testing the effect of task involvement load should consider different weights of the essential component, evaluation, in the Involvement Load Hypothesis and other task variables, such as time spent on task, chances of retrieval, as well as the individual factors, such as learners’ reading proficiency and knowledge.. Key words: Involvement Load Hypothesis, learning task, incidental vocabulary acquisition. iv.

(6) ACKNOWLEDGEMENTS I would like to acknowledge the enormous suggestion and encouragement from my research advisor, Dr. Chu Hsi-Chin. Without Dr. Chu’s assistance, my proposal hearing could not be conducted smoothly. Without Professor Chu’s suggestions, the design of my experiment would not be creative and practical. With Professor Chu’s reminders, the administration of my experiment was conducted more deliberately. My gratitude also goes to Dr. Tseng Wen-Ta for his advice on the theoretical justification and the encouragement, and Dr. Wu Mei-Cheng for her advice and suggestions. In addition, I would like to show my gratitude to the administrative specialist, Wang Mu-Han, for all the assistance from the application of the proposal hearing to the final defense. My thankfulness goes to two of my colleagues, Jane and Jamie. They spared their precious class time to conduct my experiment. Without them, my experiment could not be completed within two weeks. My gratefulness also goes to my friends around me. Thank you for all the support and the thoughtfulness in different aspects. Last but not least, I feel apologetic to my family. During my thesis writing, I spared no time for visiting my parents. I squeezed little time to accompany my two dear daughters, Sophia and Zoe. You two always tried hard to enjoy yourselves and not to interrupt me. Thanks also goes to my husband for taking care of most of the housework. Thank you all, with all my heart.. v.

(7) TABLE OF CONTENTS 中文摘要 ................................................................................................................... i ABSTRACT............................................................................................................. iii ACKNOWLEDGEMENTS ...................................................................................... v TABLE OF CONTENTS ......................................................................................... vi LIST OF TABLES ................................................................................................... ix LIST OF FIGURES .................................................................................................. x CHAPTER ONE INTRODUCTION ........................................................................ 1 Background and Motivation ............................................................................. 2 Purpose of the Study ......................................................................................... 3 Definition of Terms ........................................................................................... 4 Incidental learning .................................................................................... 4 Incidental vocabulary learning.................................................................. 4 Involvement Load Hypothesis .................................................................. 4 Significance of the Study .................................................................................. 5 Organization of the Study ................................................................................. 5 CHAPTER TWO LITERATURE REVIEW ............................................................ 6 Incidental Learning ........................................................................................... 6 Incidental Vocabulary Learning ........................................................................ 7 Incidental Vocabulary Acquisition through Reading ................................ 7 Incidental Vocabulary Acquisition via Performing Tasks ....................... 10 Involvement Load Hypothesis ........................................................................ 11 Previous Vocabulary Studies Applying the Involvement Load Hypothesis ... 17 CHAPTER THREE METHODOLOGY ................................................................ 21 Participants...................................................................................................... 21 vi.

(8) Materials ......................................................................................................... 21 Selection of Reading Texts and Target Words ........................................ 21 Three Learning Tasks .............................................................................. 23 The Reading-Comprehension Task (hereafter RC Task) ........................ 23 The Gap Fill-in Task (hereafter GF Task) ............................................... 23 The Picture-Writing Task (hereafter PW Task) ....................................... 24 Instruments...................................................................................................... 26 The Reading Proficiency Test ................................................................. 26 The Pretest .............................................................................................. 26 The Immediate and Delayed Posttests .................................................... 27 Pilot Study....................................................................................................... 28 Design and Data Collection ............................................................................ 29 Initial Group Assignment ........................................................................ 29 Ultimate Group Assignment ................................................................... 30 Design of Text and Order ........................................................................ 32 Data Collection Procedure ...................................................................... 33 Scoring .................................................................................................... 35 Data Analysis .......................................................................................... 36 CHAPTER FOUR RESEARCH RESULTS ........................................................... 37 The Preliminary Analysis................................................................................ 37 Test of Assumption for Latin Square Design: Order Effect and Text Effect .. 37 The Primary Analysis...................................................................................... 38 Effect of Task on Immediate Posttest ..................................................... 39 Interaction between Task and Proficiency Level on Immediate Posttest 40 Effect of Task on Delayed Posttest ......................................................... 43 Interaction between Task and Proficiency Level on Delayed Posttest ... 44 vii.

(9) CHAPTER FIVE DISCUSSION AND IMPLICATION ........................................ 47 Discussion ....................................................................................................... 48 Effect of Task on Incidental Vocabulary Acquisition ............................. 48 Interaction of Task and Proficiency Level on Incidental Vocabulary Acquisition .............................................................................................. 52 Effect of Task on Vocabulary Retention ................................................. 53 Interaction of Task and Proficiency Level on Vocabulary Retention ..... 56 Pedagogical Implications ................................................................................ 56 Limitations and Implications .......................................................................... 57 On Design and Treatment ....................................................................... 57 On Results ............................................................................................... 59 Conclusion ...................................................................................................... 59 REFERENCES ....................................................................................................... 60 APPENDICES ........................................................................................................ 66 Appendix A. .................................................................................................... 66 Appendix B. .................................................................................................... 68 Appendix C. .................................................................................................... 69 Appendix D. .................................................................................................... 72 Appendix E. .................................................................................................... 75 Appendix F. ..................................................................................................... 78 Appendix G. .................................................................................................... 84 Appendix H. .................................................................................................... 90 Appendix I. ..................................................................................................... 91. viii.

(10) LIST OF TABLES Table 1 Task-induced Involvement Load (Laufer & Hulstijn, 2001, p.18) .................... 15 Table 2 Task-induced Involvement Load (Laufer & Hulstijn, 2001, p.18) (Continued) 16 Table 3 The List of Target Words .................................................................................... 23 Table 4 The Involvement Loads of Tasks ....................................................................... 25 Table 5 The Distribution of Participants ......................................................................... 30 Table 6 Number of Participants in Each Treatment Group, Subgroup, and Proficiency 31 Table 7 Three-Way ANOVA on Test of Homogeneity of Proficiency Level among Groups and Subgroups .................................................................................... 32 Table 8 The Text and Order under Each Task Group ...................................................... 33 Table 9 The Schedule of Procedure ................................................................................ 34 Table 10 Two-Way MANOVA on Test of Text Effect and Order Effect ........................ 38 Table 11 Descriptive Statistics of Immediate Posttest among Three Treatment Groups 39 Table 12 Descriptive Statistics of Immediate Posttest among Three Treatment Groups and between Two Proficiency Levels ............................................................. 40 Table 13 Interaction Effect between Treatment Group and Proficiency Level on Immediate Posttest .......................................................................................... 43 Table 14 Descriptive Statistics of Delayed Posttest among Three Treatment Groups.... 44 Table 15 Descriptive Statistics of Delayed Posttest among Three Treatment Groups and between Two Proficiency Levels .................................................................... 45 Table 16 Interaction Effect between Treatment Group and Proficiency Level on Delayed Posttest ............................................................................................................ 46. ix.

(11) LIST OF FIGURES Figure 1 Self-Report Categories in the VKS Elicitation Scale (Read, 2000, p.133) ...... 27 Figure 2 VKS Scoring Categories – Meaning of Scores (Paribakht and Wesche, 1997, p.180) .............................................................................................................. 36 Figure 3 Mean Scores of IP and DP among Treatment Groups between Two Proficiency Levels .............................................................................................................. 42. x.

(12) 1CHAPTER. ONE. INTRODUCTION As an English teacher in junior high school for many years, it has always been a task to facilitate the vocabulary acquisition and lengthen the vocabulary retention of English as a Foreign (EFL) learners. EFL learners need sufficient vocabulary knowledge both on the quality and the quantity to understand a paragraph or even to comprehend a reading text. It is also a challenge for teachers in the EFL classroom to perform adequate learning activities, especially on the vocabulary with the limited lecture time. During a 45-minute class time period, three times a week, teachers in junior highs school in Taiwan suffer the pressure of the time schedule. During each class time, it is almost impossible to manipulate diverse learning activities in each of the four skills, listening, speaking, reading, and writing, not to mention the sufficient time to focus on the vocabulary. However, learning a second or a foreign language involves the learning of large number of words (Laufer & Hulstijn, 2001). Thus, we are curious about the existence of a method where the language skills and the vocabulary learning can be learner simultaneously. Previous studies discovered that the vocabulary acquisition could happen incidentally when learners carry out a task (Hulstijn & Laufer, 2001). Different learning tasks contribute to the word acquisition to a diverse extent by EFL learners (Hulstijn & Laufer, 2001). Therefore, this study aims to explore the efficiency of contextualized learning tasks on the vocabulary acquisition and retention based upon the assumptions from the Involvement Load Hypothesis, and therefore, to help EFL teachers choose the most effective and suitable task for their beginning learners.. 1.

(13) Background and Motivation In recent years, vocabulary has been considered to play a more central role in second language(L2) learning than was traditionally assumed to (Richards, J.C. & Rogers, T. S., 2003). However, it is commonly discovered that memorizing new English words is difficult for the L2 or EF leaners, especially the junior high school students in Taiwan. English teaching in elementary school focuses on listening and speaking skills while spelling and memorizing a new word is not a crucial target. Meanwhile, in junior high school, the English learning environment switches its focus to reading and writing abilities. Then memorizing new words becomes the first priority because the students need adequate vocabulary size to become a fluent reader and sufficient word knowledge to make a sentence or write a short passage. Even if memorizing the new words is an easy job for the language learners, it may be a challenge for them to maintain the spelling and the meaning of a new word in mind. Maintaining a new English word needs sufficient stimulus from the learning activities and repeated contact with the new word; however, the time constraint of English class at school restricts the adequate classroom activities experienced by the learners. It is almost impossible for the English teachers to conduct various and sufficient learning activities; therefore, they can only choose tasks that isolated target words for learning. Thus, the question that English teachers ask most frequently is, “What is an effective task or a conducive activity to practice the four skills and facilitate the vocabulary acquisition simultaneously?” Previous studies pointed that tasks were proved to be beneficial for the vocabulary acquisition, for example, looking-up words in the dictionary (Luppescue & Day,1993; Knight, 1994), marginal glosses (Hulstijn, Hollander & Greidanus, 1996), word-focused tasks (Paribakht & Wesche,1997), interactional tasks (Atay & Kurt, 2006), and 2.

(14) negotiation (Newton, 1995). These tasks were testified to be advantageous on incidental vocabulary acquisition in different aspects while they were compared with the reading-only activity. Nevertheless, the comparison of the efficacy of each task was difficult to make due to the lack of the operational criteria. Thus, which task is with the most power is still ambiguous, and in need of study. This unsolved question seemed to find its way out when the Involvement Load Hypothesis first appeared in 2001. Laufer and Hulstijn (2001) proposed the Involvement Load Hypothesis, a motivational-cognitive construct. With the operational criteria, need, search, and evaluation, researchers could manipulate the task with different degrees of the involvement loads and predict to a certain extent different tasks yield on the vocabulary acquisition. Following that, an experiment (Hulstijn & Laufer, 2001) based on the Involvement Load Hypothesis was conducted on EFL university students. The results verified this hypothesis in that the composition writing task outperformed the other two tasks, the reading task and the filling gaps task on the vocabulary acquisition. However, the difference between the gaps filling task and the reading task was partially proved. The Involvement Load Hypothesis shed the light for language teachers on the criteria to follow when choosing a task to conduct in the classroom, but it also needed more experiments to examine and adjust its assumptions.. Purpose of the Study The aim of this thesis is to investigate the efficacy of three contextualized different learning tasks, the reading-comprehension task, the gap fill-in task, and the picture-writing task, on the incidental vocabulary acquisition and retention of junior high school students in Taiwan. This study replicates the experiment design conducted by Hulstijn and Laufer in 2001, but with slight modification on the gaps filling task, the 3.

(15) composition writing task and the posttests. In this study, the Involvement Load Hypothesis is applied to measure the depth of processing when participants carry out the three learning tasks, and to predict the incidental vocabulary acquisition caused by the three different learning tasks. Besides, language proficiency is also explored to see whether the effectiveness of the three learning tasks vary according to the language proficiency.. Definition of Terms. Incidental learning An incidental learning situation is a situation in which an individual processes new information without the intention to commit this information to memory when the “the learners’ primary objective is to do something else.” (Schmidt, 1994).. Incidental vocabulary learning Incidental vocabulary acquisition is defined as the acquisition of vocabulary as a by-product of another activity or task (Laufer and Hulstijn, 2001).. Involvement Load Hypothesis The involvement load hypothesis is a motivational-cognitive construct of involvement, which consists of three components: need, search, and evaluation. The retention of an unfamiliar vocabulary depends on the degrees of prominence of the three components in processing a task (Laufer and Hulstijn, 2001).. 4.

(16) Significance of the Study The previous studies based on the Involvement Load Hypothesis are conducted among EFL university students or adults with advanced or intermediate language proficiency (Hulstijn & Laufer, 2001; Keating, 2008; Martinez-Frenandez, 2008; Jing & Jianbin, 2009; Yaqubi Rayati &Gorgi, 2010; Nassaji & Hu, 2012). This kind of experimental design is seldom applied to EFL learners, like junior high school students in Taiwan. Whether the task effectiveness will work identically with EFL beginning learners needs an exploration. Besides, the M-shaped performance of EFL novices in Taiwan prescribes the need to check whether the task effect is constant across language proficiency level. Therefore, this study adopts language proficiency as a secondary factor for analysis.. Organization of the Study This study is composed of five chapters. Chapter One describes the background and the motivation for this study. Chapter Two presented the Involvement Load Hypothesis, its empirical studies followed by the review of previous studies, and the effect of different learning tasks followed by the research questions. Chapter Three presents the research design. Chapter Four illustrates the results and discusses the findings in this experiment. Chapter Five reveals the implications on pedagogy, the limitations of the study, and the directions for future research, and a conclusion.. 5.

(17) 2CHAPTER. TWO. LITERATURE REVIEW This chapter consists of four sections. In the first section, the notion of incidental learning is discussed followed by a discussion on the incidental vocabulary acquisition in the second section. In this section, the incidental vocabulary learning from task-focused activities are compared with the ones from reading only activity. To evaluate the efficacy of different task-focused activities, the Involvement Load Hypothesis is elaborated in the third section. In the last section, the results of previous experimental studies are discussed and compared in the light of the Involvement Load Hypothesis followed by the research questions.. Incidental Learning An incidental learning situation is a situation in which an individual processes new information without the intention to commit this information to memory when the “the learners’ primary objective is to do something else.” (Schmidt, 1994). To measure the incidental learning caused by a task or an activity, a theoretically-based classification is required. A theoretical assumption on incidental learning was proposed by Craik and Lockhart (Michael & Eysenck, 1982). Craik and Lockhart (1972) argued that “trace persistence” was conditional upon “depth of analysis” in which deeper levels could result in longer lasting and stronger traces. Based on Craik and Lockhart’s original formulation, there were two types of processing, one was that learners were exposed to the learning material without learning instruction (Type I design), and the other was that learners were given instructions to conduct an activity by means of using the learning material (type II design). From then on, hundreds of experiments on the incidental and 6.

(18) intentional learning, mainly in the field of psychology, have been conducted in the operational condition where learners are informed of the coming recall in advance (Laufer & Hulstijn, 2001). In 1994, Schmidt refers the incidental learning to “the learning without an intent to learn, or the learning of one thing when the learners’ primary objective is to do something else.” Under this more general definition, no mention about the forewarning of the upcoming test, the notion of incidental learning is extensively applied in the field of second and foreign language pedagogy, such as the incidental vocabulary acquisition. This study employs Schmidt’s (1994) definition on incidental learning.. Incidental Vocabulary Learning Just like incidental learning where learners acquire new information unintentionally when processing a task or an activity, incidental vocabulary acquisition is defined as the acquisition of vocabulary as a by-product of another activity or task (Laufer and Hulstijn, 2001). Through processing the task, learners acquire the new words without deliberately committing lexical information to memory. Hence, this kind of learning is incidental, not intentional, (Laufer, 2001) is in the light of whether the lexical learning is a by-product of reading (Krashen, 1989), another learning task, or an intention to commit to memory.. Incidental Vocabulary Acquisition through Reading Incidental vocabulary learning may occur during the process of another learning activity, as in reading a text. The reading process supports the incidental vocabulary acquisition over direction instruction in three pedagogical advantages, contextualization, effectiveness, and individualization (Huckin & Coady, 1999). The first feature is 7.

(19) contextualization. The lexical meanings of new words are contextualized in an authentic material in which learners acquire new words on its full scales, not just in one dimension. They learn not only the semantic meaning of the new word but also its pragmatic understanding and syntactical structure. Thus, the learner brings out the comprehensive understanding of how new words collocate with other words, not just committing the form or the meaning to memory. Second, there is effectiveness. Language learners experience at least two kinds of learning: using strategies to perform the assigned task, such as reading, and at the same time, absorbing the lexical meaning of new words during the process of the task accomplishment. This kind of effectiveness is welcomed by language teachers because of the limited teaching time and is beneficial for language learners for the one shot with two gains. Thirdly, the incidental vocabulary learning is learner-based. Learners obtain suitable reading texts chosen by teachers according to their personal interest, language proficiency, or the learning goal. With the properly designed or chosen reading text, learners can acquire the indispensable words in a more effective way. Although learners can benefit from incidental vocabulary learning, there are three challenges left to be dealt with (Huckin & Coady, 1999), the comprehensible input, the learners’ attention, and the times of exposure to a word. First, conditional upon Input Hypothesis (Krashen, 1989), the reading text should be comprehensible input for learners to learn words incidentally. In other words, if learners are not capable of getting or guessing the meanings of new words due to the language proficiency, the lexical difficulty, or the barren context, the function of incidental vocabulary learning will decline or vanish. Another situation is learners do the guessing and derive the lexical meaning of new words, but in the wrong direction, the incidental vocabulary acquisition does occur, but is not helpful in speeding up the correct vocabulary acquisition. Secondly, when reading a text, learners may not attend to the meaning and the form of a 8.

(20) new word under the rich text content. Since the completion of a reading activity is the priority, learners will naturally neglect the new words in the text once they grasp the messages the text is trying to convey. Therefore, many scholars insist that incidental vocabulary acquisition happens only when learners’ attention to both meaning and form (Robinson, 1995). Without the interaction between the meaning and the form, it is laborious to commit a word to memory. Thirdly, learners need repeated exposure and retrieval to a word in order to build up the knowledge of a word. The whole scale of a word includes the language meaning, the language form, and the language in context; therefore, it is almost impossible to acquire a word during one exposure. Nation (1990) reviewed the previous studies and concluded that 5 to 16 exposures to a word could fulfill the vocabulary learning. Such a wide range of exposure times is dependent on the context where a word is presented. These three mentioned restrictions will hamper incidental vocabulary acquisition through the reading activity. L2 vocabulary learning had been neglected for many years, but recently it has gained the attention from researchers on different issues, including the size, the exposure frequency, the strategy, and the mechanism of the L2 vocabulary learning (Huckin & Coady, 1999). The solution of these issues will help learners to cope with the obstacle when learning a second language. The first issue is vocabulary size. All the second language learners and their teachers are well aware of the fact that learning a second language involves the learning of large numbers of words (Laufer & Hulstijn, 2001). A very influential view of vocabulary acquisition is that learners acquire most words through exposure to language input (Laufer, 2001). The language input, particularly reading input, is likely to be the main source of native language learners’ vocabulary. However, the situation is totally different to L2 learners. If most words were indeed acquired from extensive reading, L2 or FL learners would have to read approximately as much as L1 children do – that is, a million words of text a year 9.

(21) (Anderson et al, 1988). Immersing learners in this kind of learning environment is hardly possible for language teachers to create in the classroom or at school. If teachers can only conduct a small quantity of reading in class, the question turns to which effective method should be adopted in helping learners absorb sufficient vocabulary during the limited period of class time.. Incidental Vocabulary Acquisition via Performing Tasks However, that new words can be partially learned from reading does not mean that reading is the most effective method of vocabulary learning. Extensive reading for meaning does not lead to the vocabulary acquisition automatically because readers’ attention is focused on meaning or communication (Huckin & Coady, 1999). Attention means purpose, which in turn “is governed in large part by task demands” (Huckin & Coady, 1999). In this way, performing a task can be a better means of increasing vocabulary than extensive reading because by performing tasks, learners’ attention can be directed to the form or both the form and the meaning. The more attention a new word gets from learners, the more chances the word can be acquired. The general definition of task as “an activity or action which is carried out as the result of processing or understanding language”. Positive evidences from previous studies support that” text-based tasks can enhance incidental vocabulary acquisition” (Huckin & Coady, 1999) than reading a text does. Min (2008) and Paribakht & Wesche (1997) proved the reading plus activity was more beneficial to ESL learners than the reading only activity. Peters, Hulstijn, Sercu, and Lutjeharms (2009) also testified that a vocabulary task after a text reading activity was effective in receptive word learning. Laufer (2003) and Browne (2003) discovered the benefits of sentence writing task on the vocabulary retention. 10.

(22) One may argue that the results discussed above are unfair to the reading-only group since different stimuli naturally benefit language learning much more than a signal way. Therefore, Laufer conducted three experiments in 2003 to compare the efficacy between the reading activity and the word-focused task. The results revealed that the significantly worse vocabulary retention on the text reading group, compared with the word-focused group, even when the text was with marginal glosses or looking up the words in a dictionary was required. From the review of these empirical studies, we can conclude that tasks plus reading result in different degrees of incidental vocabulary acquisition. In order to learn a L2 word without attempting to commit it to memory, word-focused tasks may be a good method (Laufer, 2001). Since task is goal-oriented, which variable existing in different tasks causes distinct vocabulary learning is the question needed to be explored.. Involvement Load Hypothesis Psychologists have proposed that the better performance of learning is contingent upon “depth of processing”, or “degree of elaboration”, or “quality of attention” to information. Craik and Lockhart (1972) proposed the depth of processing hypothesis, which declares that “mental activities which required more elaborate thought, manipulation, or processing of a new word will help in the learning of that word.” Processing the lexical information of a new word more elaborately will result in the better retention than if it had been processed less elaborately (Laufer, 2001). It is the depth of processing, not with the length of exposing time that increases the chance of transforming a new message from short-term memory to long-term memory.. That is to. say, when learners “pay more careful attention to the word’s pronunciation, orthography, grammatical category, and semantic meaning”, they are increasing the chance to retain 11.

(23) the word in memory (Laufer & Hulstijn, 2001). Researchers also agree that “greater depth implies a greater degree of semantic or cognitive analysis” (Craik & Lockhart, 1972, p.675). If learners connect one property of a word with its another or existing property, the retention may last longer. The initial shallow or deep processing of new information will determine if it will be stored into long-term memory. The processing emphasizes the elaboration of the learning, which is irrelevant to the incidental or intentional vocabulary learning (Hulstijn & Laufer, 2001). A direct lexical instruction can work as well as a task-based vocabulary learning as long as learners pay enough attention, at the same deep processing. However, the notion of “depth of processing”, or “degree of elaboration”, or “quality of attention” lacks an operational definition (Baddeley, 1978). Acknowledging lack of operational definition, Laufer and Hulstijn (2001) attempted to identify the components of incidental tasks, which might provide researchers and teachers the criteria to observe, manipulate and measure the effectiveness of learning tasks. In 2001, they proposed the Involvement Load Hypothesis. It is a “motivational-cognitive construct of involvement”, which explains and predicts the degree of task effectiveness with regard to the retention of new words. The Involvement Load Hypothesis is based on the following three assumptions. The first assumption explains that “Involvement” consists of three basic components: need, search and evaluation. The need component is “the motivational, non-cognitive dimension.” It is “the attempt to achieve a goal” imposed either internally or externally (Laufer & Hulstijn, 2001, p.14). The search and the evaluation components are the two “cognitive (information processing) dimensions of involvement.” The search component is “the attempt to find the meaning of an unknown L2 word or the L2 word form expressing a concept by consulting a dictionary or another authority” (Laufer & Hulstijn, 2001, p.14). The evaluation component is a serial of comparison, “a 12.

(24) comparison of a given word with other words, a comparison of one specific meaning of a word with its other meanings, or a comparison of the word with other words in order to assess whether the word does or does not fit its context.” (Laufer & Hulstijn, 2001, p.14) The three components, need, search, and evaluation, can be absent or present with different degrees of prominence. Need can be present with moderate or strong degrees. When a task is required by an authority, the degree of the component need is moderate. Need is strong when a task is triggered by the internal will of the learners, not the exterior command. Need cannot be absent in nature since a task is the attention to complete a mission. Search can be absent or present with a moderate degree. When the unfamiliar words are glossed in the margin, search does not exist since learners do not have to search for their meanings. When learners desire to “know the meaning of a L2 word or to find a L2 word form to express the L1 meaning concept” (Laufer & Hulstijn, 2001), the degree of the component search is moderate. Evaluation can be absent, or present with moderate or strong degrees. Evaluation does not occur when the comparison and the combination of a new word does not happen; for example, the word is glossed in the margin. When learners compare a new word with other words, or the different meanings within a word, the degree of the component evaluation is moderate. Evaluation is strong when a target word is combined with other words in a sentence or in a context. “Involvement load” is defined as “the combination of these three components with their degrees of prominence” (Hulstijn & Laufer, 2001, p.15). The second assumption of Involvement Load Hypothesis is that “words which are processed with higher involvement load will be retained better than words which are processed with lower involvement load” (Laufer & Hulstijn, 2001, p.15) when other factors are equal. Hence, the retention of unfamiliar words is in conditional on the degree of the involvement load a task seduces “regardless of whether it is an input or an 13.

(25) output task” (Laufer & Hulstijn, 2001). Tasks with different involvement loads could be compared and predicted on their effect towards incidental vocabulary learning (Laufer & Hulstijn, 2001). When word complexity factors hold constant, the Involvement Load Hypothesis predicts that tasks with higher involvement load will yield better vocabulary learning results than tasks with lower involvement load. The third assumption of the Involvement Load Hypothesis is “tasks with a higher involvement load will be more effective for vocabulary retention than tasks with a lower involvement load.” Based on this assumption, tasks can be compared or manipulated by teachers or researchers according to the combination of the three components to reflect levels of involvement load. In the classroom, tasks with different involvement loads can be applied according to the learning requirements. The tasks with higher involvement load can be applied to the fundamental vocabulary which is important but unfamiliar to the learners, and the tasks with lower load can be applied when the learning focuses on the concept conveyed by the context, not the form or the meaning carried by the word. In the experimental studies, researchers could compare the effectives of different learning tasks with these manipulative criterions. This hypothesis can be verified under the same standard. In the task-based approach, “task” is defined as “an activity in which meaning is primary; there is some sort of relationship to comparable real-world activities; task completion has some priority; the assessment of the task is in terms of outcome” (Skehan, 1998, p.95). However, Laufer and Hulstijn adopted the general definition of task as “an activity or action which is carried out as the result of processing or understanding language” (Laufer & Hulstijn, 2001, p.16). They further used the three components, need, search, and evaluation, as descriptors to illustrate seven widely practiced classroom activities (see Table 1). A minus (－) indicates an absent component, which is marked as 0 in the involvement index. A plus (＋) indicates a present 14.

(26) component with a moderate degree, which is marked as 1 in the index. A double plus (＋＋) indicates a present component with a strong degree, which is marked as 2. As shown in Table 1, the first three types of task are reading with comprehension questions with (1) glossed in text but irrelevant to the task, (2) glossed in text and relevant to the task, and (3) not glossed but relevant to task; the fourth type (4) is reading and comprehension questions and filling gaps with glosses relevant to reading comprehension. The last three types of task are about writing (5) original sentences with glosses listed, (6) a composition with concept (in L1) selected by the teacher (7) a composition with concept selected by the learner. Among the seven, types relevant to the tasks employed in this study are type (2) reading and comprehension questions, (4) reading and comprehension questions and filling gaps, and (5) writing original sentences.. Table 1 Task-induced Involvement Load (Laufer & Hulstijn, 2001, p.18) Task 1. Reading and comprehension questions 2. Reading and comprehension questions. Evaluation. Involvement load. Status of target words. Need Search. Glossed in text but irrelevant to the task. －. －. －. 0. Glossed in text and relevant to the task. ＋. －. －. 1. －/＋. 3. Reading and comprehension questions. Not glossed but relevant to task. ＋. ＋. (depending on word and context). 4. Reading and comprehension questions and filling gaps. Relevant to reading comprehension. Listed with glosses at the end of the text. ＋. －. ＋. 2/3. 2. (continued on next page) 15.

(27) Table 2 Task-induced Involvement Load (Laufer & Hulstijn, 2001, p.18) (Continued) 5. Writing original sentences. Listed with glosses. ＋. －. ＋＋. 3. 6. Writing a composition. Concepts selected by the teacher (and provided in L1). The L2 learner-writer must look up the L2 form. ＋. ＋. ＋＋. 4. 7. Writing a composition. Concepts selected (and looked up) by L2 learner-writer. ＋＋. ＋. ＋＋. 5. As shown in Table 1, different learning tasks generates different involvement loads in conditional on the combination of the three components. Laufer and Hulstijn (2001) gave researchers three pieces of advice on the comparison of task effect; first, tasks with the same number of components, but with different degrees of prominence could be compared to measure whether the three components carry the equal weight; second, a comparison of the same involvement load could be done between an input task and an output task to confirm the contention that involvement load induced by a task is the determinative factor on the word retention; thirdly, a task with good quality, a higher involvement load, could possibly result in equal learning as the task with quantity, a lower load with several retrieval, hence a comparison of quality input, i.e., with high involvement load, with quantity input, namely, with repeated exposure.. 16.

(28) Previous Vocabulary Studies Applying the Involvement Load Hypothesis Immediately after the Involvement Load Hypothesis (Laufer and Hulstijn, 2001) emerged from the depth of processing, many researchers in SLA conducted experiments based on the assumptions from this hypothesis. They testified this motivational-cognitive construct and explained their study results from this contention. However, diverse experimental results emerged although they were based on the same proposal, the Involvement Load Hypothesis. Several studies (Hulstijn & Laufer, 2001; Laufer, 2003; Tu, 2004; Keating,2008; Kim, 2008; Martinez-Frenandez, 2008;Jing & Jianbin ,2009; Yaqubi et al, 2010; Marmol & Sanchex-Lafuente,2013) compared tasks with different degrees of involvement loads to echo Laufer and Hulstijn’s request for further research on this motivational-cognitive construct. Tu (2004) and Kim (2008) replicated the three tasks in Hultsiin and Laufer (2001) in two paralleled experiments and found positive effect of task on vocabulary acquisition. The composition writing task (the high involvement load) outperformed the text fill-in task (the medium involvement load), and in turn the text reading task (the low involvement load). Jing and Jianbin (2009) also compared three listening tasks and discovered listening with a composition writing excelled listening with relevant glosses, and listening with relevant glosses excelled listening with irrelevant glosses. The results confirmed the second assumption of the Involvement Load Hypothesis that “words which are processed with higher involvement load will be retained better than words which are processed with lower involvement load” (Laufer & Hulstijn, 2001). Keating (2008) and Marmol & Sanchex-Lafuente (2013) also compared three learning tasks, the sentences writing task, the text fill-in task, and the text reading task. 17.

(29) They found evidence partially supporting the Involvement Load Hypothesis. In Keating’s study, positive evidence only appeared in the active recall (L1 → L2 translation). The sentence writing task yielded a better effect than the text fill-in task, and the text fill-in task better than the text reading task. But in the passive recall (L2 →L1 translation), the significant difference only existed in the sentence writing task over the other two. Marmol and Sanchex-Lafuente (2013) successfully proved the superiority of sentence writing task over the text fill-in task, and the text fill-in task over the text reading task. However, the fourth task with the highest involvement load, sentences writing with looking up words, did not show its superiority over the previous three learning tasks. In Martinez-Frenandez’ study (2008), the sentence/composition writing task was replaced by a text reading with multiple-choice glosses. Although the text fill-in group (involvement load=2) outperformed than the text reading group (involvement load=1), the text reading with multiple-choice glosses group (involvement load=3) showed a similar performance to the text reading group (involvement load = 1) and the control group (involvement load = 0). In Yaqubi et al.’s study (2010), the fill-in text task (involvement load=2) facilitated more vocabulary learning than the text reading task with looking-up the words (involvement load=3). In order to testify the effectiveness of a task on incidental word learning is conditional upon the involvement load, not on the form of the task, Kim (2008), Yaqubi et al. (2010), Laufer (2003), and Zou (2017) compared tasks with equal involvement load but in different task types. In Kim’s second experiment (2008), positive evidence was found from the comparison of the sentence writing task and the composition writing task. These two equal-weight learning tasks resulted in similar promotion on vocabulary acquisition anticipated by the Involvement Load Hypothesis. However, in Zou’s study (2017), negative evidence was found that sentences writing facilitated 18.

(30) worse word learning than composition writing did, which required more metacognitive strategies to create a coherent context. Yaqubi et al. (2010) compared the reading task with looking-up words (input task) with the reading task with a summary writing (output task). The output task had a significantly better effect than the input task, which violated the assumption that equal load of involvement elicits equal amount of learning. The counterevidence also appeared in the third experiment of Laufer’s study in 2003. Three learning tasks with equal load revealed significant difference in promoting the vocabulary acquisition. Both the two output tasks (the sentence fill-in tasks with looking-up words and the sentence writing task) achieved a better effect than the input task (the text reading task with looking-up words). The significant different effect also appeared in the two output tasks in the delayed posttest. To testify if a task with quality could compensate for the frequency of occurrence on incidental vocabulary learning, Folse (2006) and Lu (2013) compared a triple filling-in task with a writing task. Flose (2006) proved three fill-in-the-blank exercises yielded better effect than one sentence writing exercise when the time on task was controlled. Lu (2013) discovered the triple blank-filling group with less time consuming outperformed the summary writing group with more time consuming. The superiority of frequency of occurrence also appeared in the study conducted by Hulstijn, Hollander, Greidanus (1996), which compared the effect from the glosses, the dictionary, and the inference during the reading tasks under the 25-minute time constraint. In this study, the target word with three-time appearances retained better than the one with one show-up. These findings violated the assumption that under the time constraint, a task with higher involvement load could result in equal learning as the several-retrieval task with a lower load. These various findings on the Involvement Load Hypothesis promise the significant promotion of learning tasks on the incidental vocabulary acquisition, but the 19.

(31) controversial results bring out the necessity to explore the assumptions from the Involvement Load Hypothesis, especially on the degree of prominence of the three components. Besides, with positive or negative results, most of the previous studies sampled the intermediate or advanced adult FL or L2 learners, fewer experiments were conducted on the beginning learners, especially on junior high school students in Taiwan. Moreover, almost the above studies accepted Laufer’s perspective in which time is the essence of a task, needed not to be controlled, so the time consumed on task varied decreasingly from the task with higher load to the task with lower load. This leads to a possible explanation that one task outperforming than another in the vocabulary learning could be resulted from the immerging time, not the involvement load. Therefore, in this study, three learning tasks with different degrees of prominence on the component, evaluation, were adopted to reexamine the effectiveness of tasks on incidental vocabulary acquisition under the time constraint. The reading-comprehension task, the gap fill-in task, and the picture-writing task were chosen because they are commonly practiced as classroom activities in Taiwan and the commonly researched tasks in the previous studies, so here are my research questions: (1) Are there any differences among the reading-comprehension task group, the gap fill-in task group, and the picture-writing task group on their immediate vocabulary posttest scores? (2) Is there an interaction effect between tasks and reading proficiency on immediate vocabulary posttest scores? (3) Are there any differences among the reading-comprehension task group, the gap fill-in task group, and the picture-writing task group on their delayed vocabulary posttest scores? (4) Is there an interaction effect between task and reading proficiency on delayed vocabulary posttest scores? 20.

(32) 3CHAPTER. THREE. METHODOLOGY. Participants Seventy-two EFL learners from three intact ninth-grade junior high school classes in Taichung City with 21, 21, and 30 in each class participated in this study. Participants have learned English at least for six years starting from the second graders in elementary school; however, they come with different language proficiency due to different learning experience or learning environment. The participants were informed of the study in advance and signed a consent form (see Appendix I) prior to the study.. Materials. Selection of Reading Texts and Target Words Three reading texts were selected from the supplementary reading books, Topics and Tales plus+ Book 1, edited by Trafford (2004, p. 76-79), and Topics and Tales plus+ Book 2, edited by Trafford (2005, p. 24-27 & p. 52-55), for junior high school students in Taiwan. The three reading texts were all narrative, but with different topics. The titles of the three reading texts were ”Do Bees Ride Bicycles?” “Hold on, Kitty!”, and “Just Don’t Tell Anyone“! The original length of the texts were 153-word, 124-word, and 195-word long respectively. All the three reading texts were adapted by the researcher to reach a equal length, 150 words, with the aim to set up the comparable text length. They depicted stories about an accident happening in the daily life: a bike covered by bees, a 21.

(33) cat left on the top of the car, and a king washed by the flow. They were topics familiar to junior high school students. The readability of the three texts was 4.9, 3.7, and 1.7 respectively, based on Flesch Kincaid Grade Level. Given 150 words in length, it was decided that 6 target learning words were appropriate for general reading comprehension – a 96% text coverage of known words, somewhere between Nation’s’ (1992) threshold of 98% coverage for pleasure reading and Hirsh and Nation’s (1992) proposal of 95% coverage, the least for general reading comprehension. Furthermore, it was also decided that only content words that name object of reality and their qualities would be chosen because they were easier for English as foreign language learners to capture and maintain meaning than function words. With these criteria established, the author then proceeded to select 6 target words for each text by following these steps. First, based on her 15 years’ teaching of English at junior high school, 6 target words were subjectively selected for each of the three reading texts. Then, the vocabulary from the 1200-word list for junior high school students in Taiwan, published by the Ministry of Education in 2010, was referenced to ensure that the selected target words were not on the MOE vocabulary list for junior high English. Finally, words other than target words were examined. In the case where words did not fall in the MOE list, a more frequent word that is on the list was in place. This resulted in six target words for each text, including two nouns, tow verbs, and two adjectives. In each reading text, the 6 words were boldfaced with their L1 translation and part of speech annotated on the margin (see Appendix C). Table 3 presents the target words selected from the three reading texts and their part of speech.. 22.

(34) Table 3 The List of Target Words Target word Reading text Text 1. Text 2. Text 3. Noun. Verb. Adjective. expert. mantle. faithful. hive. snare. delighted. pharmacy. perceive. horrified. claw. blink. affrighted. buffalo. flow. ferocious. field. drag. marvelous. Subtotal of the target words 6. 6. 6. Three Learning Tasks Three learning tasks that were commonly practiced in the classroom activities were adopted, the reading-comprehension task, the gap fill-in task, and the picture-writing task. The three learning tasks were described in the following sections.. The Reading-Comprehension Task (hereafter RC Task) The reading-comprehension task was a reading text followed by three reading comprehension questions displayed on one paper sheet (see Appendix C). The six target words were in bold face in the text and a L1 marginal glossary with their part of speech was provided on the right-hand side. Along with the task sheet, a piece of blank paper was provided for students to write a brief summary of five sentences in Chinese after the reading comprehension task sheet was collected.. The Gap Fill-in Task (hereafter GF Task) The gap fill-in task sheet contained a reading text in which the six target words were replaced by a blank (see Appendix D). Six target words and one distractor along 23.

(35) with their L1 translation and their part of speech were listed on the right-hand side of the paper sheet in the order of the alphabet. The participants were instructed to fill in the blank with the most suitable word to make the reading text sensible. Along with the task sheet, a piece of blank paper was provided for student to write a brief summary of five sentences in Chinese after the gap fill-in task sheet was collected. The Chinese summary was included in RC Task and GF Task for the sake of length of time alignment with the picture-writing task. Since it is in Chinese, the repeated processing of the English target words is minimal.. The Picture-Writing Task (hereafter PW Task) The picture-writing task required the participants to write a short story based on serial pictures in four frames and the six target words were provided in the frame (See Appendix E). The pictures were constructed by the researcher based on the sequence of the same story used for RC Task and GF Task. They were presented in sequence according to the corresponding reading text. The target words were listed separately with L1 translation, part of speech, and they also appeared in the pictures to indicate when to use the words and to facilitate the story writing. The purpose of pictures with target words provided was to elicit writing content and meanwhile guiding students in employing target words in constrained content. The participants were instructed to write as many sentences as they could with the target words within the 10-min time period even though they might not describe the whole story. These three tasks, although came with different formats and different degrees of involvement load, they shared a common feature, text processing at discourse level, hence, a context for comprehension and production of target words. Based on the Involvement Load Hypothesis, the involvement loads of the three 24.

(36) learning tasks varied in term of evaluation (see Table 4). All the three tasks were imposed by the teacher, which created moderate “need.” The students were not required to search since glosses of the target words were provided; therefore, no point was given to “search”. The involvement load of RC Task is 1, with moderate need, no search, but no evaluation. With the marginal glosses, the assessment of an outcome was not needed. The involvement load of GF Task is 2, with moderated need, no search and moderate evaluation. The participants were demanded to fill in the blank with the target word after selection, which required assessment of several words provided. The involvement load of the PW Task is 3, with moderate need, no search and strong evaluation. The participants were asked to write a short story with the target words with L1 translation and part of speech provided. Creating original sentences to make up a story required strong evaluation. According to the Involvement Load Hypothesis, the incidental vocabulary acquisition from the task with higher load would yield more retention of words. In other words, PW Task would yield better performance than both RC Task and GF Task. Meanwhile, GF Task would have more advantage over RC Task.. Table 4 The Involvement Loads of Tasks Components of involvement loads. Involvement. Tasks Need. Search. Evaluation. load. RC Task. ＋. －. －. 1. GF Task. ＋. －. ＋. 2. PW Task. ＋. －. ＋＋. 3. 25.

(37) Instruments The instruments applied in this study were the reading proficiency test, the pretest, and the posttests. The reading proficiency test was administered in advance in order to determine higher-lower proficiency levels. The pretest was done as a baseline prior to the experiment to ensure unfamiliarity of the target words for all the students. Two posttests were conducted to obtain the developing vocabulary learning of the participants; one immediate posttest was following each task, and the other delayed posttest was done five days after each task session.. The Reading Proficiency Test The purpose of the reading performance test was to classify the reading ability of the students into the higher-proficiency level and the lower-proficiency level and to ensure comparability of proficiency in treatment assignment. The test (see Appendix A) on the reading for the Elementary Level of GEPT was sampled from the GEPT Website (https://www.gept.org.tw/Exam_Intro/down01.asp) and was administered in the three classes before the experiment was conducted. There are 35 multiple choice questions, with 3 sections including vocabulary and grammar tests, cloze tests, and reading comprehension tests. A sample was taken from each section and shown in Appendix A.. The Pretest A pretest was administered ahead of the experiment to survey if the students had ever seen or acquired any of the target words before this experiment. The target words were listed on the test sheet (see Appendix B), and the students were required to check if they had seen it before. If so, they were also asked to write down its Chinese translation in the next column. The results of the pretest indicated that 5 students had 126.

(38) to 3- word knowledge of the 18 target words, who were then eliminated from the study.. The Immediate and Delayed Posttests The Vocabulary Knowledge Scale (VKS) was adopted in this study as the immediate and delayed posttests. The VKS was a self-report scale, from I to V, which was developed by Paribakht and Wesche (1997) to have the test takers self-estimate their own knowledge about one particular word so as to manifest the test takers’ receptive and productive vocabulary knowledge. This 5-category scale was to distinguish the developing knowledge of particular words in the language learners, ranging from totally unfamiliarity, through 3 degrees of recognition, to grammatically and semantically accurate usage of a word (see Figure 1).. I. I don’t remember having seen this word before.. II. I have seen this word before, but I don’t know what it means.. III. I have seen this word before, and I think it means_____. IV. I know this word. It means _____. (synonym or translation). V. I can use this word in a sentence: _____. (write a sentence.) (If you do this section, please also do Section IV.). Figure 1 Self-Report Categories in the VKS Elicitation Scale (Read, 2000, p.133). Since the students in this study were not English native speakers, the VKS was translated into a Chinese version by the researcher to eliminate the misunderstanding of the 5-scale category illustration. The revised VKS were used to test six target words at once in 3 rounds of immediate posttest (see Appendices F), and 3 rounds of delayed posttest, which were rearranged in accordance with alphabetic order (see Appendix G).. 27.

(39) Pilot Study A series of preliminary studies were conducted to identify potential pitfalls in the proposed methodology. Two ninth graders from the same junior high school participated in the pilot study. The sample test on the reading for the Elementary Level of GEPT was implemented to measure their reading proficiency. The results reflected that 40 minutes was enough to complete the 35 test items and the test was appropriate in difficulty level (one scored 28 and the other 29). Following that, the pretest was conducted to make sure the 18 selected target words were all beyond the participants’ knowledge. Based on the results, the target word, clinic, was replaced by another word, pharmacy because one participant provided the correct Chinese translation. Three days later, the three different learning tasks designed from text 1 were practiced by the two participants to measure how much time was properly needed for the participants. They completed the reading comprehension task in 3 minutes, and the gap fill-in task in 4 minutes; therefore, a Chinese summary writing was followed to these two tasks to equalize the time allocated, 10 minutes, required by the picture-writing task. Immediately after the task, each student was requested to fill out the revised VKS (Chinese version) to confirm the Chinese description was clear enough and the 5-min time period was sufficient for them. Two days later, the revised VKS was administered again as the delayed posttest. The materials and instruments applied in the pilot study were collected by the researcher right after each particular activity was completed. Both the students were not included in this present study.. 28.

(40) Design and Data Collection. Initial Group Assignment Before the experiment, each participant took a mocked GEPT test within 40 minutes to examine their English reading ability. For the test, one correct response to the test item was scored 1, so the max scores were 35. The mean score of the total 72 participants was 17.35, SD = 6.30, which was used as a cutting score, based on which higher-proficiency level, N=36, and lower-proficiency level, N=36, would be formed. Based on the test, the participants were numbered from No. 1 to the highest scores until No. 72 to the lowest scores so as to assign them to three tasks. In the first round, No. 1 to No. 3 were distributed to Subgroup A forwards from RC Task, GF Task, to PW Task (see Table 3). In the second round, No. 4 to 6 were distributed to Subgroup B backwards from PW Task, GF Task to RC Task. In the third round, No. 7 to 9 were distributed to Subgroup C forwards from RC Task, GF Task, to PW Task. Then the deliberate arrange was repeated again starting from Subgroup C, Subgroup B, to Subgroup A. The normal class grouping was carried out until all the three treatment groups had 24 participants in it. Within each treatment group, eight participants were assigned to a subgroup A, B, or C, according to their reading proficiency test and there was an equal number of higher-lower student distribution with 36 participants in each proficiency level. The initial class grouping was to make sure each subgroup had equal number of higher- and lower- reading proficiency participants across the three task groups (see Table 5).. 29.

(41) Table 5 The Distribution of Participants Task Subgroup Text order. No. of participants. RC Task GF Task PW Task (N=24) (N=24) (N=24) A B C A B C A B C (N=8) (N=8) (N=8) (N=8) (N=8) (N=8) (N=8) (N=8) (N=8) 123 231 312 123 231 312 123 231 312 1 6 7 2 5 8 3 4 9 18 19 36. 13 24 31. 12 25 30. 17 20 35. 14 23 32. 11 26 29. 16 21 34. 15 22 33. 10 27 28. . . . 72. . . . 67. . . . 66. . . . 71. . . . 68. . . . 65. . . . 70. . . . 69. . . . 64. Ultimate Group Assignment However, this ideal member assignment had to be revised after the prior test on the 18 target words. The target words were originally designed to be learned through task work only; thus, it was needed to make sure all the words were unknown to the participants. It turned out that 5 subjects had to be removed because they demonstrated knowledge of one word (4 participants) or three words (1 participant). All of these 5 participants were from higher-proficiency level, 2 from GF Group (Subgroup B) and 3 from PW Group ( 2 in Subgroup B and 1 in Subgroup A); hence, there were 24 participants assigned to RC Group, 22 to GF Group, and 21 to PW Group. Across the three task groups, 23 participants in Subgroup A, 20 B, and 24 C. Moreover, there were 31 participants belonging to higher-proficiency level, and 36 to lower-proficiency level. The resultant number of participants for treatment group, subgroup, and proficiency level is shown in Table 6.. 30.

(42) Table 6 Number of Participants in Each Treatment Group, Subgroup, and Proficiency Task Group. RC. GF. PW. No.. 24. 22. 21. Subgroup. A. B. C. A. B. C. A. B. C. No.. 8. 8. 8. 8. 6. 8. 7. 6. 8. Proficiency. H. L. H. L. H. L. No.. 12. 12. 10. 12. 9. 12. To ensure the comparability in proficiency across the three task groups, and the heterogeneity between the proficiency levels, I performed a three-way ANOVA, with Task Group, Subgroup, and Proficiency Level as three independent variables and GEPT score as a dependent variable, to check if there was difference in proficiency among the three treatment groups and among the three subgroups, and to ensure that the two proficiency levels were distinctive between each other. Levene’s test was insignificant, F (17,49) = .34, p > .05, indicating the homogeneity of variance across groups and thus the assumption for ANOVA analyses was met. As Table 7 shown, the results attested the appropriateness of the group assignment. Group, F (2, 49) = .85, p > .05, η2 = .003, and Subgroup, F(2,49) = .50, p > .05, η2 = .02 did not have an effect on proficiency score. Moreover, Proficiency Level made a difference on proficiency score, F (1,49) = 166.87, p = <.001, η2 = .77, revealing the differentiation between level groups, even after the reduction of five students from the high-proficiency level. Finally, no interaction was found in the rest four sets of comparisons: Group by Subgroup, F (4, 49) = .98, p > .05, η2 = .01; Subgroup by Proficiency Level, F (2, 49) = .39, p > .05, η2 = .02; Group by Proficiency Level F (2, 49) = .11, p > .05, η2 = .01, and Group by Subgroup by Proficiency Level F (4, 49) = .15, p > .05, η2 = .01. Thus, the assumption 31.

(43) of equivalence among cells for two variables, Group and Subgroup, and the heterogeneity between two levels were confirmed.. Table 7 Three-Way ANOVA on Test of Homogeneity of Proficiency Level among Groups and Subgroups. Source. Partial Eta. F. Sig. Group. .85. .92. .003. Subgroup. .50. .61. .02. Proficiency Level. 166.87. .00. .77. Group * Subgroup. .98. .98. .01. Group * Proficiency Level. .11. .90. .01. Subgroup * Proficiency Level. .39. .68. .02. Group * Subgroup * Proficiency Level. .15. .96. .01. Squared. Design of Text and Order This study was conducted in eight English sessions within two weeks. Except the first session of proficiency test, which took up almost a whole class period of 45 minutes, the rest of the 7 sessions took up only 5 to 15 minutes of the class period. Three learning tasks, the reading-comprehension task, the gap fill-in task, and the picture-writing task, were assigned to three treatment groups to practice the designated target task in three sessions. Hence, for each session, one text was developed for the presentation of 6 target words, followed by regular learning activities, and the immediate posttest of the 6 words was administered at the last 5 minutes of the class period. Altogether 3 texts, each presenting 1 set of target words, will be processed by the 3 treatment groups. To eliminate the influence from the presentation order (3) and text (3) on the acquisition and retention of the target words, a 3 x 3 Latin Square design 32.

(44) was employed, such that each treatment group was further divided into three subgroups, A, B, and C. The sequence of presentation of text-target words set were thus counterbalanced via the three subgroups within each treatment group. Subgroup A carried out the task in the order of text 1, text 2 to text 3, Subgroup B in the order of text 2, text 3 to text 1, Subgroup C in the order of text 3, text 1, to text 2 (see Table 8).. Table 8 The Text and Order under Each Task Group Task Group Subgroup Session 3 Session 4 Session 5. RC Group. GF Group. PW Group. A. B. C. A. B. C. A. B. C. Text 1 Text 2 Text 3. Text 2 Text 3 Text 1. Text 3 Text 1 Text 2. Text 1 Text 2 Text 3. Text 2 Text 3 Text 1. Text 3 Text 1 Text 2. Text 1 Text 2 Text 3. Text 2 Text 3 Text 1. Text 3 Text 1 Text 2. Data Collection Procedure During the first week, in the first session, all the 72 students took the reading proficiency test. Based on the test scores, the students were assigned to three treatment groups. In the second session on the second day, the students took a pretest on the knowledge of 18 target words. Based on the pretest, 5 students were removed from the study. The remaining 67 students went through the following three sessions (session 3 to 5) of task treatment and immediate posttest. Five days after each treatment-immediate posttest session, a delayed posttest was done in session 6 to 8 in the second week (see Table 9). In each of the 3 treatment-immediate posttest session, the participants were 33.