聽力聯合模式：說明信念與策略對聽力理解能力影響之實證模式

全文

(1)國立臺灣師範大學英語學系博士論文 Doctoral Dissertation Department of English National Taiwan Normal University. 聽力聯合模式：說明信念與策略對聽力理解能力影響之實證模式. The Conjugated Model of Listening: An Empirical Model to Explain the Impact of Beliefs and Strategies on Listening Comprehension. 指導教授 : 曾文鐽博士 Advisor: Dr. Tseng Wen-Ta 研究生: 倪強 Student: John-Michael Lopez Nix. 中華民國 104 年 5 月 May, 2015.

(2) 聽力聯合模式：聽力聯合模式：說明信念與策略對聽力理解能力影響之實證模式摘要現有研究的基本假設是學習者信念會影響語言學習策略運用的範圍，因此亦間接影響了第二語言 (second language, L2) 學習的結果。赫洛維茲（Horowitz）是這個理論傑出的支持者，且發展了語言學習量表的學說（BALLI: Horowitz, 1985; 1988）。過去數十年變項分析研究已經證實語言學習信念和學習行為之間的聯結，實務上，這些發現也經常被應用於第二語言的學習策略。到目前為止，尚無研究曾經仔細檢視 L2 聽力學習信念和聽力策略以及聽力熟練度之間的關係。假定語言學習信念會影響 L2 學生的學習行為，那麼評量 L2 聽力信念和審視它們和其他獨立變項間的交互影響是非常必要的。本研究採用赫洛維茲的概念模式，將之應用在 EFL 聽力的特定領域。透過結構方程模式（SEM）、試題反應理論（IRT）以及多向度試題反應理論（MIRT）等方法，本研究提出一個實徵上可驗證的 EFL 聽力學習的模式以瞭解影響 EFL 聽力的潛在特質變項與觀察變項間的交互作用。研究結果支持赫洛維茲的語言學習信念概念直接影響學生的學習行為，也因此被當作學習策略，並且對聽力理解能力具間接效果。因此，以信念為基礎的語言學習模式當它被應用在 EFL 聽力學習的特定領域時有了實證的依據。然而，研究結果指出幾個特殊條件必須要運用在這些通論上。所以，本研究亦具體討論並確認了上述所發現之多變項關係。本研究以經由實證之聽力聯合模式（CML）中認知和非認知各項構念之間的關聯來詮釋台灣大學生們的 EFL 聽力學習結果。研究者亦施行多次的前導研究以發展聽力學習信念（BELLA）、聽力策略理解（ELLSI）、運用聽力策略的能力（ISCEL）和被洛斯特（Rost）分類為「詮釋」的更高階、實務性類型的聽力能力等的衡量工具。衡量模式在 SEM 和 CML 內的特性與驗證過程之細節在本研究中亦被詳述及討論。最後，本研究在方法學上、理論層次以及教學應用上有更廣泛的意涵。這些可能引發的結果將按照現有的研究結果加以討論。關鍵字：聽力信念，聽力策略，聽力理解，試題反應理論，結構方程模式. i.

(3) ABSTRACT The fundamental premise underlying the present study is that learner beliefs affect the range of language learning strategies employed, thereby indirectly influencing L2 learning outcomes.. The preeminent proponent of this theory is. Horwitz, who developed the beliefs about language learning inventory (BALLI: Horwitz, 1985; 1988).. Decades of variable analytic research have corroborated the. links between language learning beliefs and learning behavior, often operationalized as L2 strategies.. To date, no research has specifically examined L2 listening. learning beliefs in relation to listening strategies and listening proficiency.. Given. that language learning beliefs can influence the learning behaviors of L2 students, the imperative to inventory L2 listening beliefs and examine their interaction with other individual difference (ID) variables exists.. The present study employs a Horwitzian. conceptual model as applied to the specific domain of EFL listening. Via Structural Equation Modeling (SEM), Item Response Theory (IRT), and Multi-dimensional Item Response Theory (MIRT) psychometric measurements, an empirically testable model of EFL listening learning was specified and identified in order to account for latent and observable trait interactions which influence EFL listening. Results support the Horwitzian notion of language learning beliefs acting directly upon learning behaviors, operationalized herein as strategies, and an indirect effect on ii.

(4) listening ability.. Therefore, the belief-based model of language learning has support. when applied to the specific domain of EFL listening learning.. However, the. findings indicate that several qualifications must be applied to these general statements.. The specific characteristics of the multivariate relations are identified. and discussed. The model identified and validated herein is designated as the Conjugated Model of Listening (CML) to denote the conjunction of cognitive and non-cognitive constructs to explain EFL listening learning outcomes for Taiwanese university undergraduate learners.. Multiple pilot studies were used to create measurement. instruments for listening learning beliefs (BELLA), listening strategic knowledge (ELLSI), listening strategic competence (ISCEL), and listening ability, i.e., higher-order, pragmatic-type listening abilities, classified by Rost (2005) as “interpretation”.. The details of measurement model specification, identification and. validation within the SEM, CML, are detailed and discussed. Finally, this study has far-ranging implications at the methodological, theoretical and pedagogical levels.. These implications are discussed in light of the current. findings. Keywords: Listening Beliefs, Listening Strategies, Listening Comprehension, Item Response Theory, SEM iii.

(5) Acknowledgments This dissertation could not be written without the wonderful outpouring of support from so many people.. First of all, thanks are extended to Dr. Thomas Tseng. (Wen-Ta), advisor to this study, who introduced me to the field of statistical modeling. Prior to his instruction, I had merely conceptualized SLA research as treatment studies, always influenced by those individual participant traits that were ever-present, but always assumed to be inscrutable.. I heartily recommend that all TESOL PhD. candidates spend the time and effort to delve into modeling and statistical pattern recognition techniques, as this knowledge can be usefully applied toward the evaluation of many complex issues both within and beyond the field of SLA. Next, I must thank all of my friends and colleagues who generously assisted me in data collection by volunteering their classes of university students.. In the course. of three years, these statistical models were honed on no less than 2,500 Taiwanese EFL learners. Thanks to Diana Lin, Dr. Eric Lin, Dr. Sherry Cheng, Dr. Zhang Yi-Ren, Dr. Annie Li, Dr. Daniel Wen, Su-Shan Chen, Daniel Rogenkamp, Dr. Chen Mei-li, Dr. Hsu Qian-Yan, Michelle Hsu, Dylan Black, Carrie Shen, and Barry Cho for helping to make this research possible. I would also like to express my gratitude to the review committee who provided invaluable suggestions for improving the layout and flow of the manuscript, as well as iv.

(6) good advice on strengthening the many stages of analyses.. Thank you to Dr. Lin Yi-. Hung, Dr. Lilian Cheng, Dr. Lee Sy-Ying, Dr. Wang Chaochang, Dr. Chen Po-Hsi and Dr. Tseng Wen-Ta. Finally, I must thank my lovely wife and better half, Vivi, who had to sacrifice many days of beautiful Taitung weather because I was trapped indoors in front of a computer all day- and who always acted as ad-hoc spokesperson whenever my in-laws asked, “Isn’t the ah-duo-ah finished with his PhD yet?”. Also, special thanks. go to my own parents who have waited patiently for eight years to call me their son, “the doctor.”. v.

(7) TABLE OF CONTENTS List of Tables. Page x. List of Figures. xii. CHAPTER ONE: INTRODUCTION Background and Motivation. 1. Rationale for Research. 5. Statement of Purposes Research Questions Significance of the Study. 6 7 8. Limitations of the Study. 9 10 10 18 20 21 23 27 28 40 46 49 52 52 54. CHAPTER TWO: LITERATURE REVIEW Definitions of Terms The Listening Process L1 Listening Research Empirical Studies of the Listening Process Qualitative Treatises on the Listening Process L2 Listening Research Strategy Use in L2 Listening Authentic Input for L2 Listening L2 Listening Test Design Summary of Listening Research The Present Study The Proposed Model: The Conjugated Model of Listening The Belief Component The Strategy Components Summary CHAPTER THREE: METHODOLOGY Research Design IRT vs. IFA Methods for Latent Trait Measurement Necessity for Multi-group Analysis IRT models for Data Analysis Subjects Instruments Beliefs on EFL Listening Ability (BELLA) Inventory EFL Listening Strategy Inventory (ELLSI) Inventory of Strategic Competence in EFL Listening (ISCEL) vi. 1. 62 77 82 82 84 88 90 95 96 96 99 109.

(8) Listening Test. 111. Test content. 112. Calibration Pilot Studies. 113 114. Validating Measurement Model of Beliefs. 115. Results of BELLA Pilot. 115. Validating Measurement Model of Strategic Knowledge Results of ELLSI Pilot. 127 128. Validating Measurement Model of Strategic Competence Results of ISCEL Pilot Validating Measurement of Listening Abilities. 139 140 147. Results of Pilot Test Construct Validity Review Results Implications for Main Study Questionnaire Instruments Listening Test Main Study Procedures Revised Model CHAPTER FOUR: RESULTS Measurement Models Beliefs on EFL Listening Ability (BELLA) Parallel Analysis EMIRT Reduction of the BELLA Model EFL Listening Strategy Inventory (ELLSI) Parallel Analysis MIRT Reduction of the ELLSI Model Inventory of Strategic Competence in EFL Listening (ISCEL) Parallel Analysis EMIRT Parallel Analysis on Reduced Inventory MIRT Listening Test Parallel Analysis IRT Summary of RQ1 Findings vii. 148 161 163 165 165 167 169 171 174 174 176 176 179 180 187 189 197 200 209 210 212 213 214 219 220 226 234.

(9) Structural Models. 239. Baseline CML (CML0). 240. First Alternative (CML1) Second Alternative (CML2). 245 248. Third Alternative (CML3). 252. Fourth Alternative (CML4). 255. Summary of Model Comparisons Research Questions 2-5. 258 259. Research Question 2 Research Question 3 Research Question 4. 259 260 262. Research Question 5. 266 268 268 270 271. CHAPTER FIVE: DISCUSSION AND CONCLUSION Overview of the Study Methodological Implications Response Labels and Scales The Consonance of Correlation- and Covariance-based Analyses Necessity of Multiple Cross-validations Identification of Test-measured Traits Theoretical Implications Future Elaborations of the CML Native English Speaking Instructors (NES) English Listening Practice (Practice) Gender (Sex) Future Research in Listening Learning Beliefs Pedagogical Implications Conclusion REFERENCES APPENDICES Appendix 1 : The BELLA Scale Items Appendix 2: Classic Reliabilities and Statistics of Original ELLSI Item Pool at Initial Trialing Appendix 3: Initial EFA of ELLSI Item Pool with JH/SH Subjects Appendix 4: Listening Exam Construct Validity Survey (sample screenshots) Appendix 5: CML Questionnaire Combining BELLA, ELLSI and ISCEL Scales Appendix 6: Histograms of Requisite and Essential Skill viii. 275 277 278 284 287 287 288 289 291 293 297 298 314 314 315 319 322 334 338.

(10) Endorsement for Individual Test Items Appendix 7: Results of Four-point Likert-type Scale Trials on ELLSI and BELLA Concurrent with ISCEL Pilot. ix. 342.

(11) List of Tables Table 1 Outline of the Study Procedures Table 2 ELLSI Indicators Used in Present Study. 85 107. Table 3 Two-Dimensional BELLA Structure in EMIRT with. 119. Undergraduates (n = 315) Table 4 Fit Comparisons of Hypothesized BELLA Measurement Models. 120. Table 5 BELLA Non-uniform DIF Test Results. 124. Table 6 BELLA Uniform DIF Test Results Table 7 ELLSI EFA Two-factor Solution for University Subjects (n = 315) Table 8 Fit Comparisons of Hypothesized ELLSI Models. 125 130 132. Table 9 ELLSI MIRT Two-dimensional Structure for University Subjects Table 10 Fit Comparisons of Second-order and First-order Factor Structures Table 11 ELLSI Non-uniform DIF Test Results Table 12 ELLSI Uniform DIF Test Results Table 13 Fit Comparisons of Hypothesized ISCEL Models Table 14 ISCEL EMIRT Estimates for One and Two Dimensions Table 15 ISCEL Non-uniform DIF Test Results Table 16 ISCEL Uniform DIF Test Results Table 17 1PL Estimates and Item Residualχ2 for Pool of Listening Test. 132 135. Items Table 18 Non-uniform and Uniform DIF Results for Stimulus-response Items Table 19 Non-uniform and Uniform DIF Results for Short Conversation Items Table 20 Content of Stimulus-response Items Flagged for Gender DIF Effects Table 21 Content of Short Conversation Items Flagged for Gender DIF Effects. Table 22 Constructs of Listening Exam Items by Listening Skills Table 23 BELLA Parallel Analysis Two-Factor Solution for Main Study, n =618 Table 24 Two-Dimensional BELLA Structure in EMIRT with main study participants , n = 618 Table 25 BELLA Non-uniform DIF Test Results in Main Study, n = 618 Table 26 BELLA Uniform DIF Test Results in Main Study, n = 618 Table 27 ELLSI Parallel Analysis Two-Factor Solution for Main Study , n x. 137 138 142 144 145 146 148 151 152 153 155 163 178 179 186 186 191.

(12) =618 Table 28 Cross-sample Comparison of Gender and Years of EFL Study. 193. with NES Instructors Table 29 Cross-sample Univariate Comparison of Means for BELLA and. 194. ELLSI Responses Table 30 Second ELLSI Parallel Analysis Two-Factor Solution for. 197. Reduced Inventory Table 31 Three-Dimensional ELLSI Structure in MIRT with main study. 199. participants (n = 618) Table 32 Original 2D ELLSI Structure in MIRT with main study participants (n = 618) Table 33 ELLSI Non-uniform DIF Test Results in Main Study, n = 618 Table 34 ELLSI Uniform DIF Test Results in Main Study, n = 618 Table 35 ISCEL Parallel Analysis Two-Factor Solution for Main Study, n =618 Table 36 Fit Comparisons of Hypothesized ISCEL Models with Main Study Participants (n = 618) Table 37 ISCEL MIRT Estimates and R2 for 1D Model Table 38 ISCEL Non-uniform DIF Test Results in Main Study, n = 618 Table 39 ISCEL Uniform DIF Test Results in Main Study, n = 618 Table 40 Listening Test Parallel Analysis Two-Factor Solution for Main Study, n =618 Table 41 Second Listening Test Parallel EFA Two-Factor Solution with 30 Items Table 42 1PL Difficulty Estimates for Listening Test with Main Study Participants (n = 618) Table 43 Measures of Item Fit for Reduced Listening Test in 1PL IRT Model Table 44 Non-uniform and Uniform DIF Results for Listening Comprehension Exam Items. Table 45 Structural Relations Among Latent Traits and Covariates in CML0 Table 46 Structural Relations Among Latent Traits and Covariates in CML1 Table 47 Structural Relations Among Latent Traits and Covariates in CML2 Table 48 Structural Relations Among Latent Traits and Covariates in CML3 xi. 199 208 209 211 213 215 218 219 222 224 227 229 234 244 246 252 254.

(13) Table 49 Structural Relations Among Latent Traits and Covariates in. 257. CML4 List of Figures Figure 1. Conceptual schematic of the currently proposed CML.. 54. Figure 2. The initial CML SR model.. 81. Figure 3. Overview of theoretical facets for construction of ELLSI. Figure 4. The unidimensional default Model 1D.. 101 118. Figure 5. The two-dimensional model previously identified, Model 2Do. Figure 6. Hypothetical alternative two-dimensional structure, Model 2D1. Figure 7. Hypothetical alternative two-dimensional structure, Model 2D2.. 118 118 119. Figure 8 Non-uniform DIF as seen in ICCs of item 18. Figure 9 Uniform DIF as seen in ICCs of item 15 Figure 10. Scree plot of ELLSI EFA with university subjects (n= 315). Figure 11. ELLSI Parallel analysis scree plot with university subjects (n =315) as observed sample. Figure 12. ELLSI Hypothetical second-order factor structure. Figure 13. ELLSI Baseline oblique, first-order factor structure. Figure 14. ISCEL Parallel analysis scree plot with university subjects (n = 475) as observed sample. Figure 15.Timeline of pilot results informing the analyses in the main study. Figure 16. The CML hypothesized at the outset of principal data collection.. 126 127 129 131. Figure 17. Parallel analysis scree plot of BELLA responses from main study sample (n =618). Figure 18. Parsimonious BELLA model with estimates for use in CML structural model. Figure 19. Uniform DIF as seen in ICCs of item 15 Figure 20. Parallel analysis scree plot of ELLSI responses from main study sample (n = 618). Figure 21. Second parallel analysis scree plot of 15 ELLSI items from main study sample (n =618). Figure 22. Parsimonious ELLSI model with estimates for use in CML structural model. Figure 23. Uniform DIF as seen in ICCs of item 5 Figure 24. Uniform DIF as seen in ICCs of item 7 Figure 25. Parallel analysis scree plot of ISCEL responses from main study sample (n = 618). xii. 134 134 142 171 172 177 182 185 190 196 205 207 208 212.

(14) Figure 26. Second parallel analysis scree plot of 15 ISCEL items from. 214. main study sample (n =618). Figure 27. Parallel analysis scree plot of listening test responses from main study sample (n = 618).. 221. Figure 28. Standardized solution to CML0 with significant paths shown.. 244. Figure 29. Standardized solution to CML1 with significant paths shown.. 247. Figure 30. Standardized solution to CML2 with significant paths shown. Figure 31. Standardized solution to CML3 with significant paths shown.. 251 255. Figure 32. Standardized solution to CML4 with significant paths shown.. 257. xiii.

(15) The Conjugated Model of Listening 1. CHAPTER ONE INTRODUCTION. Background and Motivation Theoretical assumptions of L2 listening research have traditionally been based on the dual strands of psycholinguistics and social constructivism. Within the psycholinguistic strand of research, listening is composed of sub-processes which are theoretically indifferent to the language of input.. Meanwhile research from the. social constructivist paradigm has postulated that interaction with interlocutors and situational variables palpably influence learners’ listening achievement, and similar to the psycholinguistic approach, the language of input and interaction is of marginal import to the theoretical discussion.. As Rost (2005) postulated, it is indeed true that. the physiological parameters of L1 and L2 listening processes are identical, as are the situational variables (e.g., face-to-face, dialogic, monologic, etc.); and it is the natural process of scientific inquiry to apply the knowns of L1 for comprehension of the unknowns of L2.. As such, much of the research in both domains presupposes that. L2 listening approximates L1 listening (Buck, 2001; Faerch & Kasper, 1986; Goh, 2000). Similarities aside, specific affective and cognitive variables still separate the two constructs (Bacon, 1992; Rubin, 1994). While native speakers aurally process.

(16) The Conjugated Model of Listening 2. their mother tongues rather effortlessly and automatically, L2 learners experience difficulties and delays in processing aural input (Flowerdew & Miller, 2005), of which individuals have varying degrees of tolerance.. It is universally accepted that some. individuals possess specific traits which enable them to persist in the face of learning adversities while others, lacking those traits, quickly relent in learning L2.. Thus,. variables under the rubric of individual differences (ID) create a large gulf between the respective knowledge bases of listening processes, which remains despite the borrowing of L1 constructs that are analogous to L2. The advent of ID research was intended to elucidate this gulf. A number of studies situated in the domain of ID research have found evidence of the discrete correlations that strategies, aptitude and attitude exhibit with L2 achievement in general, and L2 listening proficiency, specifically (Bacon, 1992; Elkahfaifi, 2005; Park, 2010; Teng & Chan, 2008; Vandergrift, 2006).. Yet, specification of the. interrelationships among these variables remains unsupported, thus the antecedent contributions of these variables to L2 listening ability remain tentative. As a consequence, the majority of extant research literature to date is “variable analytic” (Bodie, 2009), meaning that although empirical and informative, it remains unguided by comprehensive theory.. In other words, this research is grounded on hypotheses,. resulting in descriptions of pieces of listening theory, yet lacks the wider context.

(17) The Conjugated Model of Listening 3. within which to assemble these pieces.. As Bodie (2009) states, “Theory refers to a. systematic accounting of interrelated phenomena and why their relationships exist” (Bodie, 2009: p. 83). Thus, the present study aims to move toward the development of L2 listening theory by building a structural model of ID variables posited as antecedents to listening comprehension ability. Creation of a structural model of L2 listening requires extensive groundwork to determine the proper direction for research, i.e. creation and validation of measurement models for ID variables which comprise the structural model, all of which are based on conceptual models.. The present study will adopt a ground-up. approach to model construction, thereby allowing for meticulous and reliable measurement of the relevant constructs before integration into a complete covariate network to explain the interactions that yield L2 listening performance outcomes.. In. addition to creating an empirically testable model of L2 listening, it is expected that this research manuscript can also instruct future researchers with regard to latent trait measurement methodologies.. From a broad perspective it can be stated that this. dissertation fulfills both theoretical and didactic research needs in L2 listening and latent trait measurement. The starting point of the present study is the nexus of language learning beliefs, a construct under the umbrella term personal epistemology, and language strategies.

(18) The Conjugated Model of Listening 4. which enable or facilitate language processing and comprehension.. The. fundamental premise underlying language learning belief research is that learner beliefs affect the range of language learning strategies employed and also affect the motivation to learn, thereby indirectly influencing L2 learning outcomes.. The. preeminent proponent of this theory is Horwitz who developed the beliefs about language learning inventory (Horwitz, 1985: BALLI).. Other studies utilizing. BALLI have corroborated the link between beliefs and learning behavior, often operationalized as L2 strategies (Chang & Shen, 2010; Yang, 1999), and BALLI-based research has spurred more belief-on-motivation and belief-on-learning behavior research.. Graham (2006) found that self-efficacy beliefs determined motivation for. language learning among teen French as FL learners while Zhong (2010) found that Chinese ESL learners possess an array of language learning beliefs with mixed conducive and debilitative effects on self-regulated learning. Studies along this vein examine language learning beliefs in general, or self-efficacy beliefs in relation to language learning progress assessed via general tests of proficiency, or learner’s perceptions of progress.. To date, no research has. specifically examined L2 listening learning beliefs in relation to listening strategies and proficiency. Given that language learning beliefs can influence the learning behaviors of L2 students, the imperative to inventory L2 listening beliefs and examine.

(19) The Conjugated Model of Listening 5. their interaction with other ID variables exists. The present study will break new ground in this direction and will proffer an approach to exploring this issue.. Rationale for Research Variable analytic research has been informative to the extent that a number of ID variables such as strategic knowledge, strategic competence, beliefs and gender have been shown to possibly contribute to L2 listening performance.. However,. integration of these variables into a single theoretical model to investigate multivariate interactions antecedent to listening performance has yet to occur.. As a. case in point, gender, arguably one of the most crucial ID variables across SLA research has been shown to have contradictory results in numerous effect studies.. It. is now widely accepted that gender alone cannot reliably explain learning performance outcomes, but must be examined in concert with intervening variables. Likewise, treatment studies incorporating language learning strategies often yield mixed or contradictory results (Beal, Gill-Rossier, Tate & Matten, 2008). Consistently inconclusive results are the sign of interaction effects or effects of other unknown variables, and these cannot be adequately examined with the variable analytic research paradigm.. A more informative research approach would be to. utilize structural equation modeling (SEM) and measurement to account for these.

(20) The Conjugated Model of Listening 6. effects with empirical models which can also measure errors and model disturbances. In addition, SEM analysis allows for comparison of alternative models within a single study, which can shed light on the structure of the multivariate relations. SEM analysis is the most efficient means to test theoretical models.. In this way. To date,. structural modeling is not a de-rigeur analytical approach within the field of L2 listening research, and there are no structural models incorporating putative independent variable and covariate interactions to explain listening performance outcomes.. Statement of Purposes The present study will first measure four latent trait variables: knowledge of listening strategies, listening strategic competence, beliefs of listening learning and listening ability, using IRT analysis of questionnaire data (Graded Response Model: GRM) to verify invariance of and identify sources of differential item functioning (DIF) within the measurement models, meanwhile one parameter logistic (1PL) IRT analysis of dichotomous listening item responses will verify invariance of and identify sources of DIF within the measurement of listening ability.. Subsequent to. measurement, a structural model will be constructed, incorporating the theoretical latent variables and DIF variables into an empirically testable model to explain EFL.

(21) The Conjugated Model of Listening 7. learners’ listening learning outcomes.. In the process of conducting the present. research three novel latent trait questionnaire inventories will be developed and validated.. Thus, this study is proposed with two broad purposes in mind: to advance. L2 listening research by developing both measurement instruments and a testable model, and to serve as a handbook for construction of listening comprehension models.. Research Questions Extant research and theory clearly indicate that language learning beliefs interact with other variables to influence learning outcomes.. Applying this proposition to the. specific domain of L2 listening yields the research hypothesis that a model of L2 listening processing must invariably take into account L2 listening learning beliefs. Thus, the main research question guiding this study is: Do listening learning beliefs impart measurable and significant effects on listening learning outcomes via listening strategic knowledge and listening strategic competence? Investigation of this question entails a number of subordinate questions which undergird the study. 1) What are the dimensional structures underlying the measurements of listening.

(22) The Conjugated Model of Listening 8. learning beliefs, listening strategic knowledge, listening strategic competence and listening comprehension? 2) What are the relative relationships of listening learning beliefs to listening strategic knowledge and listening strategic competence? 3) What are the relative relationships of listening learning beliefs, listening strategic knowledge and listening strategic competence to listening comprehension? 4) What are the relative relationships of listening learning beliefs, listening strategic knowledge, listening strategic competence and listening comprehension to the pertinent covariates of gender, listening practice and native English speaker (NES) instruction? 5) Can the identified structural model adequately explain the pattern of observed responses and be validated?. Significance of the Study Whereas no empirically testable, structural model of L2 listening theory exists to date, and consequently, understanding of listening learning outcomes resultant from multivariate interactions is hampered, the present study serves as an important step forward for informing listening theory and pedagogy.. This study will provide a. prototypical structural model of listening learning which can be further developed and.

(23) The Conjugated Model of Listening 9. refined by future scholars.. Limitations of the Study The main objective of developing a prototypical model imposes an inherent limitation on the scope of the study with respect to the number of explanatory variables to listening comprehension.. Only three latent variables and three observed,. causative variables are proffered as antecedent to listening comprehension outcomes. It is not suggested that such a small number of variables are sufficient for explanation of listening outcomes.. Instead, corroboration / rejection of a parsimonious model. identifies the factors necessary for listening comprehension and validates the theoretical presumptions of antecedence/causation.. As SEM analysis invariably. involves the comparison of alternative models, the inclusion of additional theoretical variables can be validated in the future, subsequent to establishing the prototypical structure..

(24) The Conjugated Model of Listening 10. CHAPTER TWO LITERATURE REVIEW. The chapter first provides definitions of terms used throughout the manuscript. Next, the chapter provides a description of the listening process which constitutes the theoretical bases of the latent constructs, and subsequently reviews, analyzes, and critiques extant research on L1 and L2 listening.. Finally, this chapter presents the. basic conceptual model which informed the procedures adopted during the piloting and main stages of analysis.. Definitions of Terms The two major constructs which comprise the model of L2 listening proposed herein are often accompanied by confusion regarding overlapping terminology.. This. section will proffer definitions of the various constructs and nomenclature as situated in the present research.. Beliefs The concept of beliefs is very broad within the domain of SLA theory, often overlapping with perceptions and opinions.. Partly as a consequence, various.

(25) The Conjugated Model of Listening 11. researchers have proposed a number of specific types of beliefs in order to precisely delineate the construct in question.. It should be made clear that all belief types are. considered trait variables in the present study, meanwhile perceptions and opinions are state variables which change according to contexts. This does not imply that beliefs are absolutely immutable; rather, they are sufficiently resistant to change so as to be distinguishable from opinions. Beliefs may be malleable if an individual encounters sufficient contradiction in observations and experience to effect a cognitive dissonance with the sum of personal knowledge undergirding the beliefs. The definition of general beliefs applied to this paper posits them as a way of viewing the world, and the individual’s relation to the world, which is based on prior knowledge or enculturation, and which is stable across situational contexts. Furthermore beliefs involve an affective judgment which is the stabilizing force that often bolsters them against contrary observations made by the individuals holding the beliefs.. Epistemic Beliefs Epistemic beliefs, also referred to as epistemological or learning beliefs by Schommer’s (2004) framework, are a specific type of beliefs concerning the nature of learning, including one’s capability to learn and speed of learning, and the nature of.

(26) The Conjugated Model of Listening 12. knowledge, including the stability of knowledge, the structure of knowledge, and the source of knowledge.. A number of scholars disagree over the proper conceptual. scope of epistemic beliefs; most notably, Pintrich and Hofer (1997) and later works by said researchers propose that learning beliefs comprise a construct highly related to epistemic beliefs, but are not strictly epistemological, as learning of skills and abilities may be distinct from the learning of knowledge. said to span two different constructs.. As such, Schommer’s framework is. To maintain conformity with the bulk of. research, “epistemic/epistemological” beliefs as mentioned herein pertain to personal theories of knowledge and knowledge acquisition, while “learning” beliefs pertain to the nature and acquisition of skills/abilities, only a subset of Schommer’s belief system.. The present study will utilize a domain specific system of learning beliefs,. although numerous citations to epistemic beliefs will also be found in the literature review.. Listening Learning Beliefs Listening learning or listening beliefs, comprise a focal construct in the present study, which are proposed to be a domain-specific type of language learning belief concerning the nature of learning to listen.. Listening learning beliefs include beliefs. of optimal ways to develop and enhance listening skill and evaluation of one’s.

(27) The Conjugated Model of Listening 13. potential for developing skill, and beliefs concerning the nature of listening ability, which includes judgments of requisite skills and the necessity of listening to facilitate learning processes.. Self-efficacy Beliefs Self-efficacy beliefs are part of an individual’s ego with respect to personal abilities.. Self-efficacy beliefs are often posited in educational and SLA research as. beliefs in one’s ability to perform a specific task.. Therefore, self-efficacy beliefs are. sometimes portrayed as state variables as they are context-specific.. However, it is. best to consider self-efficacy beliefs as quasi-state variables that are half-way between traits and states.. This is because the individual evaluates his/her capability within a. given context relative to a stable, trait-based belief of general ability that is integrated with self-image, and is resilient to change.. Because the ego is the underlying. construct of self-efficacy, these belief statements are typically characterized by the propositions of “I can”.. Strategies The generic term “strategies”, as used herein, refer to learning strategies, i.e., mental and /or observable, goal-directed procedures to facilitate a learning task..

(28) The Conjugated Model of Listening 14. Recently acquired strategies are consciously evoked by the learner, but may become part of automatic processing after the learner develops familiarity with learning tasks. Despite automaticity of strategies, they may be recalled to awareness if necessary. Also, recalled strategies are presumed to illustrate some aspects of cognitive processing (Chamot, 2005).. Strategic Knowledge The two aspects of strategies extant in research literature are strategic knowledge and strategic competence.. The latter is an ability to use strategies to good effect, and. this element of success causes construct ambiguity between competence and skills (defined below). and state variants.. Furthermore, strategic knowledge and competence both have trait The present study utilizes trait strategic knowledge, which is. knowledge of extant strategies, and knowledge of the suitability of strategies for situations.. As strategic knowledge is knowledge of mental processing, strategies and. heuristics, it also appears in the research literature as “metacognition”.. Strategic. knowledge is a trait variable because knowledge is stored in long-term memory, rendering it stable across contexts.. In short, trait strategic knowledge is experiential. memory of habitual strategic use.. The term strategic knowledge is adopted herein to. avoid confusion with the state strategic use variable which also appears often in the.

(29) The Conjugated Model of Listening 15. research literature.. Strategic Competence The term strategic competence refers to one’s personal evaluation of the ability to use strategies to good effect.. Strategic competence occurs in both trait and state. aspects, with the state competence variable also referred to as strategic efficacy.. The. key distinctions between strategic competence (trait) and strategic efficacy (state) are the temporal reference for the evaluation of strategy use and the breadth of strategy use.. Strategic competence reflects one’s confidence in applying extant strategies. based on summative experiential evidence.. In this regard, strategic competence is. akin to strategic knowledge in its dependence on long-term memory for manifestation on questionnaire instruments and reflects individuals’ qualitative evaluations of habitual implementation of one’s full strategic repertoire.. In contrast, strategic. efficacy reflects one’s confidence in utilizing strategies to good effect on a specific task or instantiated strategy use, thereby necessarily limiting the scope of the construct by linking it to the relevant subset of strategies.. Listening Strategies Listening strategies are a type of strategy invoked when faced with a listening.

(30) The Conjugated Model of Listening 16. task.. They are mental and /or observable, goal-directed procedures to facilitate a. listening task.. Listening strategies also share the same feature of quasi-automaticity. as general learning strategies, and recalled listening strategies are presumed to illustrate some aspects of listening comprehension processing.. Knowledge of Listening Strategies This is a type of strategic knowledge specific to the domain of listening, which is comprised of knowledge of extant listening strategies, and knowledge of the suitability of such strategies for situations.. This construct, also referred to as. knowledge of strategic listening, is a trait variable because such knowledge is stored in long-term memory, rendering it stable across contexts.. Competence in Listening Strategies This is a type of strategic competence specific to the domain of listening, which is comprised of one’s qualitative evaluations of listening strategic knowledge.. This. construct is also a trait variable as evaluations of strategic competence are referenced with habitual listening strategy use which is stable across contexts.. Skills.

(31) The Conjugated Model of Listening 17. The generic term skills, as used herein, refer to the mental processing of new information or knowledge together with previously acquired knowledge that successfully accomplishes a learning task.. Thus, skills are abilities to draw upon. knowledge to yield successful outcomes.. Skills are dependent to some degree on. mental processing which encompasses strategies, and also breadth and depth of knowledge; therefore, researchers and laypeople often commingle the construct of skills with knowledge and strategies.. Skills can be distinguished from automatic. strategies in that skills by definition produce successful outcomes, whereas strategies may fail, and the subtle nature of skills renders them intractable to recall.. Listening Skills The term listening skills, a sub-set of language skills, refer to the mental processing of aurally received information wherein input is combined with previously acquired knowledge to successfully accomplish a listening task.. Thus, listening. skills are abilities to perform all stages of the listening process resulting in successful comprehension of the message..

(32) The Conjugated Model of Listening 18. The Listening Process A basic understanding of scholars’ conceptualizations of listening is requisite for determining the foci of theoretical models of listening ability.. A number of scholars. (See Goh, 2000; Nunan, 2002; O’Malley, Chamot & Küpper, 1989; Sun, 2002) have adopted a three-stage process of listening which is loosely based on two separate but contemporary researchers’ conceptualizations: J.R. Anderson’s process of perception, parsing and utilization and A.D. Wolvin and C. G. Coakley’s process of reception, attention, and meaning assignation.. The processing stages of both conceptual. frameworks do not neatly overlap, but rather meld into a generalized three-step mental procedure with Anderson’s processes subsuming Wolvin and Coakley’s. Anderson proposed that perception is characterized by the listener’s attention on sounds and decoding. In other words, the listener first hears and begins to listen in earnest by differentiating speech input from other aural signals such as random background noise.. Decoding of speech also occurs in this first process, and thus. Anderson’s perception combines Wolvin and Coakley’s two stages of receiving and attending to aural stimuli.. Anderson’s parsing stage refers to the transformation of. perceived words into meaningful representations wherein the listener must recall and correctly match the referents (tangible and abstract) to the words.. This stage. corresponds to Wolvin and Coakley’s assignation of meaning insofar as parsed words.

(33) The Conjugated Model of Listening 19. are concerned.. However, Wolvin and Coakley spoke generally of meanings of aural. stimuli, thus their final stage also overlaps with Anderson’s more precise concept of utilization, wherein a mental representation of the entire message is mapped to the listener’s background knowledge.. Taken together, these hypotheses form the. research assumptions for a wide swathe of L2 listening research to date wherein separate stages and even the entire process of listening comprehension are examined. However, as some scholars have correctly noted, the three-step listening comprehension process described above is limited to one-way listening, and quite frequently listening is a two-way interaction; i.e., it is one half of interpersonal communication which involves real-time, attentive listening to an interlocutor’s speech.. Research into listener responses and outward signs of comprehension (or. comprehension break-down) contributes to our understanding of the meaning construction processes in listening (Gardner, 1998; Lynch, 1997; Nunan, 2002; Oprandy, 1994).. Naturally, the present study must extend the conceptualization of. listening beyond three sub-processes.. Feyten (1991) in seminal research on listening. comprehension ability and L2 proficiency recognized the need for a response phase with verbal or non-verbal cues to indicate the listener’s comprehension of the message. Rost (2005) echoes this idea and provides a succinct combination of the notions described thus far.. He suggests that the listening process consists of three processing.

(34) The Conjugated Model of Listening 20. phases which are simultaneous and parallel: decoding, comprehension, and interpretation, and which should be accompanied by a final phase of listener response. The three processing phases are reformulated nomenclature for the stages previously known as perception, parsing, and utilization, while the listener response phase situates the listening process into the paradigm of listening comprehension research.. In Rost’s notion, the construct of listener response is theoretically vague,. posited as a requisite component because listening is an unobservable mental process which must be indicated to other entities.. In social constructivist strands of research,. listener response is crucial for the construction of discourse and negotiation. Likewise in constructivist-based research, it is the sine-qua-non, as there can be no measurement of comprehension without listeners’ selection of answer options based on their interpretation/utilization of the aural stimuli.. It should then be understood. that the undergirding framework of listening which shapes the construction of the instrumentation for the present study is Rost’s four-stage listening process comprising the unobserved mental activity at the heart of listening proficiency.. L1 Listening Research After reviewing the prevailing conceptualizations of listening, the respective.

(35) The Conjugated Model of Listening 21. contributions of the four listening stages in supporting research frameworks can be seen with greater clarity.. The following sections will briefly review key findings in. L1 research literature which inform our current understanding of L2 listening.. Empirical Studies of the Listening Process Seminal L1 research (McDevitt et al., 1990; McDevitt, 1990), investigated children’s and caretakers’ beliefs on effective listening and proper listening behavior, and found that at early school ages, individuals equate effective listening with attention and that attentive listening is a good academic habit.. Data was gathered via. interviews of young children between the ages of seven and eleven and their mothers; therefore, strict adherence to deconstruction of “attention” into the previously mentioned theoretical nomenclature was impractical.. However, it is certain that. participants asserted the primacy of attention as a necessary factor for understanding the aural input, i.e., classroom instruction of elementary school teachers. In McDevitt’s line of research, the participants’ meta-cognition of three stages of aural processing, conglomerated into the common parlance “attention”, is shown to be present at early ages of learning, and crucially, the importance ascribed to attentiveness is shared by both youngsters and adults.. Furthermore, that the. participants seek to understand the aural message shows that listening is a.

(36) The Conjugated Model of Listening 22. goal-directed activity, a theme that recurs throughout the breadth of L1 and L2 listening research literature. In the case of McDevitt (1990) and colleagues (McDevitt et al., 1990), the results corroborated the accepted notion of the listening process, as both researchers and lay people recognize that mental processing must occur for the desired listening outcomes to be manifested.. Other researchers soon began to query the nature of listeners’. attention and comprehension goals.. Research into the nature of listeners’ goals. posited the existence of listening orientations, a type of ID variable, which determines the parts of the aural message to receive primary attention. Bodie and Worthington (2010) describe the construction of the listening styles profile comprised of four latent constructs: people-, action-, content-, and time-orientations of listening wherein people-oriented listeners primarily attend to the affective states of the interlocutor; action-oriented listeners attend to and understand the logical structures of the interlocutor’s message, and even listen for errors or inconsistencies; content-oriented listeners attend to the whole message by listening for details which support the interlocutor’s opinions; and time-oriented listeners primarily attend to the time constraints which they impose on the interlocutor to convey the aural message. last orientation is the least desirable type from a social-interactive perspective, as time-oriented listeners are characterized as interrupting others and exhibiting. This.

(37) The Conjugated Model of Listening 23. outwardly hurried behavior while listening.. As a consequence, the listener’s. attention is drawn away from the aural input and processing of the message is impaired. The significance of examining listening goal orientations lies in understanding that individuals’ end goals for comprehension and communication attenuate the attention paid to the message.. In other words, when individuals activate attentive. listening processes, they are seeking to comprehend specific types of information embedded in the aural input such as the gist, details, or even the tone and attitude of the speaker.. Such a directed execution of attention suggests that some variance in. the quantity of information gleaned through aural transmission is accounted for by ID variables, in this case, referred to as listening goal orientations or styles.. Listening. goals are not equivalent to the notion of listening beliefs that will be enumerated in later sections, but the existence of an antecedent to the cognitive processing variables entailed by the listening process, makes a Horwitz-type, theoretical conjecture plausible.. Qualitative Treatises on the Listening Process In a position paper on the state of L1 listening training research, Beal et al. (2008) asserted the need for continuing study of listening training programs and strategy.

(38) The Conjugated Model of Listening 24. instruction in order to clarify the factors which confound the treatments.. Beal et al.. postulated that listening orientations as well as listener beliefs impact the efficacy of listening strategy training and listening education treatments.. Summarizing the. findings of earlier research, they claim that “many students do not have a clear concept of listening as an active process that they can control” (p. 128).. Furthermore,. when students believe they have no control over listening processes, strategy instruction has ambiguous effects. With this in mind, Beal et al. continue to call for increased study of meta-cognitive strategies, as this variety of strategic knowledge applies directly to control of cognition and listening processing. Beal et al.’s (2008) position is to address the problems in listening treatment research with an activist approach.. Recognizing that certain strategy training. programs fail to yield reliable results, they call out to listening researchers for a different tack in strategy training, i.e., emphasis on meta-cognitive strategies to counteract learners’ lack of self-control over their listening processes.. One should. pause to take in the implications of the scholars’ position: when individuals do not hold beliefs that they actively control the listening process, listening training programs lose efficacy.. Thus, the express purpose of meta-cognitive training, Beal. et al.’s proposed remedy, is to imbue learners with beliefs about the manner of controlling the listening process.. Thus, Beal et al. essentially posits an inchoate.

(39) The Conjugated Model of Listening 25. construct of listening beliefs as antecedent to listening processing, the conceptual model to be utilized in the present study with respect to L2 listening.. In this way, the. Horwitzian notion of language learning beliefs acting on language learning processes is transferred to the specific domain of L2 listening learning. Interestingly, a subsequent paper on listening theory by Haroutunian-Gordon (2011), which is unique among the literature due to its use of Plato’s dialogues to derive a listening theory, directly points out the primacy of listeners’ beliefs with respect to the listening process.. Haroutunian-Gordon applies a discourse analysis to. Plato’s Symposium wherein Socrates primarily assumes the role of listener, thus, positioning the paper into the social-constructivist strand of research.. Her approach. is to construct a “listening philosophy”, a component in the wider theory of dialogic interaction, which posits four categories of listening beliefs: the aim of listening, the nature of listening, the role of the listener, and the relation between the listener and speaker.. The first two categories of beliefs have clear implications for listening. theory should they be substantiated by empirical data.. This is because incorporation. of beliefs on the aim of listening places the construct of listening belief antecedent to goals, that is, one must believe that listening during dialogic interaction will serve the learning goal which brings one into the dialog. Thus, every individual who listens with a purpose in mind is acting on a latent belief. The nature of listening, according.

(40) The Conjugated Model of Listening 26. to Haroutunian-Gordon, is the drawing of inferences with regard to the listening aim and mentally reformulating inquiries if the reasoning is occluded by situational idiosyncrasies.. In this manner, the two dimensions of aim and nature are theorized. as interdependent.. From the quantitative perspective, when the data reflect this. relationship, we say that the dimensions co-vary, an important a priori assumption when constructing a structural model of listening. The latter two categories of listening beliefs are specific to dialogic interactions, which limits the utility of incorporation into a structural model of listening that must also accommodate modern-day academic listening tasks, often one-way, audio recordings for practice or exams.. Yet, it is interesting to note the intersection of. Haroutunian-Gordon’s philosophical analysis of the listening process with the definition used herein. The role of the listener is to endeavor to make the inferences which comprise the nature of listening and reformulate questions when necessary and to endeavor “to provide an appropriate response” [emphasis added] (p. 135). Meanwhile, the relation between listener and speaker, according to HaroutunianGordon, sums up the dialogic interplay wherein the speaker must say things that engage the listener to attend the message, while the listener must convey to the speaker whether the inferences can be made based on what was heard, again harkening to the response phase of the listening process..

(41) The Conjugated Model of Listening 27. It is remarkable that Haroutunian-Gordon (2011), independent of cognitive research of listening processing, was able to construct a philosophy of listening which mirrors the four-step listening process.. In this construct of listening, beliefs play a. crucial role at the outset of the process.. Individuals must first believe that attentive. listening will lead to comprehension of information to meet their aims.. They must. then actively process the aural message to draw inferences and further match these inferences to a matrix of assumptions constructed of prior discourse turns and background knowledge.. These dimensions of Haroutunian-Gordon’s listening. philosophy correspond to the three active processing phases of listening, i.e. Rost (2005)’s decoding, comprehension and interpretation- also known in common parlance as the mental “attention” identified by McDevitt (1990).. Finally, according. to Haroutunian-Gordon, the listener must continually provide responses within a dialogue so that the interlocutor may realize the listener’s comprehension or lack thereof, which is analogous to Rost’s response phase.. L2 Listening Research With the above understanding of pertinent L1 listening research in place, the following section will briefly summarize L2 studies which extend from L1 listening theory.. The reader may already be aware that variable analytic studies of strategy.

(42) The Conjugated Model of Listening 28. use and efficacy of interventions comprise a large portion of L2 listening research. However, in the following section it will be shown that listening learning beliefs implicitly comprise a major portion of the research despite the overt focus on strategies and strategic knowledge.. Strategy Use in L2 Listening The predominant topic of L2 listening research is the relationship of strategy use and listening comprehension outcomes.. Stepping back from the literature. meta-analytically begs the question of why strategies are so actively studied.. It is. obvious that the research goals and hypotheses attest to the researchers’ collective belief that strategies are useful, and therefore should be taught.. Seminal strategy. research has indirectly posited that listening strategies are necessary for development of L2 listening proficiency (Graham & Macaro, 2008; Oxford & Nyikos, 1989; Seo, 2005; Thompson & Rubin, 1996) via the research assumptions, and currently the academic community at large reflects the belief in the utility of listening strategies. As Seigel (2011) puts it, teaching listening strategies is equivalent to teaching “how to listen” and is therefore “valuable” (p.320). A long road of research has led to current valuation of listening strategies. Listening strategy research began as an outgrowth of L2 strategy research,.

(43) The Conjugated Model of Listening 29. characterized by studies that were descriptive in nature, seeking to observe differences in strategic repertoires of focal groups such as gender (Bacon, 1992) high and low achievers (Teng & Chan, 2008) or effective and ineffective L2 listeners (Park, 2010). Bacon (1992) examined female and male adult learners of Spanish to ascertain differences in listening comprehension and strategy use when challenged with authentic listening passages.. Subjects were randomly placed in cohorts to listen to. counterbalanced passages on two separate topics, which putatively represented distinct difficulties.. Data on listening strategy use and perceptions of difficulty were. collected via retrospective interview protocols.. It was found that despite parity in. listening comprehension across both passage types, evidenced by post-listening recall, difficulty perceptions and the strategic repertoires of men and women differed considerably, with both genders adjusting metacognitive and cognitive strategies to suit perceived difficulty.. Overall, women displayed a predilection for metacognitive. strategies, while men preferred cognitive strategies, but women tended to utilize metacognitive strategies such as monitoring and evaluation on the reportedly easier passage.. In contrast, the more difficult passage heightened men’s tendency to rely. on cognitive strategies such as translation and bottom-up processing.. Thus, passage. difficulty had an interaction with gender wherein it amplified the gender’s respective tendencies for either metacognitive or cognitive listening strategies. Bacon surmised.

(44) The Conjugated Model of Listening 30. that strategy training was still necessary for learners to become aware of the situations which render certain strategies more effective than others, and that learners must be exposed to listening tasks of varying difficulties in order to develop flexibility in strategy use, a kind of metacognitive prowess. Thus, in the final remarks of the study Bacon implies that metacognition, i.e., awareness of listening stratgeies, is the key for learners to improve their listening comprehension. Teng and Chan (2008) provide a more updated view which bolsters the emphasis that Bacon (1992) placed on metacognitive listening strategies.. Teng and Chan. borrowed many elements of Bacon’s seminal study including passage recall as a measure of performance and verbalized retrospections of strategy use.. Teng and. Chan’s study differs from Bacon in that the scope of strategies surveyed was narrowed down to only four categories of metacognitive strategies: planning, monitoring, evaluation, and problem identification; the subjects were Taiwanese university EFL learners; and grouping of cohorts was by proficiency level, rather than gender.. It was found that proficient listeners utilized more planning, monitoring and. evaluation strategies than less proficient learners.. A point mentioned by the. researchers is the increased tendency for less proficient listeners to report the strategy of problem identification.. The researchers surmised that the finding is a. consequence of the cohort’s lower listening proficiency, i.e., they simply encountered.

(45) The Conjugated Model of Listening 31. more problems in listening, hence they identified more.. The researchers also noted. that strategic knowledge by itself may be insufficient to affect listening outcomes. Rather, learners must also actively utilize the most appropriate strategy for the given situation, a commonly mentioned point in language learning strategy research. Park (2010) combined quantitative and qualitative methodologies to analyze the strategic listening repertoires of Korean tertiary EFL students.. The effective and. ineffective EFL listeners were identified by TOEIC scores and then administered a translated version of Rebecca Oxford’s strategic inventory for language learning (SILL) questionnaire (the 1990 version as cited by Park).. It was found that effective. listeners utilized significantly more strategies than ineffective listeners as delineated by the SILL.. Subsequently, a small sample of eight individuals (four effective and. ineffective listeners, respectively) was selected for qualitative analysis of strategy implementation via think-aloud protocol.. Qualitative analysis was framed by a. simpler framework of metacognitive, cognitive, and socio-affective strategies, the well-known trichotomy often attributed to O’Malley and Chamot. Ineffective listeners reported acquiescence and distraction of thoughts after judging the aural text as being too difficult, whereas effective listeners maintained attention even in the face of distraction by nearby construction work (p. 7).. Both groups reported the cognitive. strategy of prediction; however, effective listeners made accurate predictions while.

(46) The Conjugated Model of Listening 32. ineffective listeners predicted erroneously, leading to confusion further in the audio stream.. Also, although both groups attempted to parse the audio stream, effective. listeners comprehended the utterances at the sentence level, while ineffective listeners parsed at the word level and frequently obstructed themselves by attempting to recall Korean translations.. Thus, when viewed in the tripartite framework of strategies,. effective and ineffective listeners both avail themselves of the same types of strategies, e.g., meta-cognitive and cognitive strategies, but the aspect which discerns the two groups is how to implement those strategies.. In this case, a tripartite taxonomy of. strategies alone was insufficient to identify the difference between participant groups. When applying an additional facet to the taxonomy of strategies (top-down versus bottom up) it could be seen that effective and ineffective listeners differed in implementation of bottom-up (parsing) listening strategies. Although Park (2010) is not as renowned as seminal work by the likes of Oxford or Bacon, the findings warrant recognition in that effective L2 listeners not only have more strategies available than ineffective listeners, but they are also more adept in using them.. In other words, Park’s findings imply a two dimensional structure to. listening strategies: knowledge of extant strategies and knowledge of appropriate implementation.. The reader should also recall that this implication echoes the. conclusion of Teng and Chan (2008) wherein knowledge of listening strategies is.

(47) The Conjugated Model of Listening 33. insufficient; learners must also have procedural knowledge to use the strategies to good effect. By viewing strategies with this dichotomy, i.e., the “what” and the “how”, effective L2 listeners can be discerned from ineffective listeners. These three studies span two decades of L2 listening research yet they encapsulate the essence of the descriptive, strategy-analytic genre of studies. Within these research papers it can be seen that listening strategies are deemed valuable, as they are associated with successful listeners.. In the concluding remarks of all the. studies, the authors call for continued efforts to teach listening learning strategies to EFL learners to facilitate their progress.. This is a salient belief held by researchers. which spurred the research for such an extensive length of time, and even continues to do so today.. However, the limitation of this research genre is that dependency/. independency of the variables can never be established.. Strategy use is associated. with better listeners, but within these study designs it is unclear whether the superior groups listen better on account of strategies, or whether the strategy use is an artifact of the L2 proficiency. As research of listening strategies continued to evolve, focus turned to treatment/intervention studies to effect growth of strategic knowledge and competence (Carrier, 2003; Cross, 2009; Graham & Macaro, 2008; Seo, 2005).. The. impetus behind this genre of research was to confirm or disconfirm the belief in the.

(48) The Conjugated Model of Listening 34. facilitative effects of strategies posited by the descriptive, strategy-analytic studies. Research under the treatment study paradigm attempts to discern effects on listening performance attributable to strategy use by measuring gains on listening assessments subsequent to an intervention of strategy training.. By using listening achievement. tests as the external criterion of validity, researchers were in effect positing first, that strategies had a significant effect on listening learning outcomes, and, second, that strategy instruction had useful payoffs in achievement vis-à-vis the time and effort required to learn the strategies. Carrier (2003) investigated the efficacy of a listening strategy intervention program on bilingual Hispanic high school students in a mandatory ESL program and found significant gains in discrete and video listening.. Carrier also acknowledged. the primacy of metacognitive components in strategy training by making explicit instruction the defining feature of the intervention, which was based on O’Malley and Chamot’s proposition that a learning strategy requires deliberate and conscious effort to be applied effectively.. The strategy intervention program consisted of bottom-up. and top-down cognitive strategies for one-way listening, such as listening for stress and intonation on focus words or inferencing from titles and visual cues. Metacognition was imparted to the subjects solely through the training procedure which incorporated definitions of the strategies and modeling by the instructor with.

(49) The Conjugated Model of Listening 35. think-aloud protocols. Carrier (2003) notes that the success of the intervention was due to the explicit instruction which taught the situations that require specific strategies and the manner of using them.. Thus, the primary implication of Carrier’s. study is that effective instruction of cognitive listening strategies should not be excised from metacognitive instruction. Cross (2009) investigated the listening learning gains of an intervention group receiving listening strategy training versus a control group which received conventional listening training.. The learners were adult Japanese, advanced level. EFL students at a private language learning center.. The listening training common. to both groups consisted of BBC video news segments, i.e., authentic input without pedagogical enhancement.. It was found that significant gains were made within. groups, while the gains of the intervention group over the control were not significant. Cross explained these results as being caused by the control group’s natural proclivity for using strategies similar to the intervention group and the tendency of some intervention group members to poorly utilize strategies. That Cross (2009) yielded inconclusive results is not surprising.. In the. literature review to his study, it was mentioned that inconclusive results of previous listening strategy intervention studies was one element of the motivation behind his research.. Citing previous studies, Cross pointed out that ID variables such as beliefs.

(50) The Conjugated Model of Listening 36. and habits also modulated the efficacy of strategy utilization.. Moreover, Cross. reiterated that application of the appropriate strategy for the given situation is as important as the breadth of strategies in one’s repertoire- the point made earlier by Teng and Chan (2008) and also the conclusion reiterated later by Park (2010) discussed above.. A final point about Cross’s research should be stressed: at the. outset of the study, the participants already had a repertoire of listening strategies available.. This observation means, first, that L2 learners may develop their own. notions of how to listen independently of and prior to any explicit instruction provided by educators or researchers, and second, learners’ unique proclivities for certain strategies are based on their own beliefs of how to pay attention to L2 aural input.. It should be recalled that this assumption has been corroborated by. McDevitt’s line of research in L1 listening, wherein young learners had already engendered nascent listening learning beliefs.. The crux of the issue, then, for L2. strategy intervention studies is whether the researchers can persuade the learners to adopt and utilize the target strategies to good effect, i.e., to replace an old belief set with a new one. In contrast to the ambiguity of Cross (2009), Graham and Macaro (2008) reported success in teaching listening strategies to British learners of French as a second language (approximately third year of high school in the American system)..

(51) The Conjugated Model of Listening 37. Their study incorporated a number of methodological and theoretical considerations which were claimed as being responsible for the positive learning outcomes.. First,. they incorporated a needs analysis component wherein participants were surveyed on the strategies they used a priori, so that the strategy intervention would instruct participants in novel strategies.. Second, they utilized a listening test with tasks. different from the training tasks, so that there would be no bias favoring the intervention group over the control group.. Third, they utilized a strategy training. program which taught clusters of strategies, rather than strategies in isolation.. And. finally, they incorporated self-efficacy training in the intervention, so that participants constantly monitored and evaluated their own successes in using the intervention techniques.. It should be recalled that self-efficacy is a belief in one’s ability to. perform a specific task (Bandura, 1986, 1993).. Essentially, Graham and Macaro. (2008) were reformulating the participants’ pre-existing belief systems by linking successful task completion with the utilization of the novel, intervention strategies. Participants increased their perceptions of self-efficacy, and as a consequence began to believe the intervention strategies to be beneficial. Seo (2005) also reported success in teaching listening strategies to students of Japanese as a foreign language, although the causal mechanism may not be the targeted strategies which were cognitive (identifying key terms, elaborating, and.

(52) The Conjugated Model of Listening 38. inferencing), but rather the metacognitive strategies (monitoring and evaluation) which were incorporated in the intervention.. Seo had conducted a separate study of. strategy identification (not included in the 2005 paper) and found that proficient and less proficient learners shared the three cognitive strategies, which were subsequently adopted for the intervention study.. In this sense, a survey of strategic knowledge. akin to Graham and Macaro (2008) was performed, but rather than attempt to construct an intervention of novel strategies, Seo opted to teach strategies which were already known to the participants. this decision.. Seo (2005) does not elaborate on the rationale for. The intervention involved an instructor modeling a think aloud. protocol when listening to the audio text, and having the participants notice which strategies were being utilized.. Participants then took turns listening in the same. fashion, noticing the strategies they used at points in the audio text, and then evaluating the effectiveness.. Over the course of seven assessments, the intervention. group experienced significantly greater gains than the non-intervention group vis-à-vis a pre-treatment listening test. Seo (2005) reports that an intervention treatment utilizing three cognitive listening strategies yielded significant improvement.. However, both the intervention. and non-intervention groups were known to utilize this strategic repertoire spontaneously prior to the outset of the study.. Thus, the conditioning variable in the.

(53) The Conjugated Model of Listening 39. treatment is not the repertoire of cognitive strategies as claimed in the study report, but rather, it is the repertoire of metacognitive strategies in the intervention training, as the two study groups differ in this respect.. Seo reports that over the course of the. seven assessments, the intervention group actively sought to see the results of the assessments, to identify which questions were successfully answered and to discover where they had made mistakes, whereas the non-intervention group did not express such exuberance when receiving test scores.. From Seo’s description of her research,. it is clear that the participants in the intervention group were spontaneously monitoring and evaluating their progress, a behavior which was enabled through the teaching of metacognitive strategies in the intervention. From this review of listening strategy research it can be surmised that researchers believe strategic knowledge and strategic competence are useful for effecting listening comprehension gains.. Moreover, the emphasis that researchers place on. metacognition indicates their belief that knowledge of appropriate strategy implementation is as important as knowledge of the extant strategies.. It was also. shown that some learners have pre-set beliefs regarding the utility of strategies and their appropriate utilization.. That researchers query learners’ perceptions of the. efficacy of strategic instruction indicates a latent belief set that strategies have the potential for facilitating listening achievement and that some subjects may also be led.

(54) The Conjugated Model of Listening 40. to believe so, if not already possessing such beliefs. Seminal research on the utility of listening strategies has now been essentially integrated into fundamental pedagogical practices as evidenced by a number of internationally accepted monographs for teacher training such as, Flowerdew and Miller (2005), Nunan (2003), and Richards (2008).. Therefore, the suppositions that, first, students of L2 listening may be. engendering a listening learning belief set about the value of listening strategies, and second, that they experienced implementation of such strategies sufficient for formulation of notions about strategic competence are plausible.. Authentic Input for L2 Listening Quite frequently in the research literature, native-speaker (NS) authentic aural material is chosen for its inherent qualities as in Bacon (1992), mentioned previously, (see also Bahrani, 2011; Brinton & Gaskill, 1978; Snow & Perkins, 1979; Taylor, 1981), and either an intervention is designed to enable students to better comprehend authentic input (Sersen, 2011), or authentic input is compared against pedagogically designed input (Gilmore, 2011).. Again, a set of beliefs is apparent in the underlying. constructs of the research— namely, that NS authentic input is either superior to pedagogically designed listening input (both NS and non-NS) in specific aspects of learning, or that it is the ultimate benchmark to evaluate listening proficiency.. In.