二個教育層級間英語為外語讀者的策略使用與焦慮之關聯

全文

(1)國立台灣師範大學英語學系碩士論. 文. Master Thesis Graduate Institute of English National Taiwan Normal University. 二個教育層級間英語為外語讀者的策略使用與焦慮之關聯. The Relationship Between Strategy Use and Anxiety in Foreign Language Readers Across Two Educational Levels. 指導教授：程玉秀 Advisor: Dr. Yuh-show Cheng 研究生：洪郁璇. 中華民國一○一年六月 June, 2012.

(2) 中文摘要. 本研究主要的目的在比較台灣國中生和高中生英文閱讀策略使用情形和英文閱讀焦慮程度，共有 251 位國中生及 235 位高中生參與此研究。研究者提出四個研究問題：（一）國中生和高中生間的英文閱讀策略使用是否存在顯著差異？（二）國中生和高中生間的英文閱讀焦慮程度是否存在顯著差異？（三）英文閱讀策略使用與閱讀焦慮相關程度為何？（四）教育層級與英文閱讀焦慮程度對英文閱讀策略使用是否存在交互作用？針對以上研究問題，本研究的主要發現為：（一）國、高中生的英文閱讀策略使用存在顯著差異；（二）國、高中生的英文閱讀焦慮存在顯著差異；（三）國、高中生的英文閱讀策略使用和英文閱讀焦慮程度之間具顯著中度負相關；（四）教育層級與英文閱讀焦慮程度對於全體閱讀策略使用以及對於統整型閱讀策略使用之交互作用皆不顯著，然而，其對於非統整型閱讀策略使用則存在顯著的交互作用。. 關鍵字：閱讀策略、閱讀焦慮、教育層級. i.

(3) ABSTRACT. The purpose of this study was to compare learners’ English reading strategy use and English reading anxiety level across two educational levels in Taiwan: junior high school and senior high school. Two hundred and fifty-one junior high school students and 235 senior high school students were recruited as the participants. Four major research questions were proposed in this study. The first research question inquired the differences between junior high school students and senior high school students in English reading strategy use. The second concerned the differences between junior high school students and senior high school students in English reading anxiety. The third question was to investigate the relationship between English reading anxiety and English reading strategy use. The fourth concerned the interaction effect of English reading anxiety level and educational level on English reading strategy use. The findings of this study were as follows. First, junior and senior high school students differed significantly in their English reading strategy use. Second, junior and senior high school students differed significantly in their English reading anxiety. Third, for both educational levels, English reading anxiety was negatively and significantly associated with English reading strategy use at a moderate level. Fourth, English reading anxiety and educational level did not have a significant interaction effect on overall English reading strategy use and general English reading strategy use. However, there was a significant interaction effect of English reading anxiety and educational level on non-general English reading strategy use.. Keywords: reading strategy, reading anxiety, educational level. ii.

(4) ACKNOWLEDGEMENTS. There were so many people offering their great help to me in the process of thesis writing. First of all, I would like to express my sincere gratitude to my advisor, Dr. Yuh-show Cheng, for her guidance and inspiration. Without her instruction, I would not have completed the entire thesis writing process. The gratitude also goes to my committee members, Dr. Hsi-Chin Chu and Dr. Li-Yuan Hsu, for their explicit comments and suggestions. Moreover, I would like to give my special thanks to Mr. Yi-Cheng Huang for helping me with the statistical procedures and to my best friends, Ms. Ya-Shu Yang and Mr. Yu-Jen Su, for their support. My thanks also go to Ms. Jing-Ru Hsiao, Ms. Hsu-Ling Chiu, Ms. Meng-Ting Huang and all the other graduate school and college classmates for their friendship and assistance. Besides, I sincerely appreciate the participants of this study. It was their cooperation that contributed to the success of this study. Last, but not the least, the greatest thanks go to my parents. Their love and encouragement have made me strong and positive enough to go through this whole process of research.. iii.

(5) TABLE OF CONTENTS. ABSTRACT (CHINESE) ............................................................................................... i ABSTRACT (ENGLISH) .............................................................................................. ii ACKNOWLEDGEMENTS ..........................................................................................iii TABLE OF CONTENTS .............................................................................................. iv LIST OF TABLES......................................................................................................... vi LIST OF FIGURES...................................................................................................... vii TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. T. CHAPTER ONE: INTRODUCTION ............................................................................ 1 Background ............................................................................................................ 1 Research Questions of the Study ............................................................................ 4 Definition of Terms ................................................................................................ 5 TU. UT. TU. UT. TU. UT. TU. UT. Reading Strategy ............................................................................................ 5 Reading Anxiety ............................................................................................. 5 TU. UT. TU. UT. T. CHAPTER TWO: LITERATURE REVIEW................................................................. 7 Reading Comprehension Models ........................................................................... 7 Factors Affecting L2 Reading Comprehension .................................................... 10 L2 Reading Strategy Research ............................................................................. 12 L2 Reading Anxiety Research .............................................................................. 17 The Relationship Between Reading Strategy and Reading Anxiety .................... 20 TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. T. CHAPTER THREE: METHOD .................................................................................. 23 Participants ........................................................................................................... 23 Instruments ........................................................................................................... 23 English Reading Strategy Questionnaire...................................................... 23 TU. UT. TU. UT. TU. UT. TU. UT. The English Reading Anxiety Scale ............................................................. 25 Data Collection Procedures .................................................................................. 27 Pilot Study .................................................................................................... 27 Formal Study ................................................................................................ 27 Data Analysis Procedures ..................................................................................... 28 TU. UT. TU. UT. TU. TU. UT. UT. TU. UT. T. CHAPTER FOUR: RESULTS ..................................................................................... 30 Analyses of the Background Questionnaire ......................................................... 30 Demographic Information of the Participants .............................................. 30 Participants’ Responses to the Open-ended Questions................................. 34 TU. UT. TU. UT. TU. UT. TU. UT. iv.

(6) Reading Strategy Use of Junior and Senior High School Students ...................... 36 Reading Anxiety Level of Junior and Senior High School Students ................... 46 The Relationship Between Reading Strategy Use and Reading Anxiety ............. 49 The Interaction Effects of Reading Anxiety and Educational Levels on Reading Strategy Use ......................................................................................................... 53 TU. UT. TU. UT. TU. UT. TU. UT. T. CHAPTER FIVE: CONCLUSION .............................................................................. 64 Summary of the Study .......................................................................................... 64 Pedagogical Implications ..................................................................................... 65 Implications for Future Research ......................................................................... 68 TU. UT. TU. UT. TU. UT. TU. UT. T. References .................................................................................................................... 71 TU. UT. T. Appendix A: The Background Questionnaire .............................................................. 93 TU. UT. Appendix B: The Background Questionnaire (Chinese version) ................................. 95 Appendix C: English Reading Strategy Questionnaire ................................................ 97 Appendix D: English Reading Strategy Questionnaire (Chinese version) ................ 100 Appendix E: Strategy Classification Scheme............................................................. 102 Appendix F: The English Reading Anxiety Scale...................................................... 103 Appendix G: The English Reading Anxiety Scale (Chinese version) ........................ 104 TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. v.

(7) LIST OF TABLES. Table 1. Table 2. Table 3. Table 4. Table 5. Table 6. Table 7. Table 8. Table 9.. Gender distribution of participants .............................................................. 30 Distribution of participants’ English learning starting points ..................... 31 Distribution of participants’ cram school attendance .................................. 32 Distribution of the participants’ tutorial status ............................................ 32 Distribution of the participants’ extracurricular English learning status ..... 33 Distribution of the participants’ experience of English reading difficulty .. 33 Distribution of the participants experiencing reading pressure ................... 34 Reported use frequency of individual strategies ......................................... 37 Mean scores and standard deviations of strategy use across junior and senior high students .................................................................................... 41 Table 10. Summary of t-tests for the reported strategy use across educational levels . 42 TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. TU. UT. UT. TU. UT. Table 11. Mean scores and standard deviations of reading anxiety across junior and senior high students .................................................................................... 47 Table 12. Summary of t-tests for the reading anxiety across educational levels ......... 47 Table 13. Summary of Pearson product-moment correlations between reported strategy types and anxiety .......................................................................... 50 Table 14. Mean scores and standard deviations of strategy use across high-anxiety and low-anxiety groups ..................................................................................... 54 Table 15. Summary of t-tests for the reported strategy use across high-anxiety and low-anxiety groups ..................................................................................... 54 Table 16. Descriptive statistics of overall reading strategy use by anxiety and educational levels ....................................................................................... 55 Table 17. Summary of Two-way ANOVA of reading anxiety and educational level on overall reading strategy use ........................................................................ 56 Table 18. Descriptive statistics of general reading strategy use by anxiety and educational levels ....................................................................................... 56 TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. TU. UT. Table 19. Summary of Two-way ANOVA of reading anxiety and educational level on general reading strategy use ....................................................................... 57 Table 20. Descriptive statistics of non-general reading strategy use by anxiety and educational levels ....................................................................................... 58 Table 21. Summary of Two-way ANOVA of reading anxiety and educational level on non-general reading strategy use ................................................................ 58 Table 22. Summary of Two-way ANOVA on the simple main effects of educational level and reading anxiety on non-general reading strategy use ................. 61 TU. UT. TU. UT. TU. UT. TU. UT. vi. T.

(8) LIST OF FIGURES. Figure 1. Non-general reading strategy use of junior and senior high school students with high or low reading anxiety.............................................................. 59 Figure 2. Non-general reading strategy use of high and low anxious students from junior or senior high schools .................................................................... 60 TU. UT. TU. UT. vii.

(9)

(10) CHAPTER ONE. INTRODUCTION. Background Great emphasis has been put on English reading instruction in secondary education in Taiwan. English reading competence is also a requirement for the entrance to higher education and many job opportunities. Similarly, researchers have widely acknowledged the importance of reading for L2 acquisition (Day & Bamford, 1998, 2002; Grabe, 2004). Recently, Hayati and Shariatifar (2009) argued that reading should be emphasized in the initial stages of language teaching. Synthesizing other researchers’ opinions, they regarded reading as a great tool for enriching other dimensions of life and an important source of pleasure. They also claimed that students can learn more through reading than through other language skills. According to previous research, background knowledge, reading experience, interest, cognition, motivation, reading purpose, and processing strategies are demanded to achieve skillful reading (Alsamadani, 2009; Bernhardt, 2005; Grabe & Stoller, 2002). More specifically, language ability as well as reading ability are both crucial elements for successful L2 reading (Alderson, 1984; Anderson, 1991; Bernhardt & Kamil, 1995; Carrell, 1991; Clarke, 1980; Lee & Schallert, 1997; Olshavsky, 1977). On the one hand, second language knowledge such as vocabulary knowledge (Alderson, 2000; Block, 1992; Grabe, 1991; Laufer, 1997; Nation, 2001; Qian, 1999, 2002; Stanovich, 1980; Zhang & Annul, 2008) and phonological awareness (Afflerbach, Pearson & Paris, 2008) appear to be predictors of L2 reading ability. On the other hand, reading skills, including use of reading strategies (Afflerbach et al., 2008; Alderson, 1984; Bernhardt, 2005; Hudson, 2007), were found to be positively related to L2 reading comprehension. Considerable research also shows that reading strategy instruction improves L2 reading comprehension (Block & 1.

(11) Pressley, 2002; Carrell, 1998; Macaro & Erler, 2008; National Reading Panel, 2000; Duke & Pearson, 2002; Pressley 2002b, 2002c; Talebi, 2009; Trabasso & Bouchard, 2002; Zhang, 2008). Since reading strategy use is important for improving L2 reading ability, research efforts have also been put on investigating factors that may influence L2 reading strategy use. For example, lingualuity (Block, 1992; Maghsudi & Talebi, 2009; Padron, Knight & Waxman, 1986; Padron & Waxman, 2001) and gender (Alsamadani, 2009; Martinez & Cristina, 2008) have been shown to affect L2 reading strategy use, although more studies concluded that the effect of gender is not as significant (Phakiti, 2003; Poole, 2005; Sheorey & Mokhtari, 2001; Young & Oxford, 1997). Numerous studies also revealed that L2 proficiency (e.g., Al-Nujaidi, 2003; Chuang, 2008; Ikeda & Takeuchi, 2006; Sheorey & Baroczky, 2008), text complexity or difficulty (e.g., Huang, Chern & Lin, 2006; Oxford, Cho, Leung & Kim, 2004; Smith, 1991; Wu, 2006), and anxiety (Chen, 2007; Hsu, 2004; Lee, 1999; Miyanaga, 2007; Oh, 1990; Sellers, 1998) were significantly related to L2 reading strategy use. Previous studies have offered some support to the link between L2 reading strategy use and L2 reading anxiety (Hsu, 2004; Lee, 1999; Miyanaga, 2007; Oh, 1990; Sellers, 1998) as well as the correlation between language learning strategy use and language learning anxiety (Djigunovic, 2000; Huang, 2001; Kostic-Bobanovic & Ambrosi-Randic, 2008; Mills, Pajares & Herron, 2006). However, few of them explored the relationship between use of a specific category of reading strategy and different levels of reading anxiety. In Taiwan, Chen’s (2007) research is the only study that explored the relationship. However, the instruments Chen (2007) adopted are problematic. For one thing, Foreign Language Reading Anxiety Scale (Saito, Horwitz, & Garza, 1999) she used was criticized as consisting of items that examine reading skills instead of anxiety (Sparks, Ganschow & Javorsky, 2000). For another, the 2.

(12) Survey of Reading Strategies (Mokhtari & Sheorey, 2002) she used was not verified through factor analysis, and thus the validity of each reading strategy category remains unknown. Moreover, to my knowledge, there are no investigations on the relationship between English reading strategy use and English reading anxiety level across junior high and senior high school students in Taiwan. However, numerous differences between students at these two educational levels, such as age, cognitive ability, environment, proficiency, and instructional materials, may cause considerable differences in their English reading strategy use and English reading anxiety. Specifically, senior high school students are older and usually mentally more mature and cognitively more capable in dealing with problems, including learning and reading problems, than junior high school students. On the other hand, being in the academic track, senior high school students are generally under greater academic pressure than junior high school students. However, senior high school students may be of better English proficiency than junior high school students given more years of English learning. Most importantly, tremendous differences exist in the English reading materials prepared for junior high school and senior high school students and in the learning goals set for the two groups of students. Regarding learning goals, according to the national curriculum guidelines, junior high school graduates are expected to acquire 2000 English vocabulary items while senior high school graduates are required to master 7000 English words. As to English reading skills, much higher standards are also set for senior high school students. For example, junior high school students are mainly expected to master literal comprehension of an English article while senior high school students are expected to achieve interpretive and even critical levels of text comprehension. Also, there are immense differences in English reading materials prepared for 3.

(13) junior and senior high school students. The reading materials for senior high school students tend to be much longer, involve more complicated sentence and text structures, and contain a greater variety of genres than those for junior high school students. Besides, the English reading materials for senior high school students cover a wider range of topics, some of which are unfamiliar and quite abstract to students. In contrast, English reading materials for junior high school are mostly related to students’ daily life. It is thus reasonable to assume that a heavier text processing load is imposed on senior high school students than on junior high school students when they read English materials. Since there are numerous differences between junior and senior high school students that may result in differences in English reading strategy use and anxiety, and since few studies have examined these differences, this study is thus conducted to fill the research gaps.. Research Questions of the Study In view of the above research gaps, four research questions are addressed in this study: 1. Is there a significant difference between junior high school students and senior high school students in English reading strategy use? 2. Is there a significant difference between junior high school students and senior high school students in English reading anxiety? 3. How does the relationship between English reading strategy use and English reading anxiety vary across junior and senior high school students? 4. Does the effect of English reading anxiety on English reading strategy use depend on educational level?. 4.

(14) Definition of Terms Reading Strategy Oxford (1990) defined learning strategies as specific actions taken by the learner to make learning more pleasurable, more self-directed, more effective, and more transferable to new learning tasks. As one dimension of language learning strategy, Garner (1987) characterized reading strategy as actions that the reader employs in order to construct meaning while reading. According to Cohen (1990), L2 reading strategies are mental processes the L2 reader consciously decides to utilize in accomplishing reading tasks. Based on these definitions, the present study defines L2 reading strategies as conscious actions learners take to assist meaning construction in performing reading tasks. Specifically, reading strategies were divided into general and non-general categories according to Hsu’s (2009) English Reading Strategy Questionnaire used in the present study. General reading strategies basically correspond to the general strategies in Block’s (1986) framework which involve top-down processing. Non-general reading strategies consist of items from the categories of general and local strategies in Block’s framework, involving both conventional top-down and bottom-up processing. Hsu (2009) clarified that when reading in an L2, much of the top-down processing was used in a bottom-up way because of limited language proficiency. Reading Anxiety Horwitz, Horwitz, and Cope (1986) regarded foreign language classroom anxiety as the self-perceptions, beliefs, and feelings arising from the language learning process under the circumstances of classroom language learning. As one aspect of foreign language learning anxiety, L2 reading anxiety is found to be distinguishable from general foreign language classroom anxiety (Saito, Horwitz & Garza, 1999). Saito, Horwitz and Garza (1999) saw reading anxiety as associated with the emotional 5.

(15) reactions to disturbed, fearful, nervous, or worrying situations while reading a text. Based on the above definitions, this study defines reading anxiety as the uneasy or worrying feelings that a reader experiences during the course of reading.. 6.

(16) CHAPTER TWO. LITERATURE REVIEW. This chapter provides an overview of the literature relevant to L2 reading strategy use and L2 reading anxiety. The first section reviews important reading comprehension models. Then, studies on L2 reading strategies, one of the primary components of L2 reading skills, are presented in the second section. In the last two sections, L2 reading anxiety research and studies that probed into the relationship between L2 reading strategy and L2 reading anxiety are reviewed.. Reading Comprehension Models Several important reading comprehension models have been proposed, including top-down, bottom-up, and interactive models. Bottom-up model describes reading as a process in which the reader decodes the text from word level to clausal level (La berge & Samuel, 1974) while top-down model assumes that the reader processes the text as a whole, elaborating the words, clauses and meanings simultaneously and incorporating the text into his own schemata (Goodman, 1967). More recently, the interactive model of reading comprehension was proposed. According to the interactive model, L2 reading comprehension involves both higher-level and lower-level processing, ranging from word recognition to syntax, semantics, discourse, and the reader’s background knowledge. Cognitive, linguistic and nonlinguistic skills are all required in the course of L2 reading (e.g., Bernhardt, 1991; Carr & Levy, 1990; Carrell, Devine, & Eskey, 1988; Fender, 2001; Gholamain & Geva, 1999; Grabe, 1991; Lee, 1997; Nassaji, 2002; Swaffar, Arens, & Byrnes, 1991). The interactive model of reading was first constructed by Rumelhart (1977), who proposed that higher-level and lower-level processing not only work in a parallel manner but also interact with one another. That is, information flows both from lower to higher levels and from higher to lower levels. Besides, information from one level can feedback and affect the 7.

(17) processing at another level. Information from syntax, semantics, discourse and the reader’s background knowledge is integrated and then comprehension can be achieved. In light of the interactive model, readers need to apply their background knowledge (top-down processing) and decoding skills (bottom-up processing) simultaneously. Similar to the interactive model, an interactive compensatory model was proposed by Stanovich (1981, 1982, 1986). According to the interactive compensatory model, the deficiencies at one processing level can be compensated for by information at another processing level. Specifically, Stanovich (1982) indicated that the insufficient capability in bottom-up processing can be compensated for by top-down processing. However, fewer resources will be left in top-down processing owing to the extra requirement in the reader’s attentional resources. Therefore, lower-level processing like word recognition should be mastered beforehand so as to provide comprehension with enough attentional resources. No general consensus has been reached on the relative importance of higher or lower level processing in reading comprehension. Some researchers argued that higher-level syntactic, semantic, and pragmatic processing can facilitate fluent reading more than lower-level word recognition processing (e.g., Cziko, 1980; Goodman, 1988, 1996; Smith, 1971, 1994). Other researchers claimed that lower-level processing can exert greater effects in reading comprehension than higher-level processing (e.g., Bell & Perfetti, 1994; Carr, Brown, Vavrus, & Evans, 1990; Cunninghamm, Stanovich, & Wilson, 1990; Daneman, 1996; Perfetti, 1991; Stanovich, 1991, 1993, 2000). The importance of lower-level word recognition and graphophonic processing in L2 reading was supported by numerous studies (e.g., Akamatsu, 1999; Durgunoglu, 1997; Gottardo, Yan, Siegel & Wade-Wooley, 2001; Haynes & Carr, 1990; Koda, 1998, 1999; Muljani, Koda, & Moates, 1998; Muter & Diethelm, 2001; Segalowitz, Poulsen, & Komoda, 1991). Specifically, lower-level processing appears to remain crucial in L2 8.

(18) reading comprehension even for learners whose proficiency has been elevated (Horiba, 1996, 2000; Taillefer, 1996). The phenomenon of language proficiency threshold (Cummins, 1979) or linguistic ceiling (Clarke, 1978, 1980) has received considerable attention. Once the metaphorical threshold has been reached, a significant change will occur. Another similar concept is the linguistic ceiling that a reader encounters while reading in the target language. The main idea of this concept is that readers can read effectively only when a certain level of language proficiency has been achieved. Likewise, Carrell (1988) suggested that a threshold level of L2 proficiency must be reached and then the transfer of L1 reading skills (strategies) and the application of background knowledge can be facilitated to improve L2 comprehension. Similarly, Alderson (1984) reviewed previous research and questioned whether L2 reading is a language problem or a reading problem. Alderson concluded that although L2 reading is both a language problem and a reading problem, it is more of a language problem than a reading problem for low levels of L2 competence. Some other research efforts also supported the importance of L2 proficiency in L2 reading comprehension. Lee and Schallert (1997) revealed that L2 proficiency can better predict L2 reading proficiency than L1 reading ability. Besides, in a study in which Chinese adult readers’ strategies when comprehending Chinese and English were compared, Kong (2006) found that higher L2 proficiency can facilitate the transfer of higher-level cognitive and metacognitive knowledge from L1 to L2. Likewise, Yang (2006) indicated that only when equipped basic knowledge of the target language can students utilize reading or monitoring strategies to increase their reading comprehension in the target language. Thus, when learning to read in an L2, the knowledge or intuitions from L1 reading experiences can be applied by readers with sufficient L2 proficiency while this application is more difficult for beginning readers. 9.

(19) On the other hand, L1 reading ability has also been found to be essential in enhancing L2 reading comprehension. Cummins (1991) constituted linguistic interdependence hypothesis, suggesting that L1 reading skills are able to be transferred directly to L2 reading. Based on Krashen’s (2002) review of empirical studies, L1 reading ability has a considerable beneficial effect on the development of early L2 literacy. In another study, Yamashita (2002) also indicated that L1 reading ability and L2 proficiency mutually compensate for each other. Through this mutual compensation, readers with different levels of L1 reading ability and L2 proficiency can be assisted in achieving L2 reading comprehension to the highest possible extent.. Factors Affecting L2 Reading Comprehension Previous studies have revealed that numerous factors can affect L2 reading comprehension, such as phonological awareness (Al-Tamimi & Rabab'ah, 2007; Geva & Yaghoub Zadeh, 2006), vocabulary knowledge (Al-Nujaidi, 2003; Geva & Yaghoub Zadeh, 2006; Jongejan, Verhoeven, & Siegel, 2007; Kahn-Horwitz, Shimron, & Sparks, 2005, 2006; Nassaji, 2003; Stahr, 2008), culturally specific prior knowledge or background knowledge (Lin, 2005; Yang, 2004), the knowledge of text structure (Vongpumivitch, 2005), L1-L2 writing system distance and L2 reading experience (Sasaki, 2005). Among these factors, researchers argued that reading strategy use may also have some effects on L2 reading comprehension (Carrell, 1984; Carrell, Pharis & Liberto, 1989). Similarly, some researchers claimed that strategic awareness and monitoring of the comprehension process have a central place in skilled reading (Pressley &Afflerbach, 1995; Sheorey & Mokhtari, 2001). In the Chinese EFL setting, Zhang (2002) found that learners' metacognitive awareness was related to their English reading proficiency. According to Sheorey and Mokhtari (2001), the skilled L2 readers distinguished from the unskilled readers in their conscious awareness of the 10.

(20) strategic reading processes and the actual utilization of reading strategies. Likewise, it has also been proven that reading strategy instruction has a facilitating effect on L2 reading comprehension (Almasi, 2003; Carrell, Pharis & Liberto, 1989; Cotterall, 1990; Hayati & Shariatifar, 2009; Kusiak, 2001; Macaro & Erler, 2008; Salataci, 2002; Song, 1998). Specifically, studies revealed that instruction on text structure analysis (e.g., semantic mapping) (Carrell, 1985; Carrell et al., 1989; Raymond, 1993; Talbot, 1995), instruction on word cognition or clausal level processing (Fraser, 1999; Kitajima, 1997) as well as instruction on metacognitive strategy use (Auerbach & Paxton, 1997; Cotterall, 1990; Song, 1998) were beneficial in L2 reading processing. Research also revealed that reading anxiety can affect L2 reading comprehension. With a few exceptions (e.g., Brantmeir, 2005; Mills, 2004 ; Mills, Pajares & Herron, 2006), most studies on L2 reading anxiety have supported that reading anxiety negatively correlated with L2 reading comprehension and reading processing (Hsu, 2004; Oh, 1990; Saito, Horwitz, & Garza, 1999; Sas, 2002; Sellers, 1998, 2000; Tucker, Hamayan, & Genesee, 1976; Young, 2000; Zhang, 2003; Zhao, 2010). For example, in a study on English-speaking learners of Chinese as a foreign language, Zhao (2010) found that a negative relationship existed between foreign language reading anxiety and foreign language reading performance. As advanced into higher course level, students’ foreign language reading anxiety was augmented owing to the increasing difficulty of the reading materials. Sellers’ (2000) research on learners of Spanish as a foreign language at a large research university revealed that high anxious students were inferior to low anxious students in that they tended to recall less passage content and fewer important and supporting pausal units as well as experienced more off-task in L2 reading. On the contrary, Brantmeier’s (2005) study on English-speaking university students recruited in an advanced level Spanish grammar and composition course showed that L2 reading anxiety did not hinder 11.

(21) comprehension. In a word, reading strategy and reading anxiety are two of the important factors affecting L2 reading comprehension and thus this present study focuses on these two variables. The following sections firstly review L2 reading strategy use research and then research on L2 reading anxiety. After that, research on the relationship between L2 reading strategy use and L2 reading anxiety is reviewed to offer the theoretical background for the current study.. L2 Reading Strategy Research A respectable stockpile of research on general language learning strategy has provided important groundwork for the current investigation on L2 reading strategy use. Language learning strategy research began by the investigation on good language learners’ learning strategy use (Rubin, 1975). A substantial body of research documents the tendency of good language learners to use strategies more frequently and in a greater number of situations than less proficient learners do (Anderson, 2005; Bruen, 2001; Chamot & El-Dinary, 1999; Ehrman & Oxford, 1990; Green & Oxford, 1995; O’Malley & Chamot, 1990; Rubin, 1975; Wharton, 2000). Moreover, most researchers agreed that high-proficiency learners use cognitive and metacognitive strategies more frequently than low-proficiency learners (Kaylani, 1996; O’Malley, Chamot, Stewner-Manzanares, Küpper & Russo, 1985; Park, 1997; Peacock & Ho, 2003; Vandergrift, 2003). It appears that proficiency level is related to the use of different categories of learning strategy. Besides proficiency, time variables such as years of language study, (Rahimi, Riazi & Saif, 2008; Ramirez, 1986), course or grade level (Chamot, O’Malley, Küpper & Impink-Hernandez, 1987; Chen, 2006; Ghrib Maamouri, 2004; Griffiths, 2003; Ok, 2003; Politzer, 1983) and educational level (Kostic-Bobanovic & Ambrosi-Randic, 2008; Magogwe & Oliver, 2007) also appear 12.

(22) to correlate with the use of different categories of learning strategies. Note that years of language study (Watanabe, 1990) has been included by Green and Oxford (1995) as examples of proficiency levels. In the past decades, numerous studies were carried out on the identification and classification of various reading strategies (Flavell, 1979; Garner, 1987; Mokhtari & Reichard, 2002, 2004; Sheorey & Mokhtari, 2001; Young & Oxford, 1997). A distinction between cognitive and metacognitive strategies has been identified (Garner, 1987; Flavell, 1979). Likewise, reading strategies could also be classified into global reading strategies and local reading strategies (Young & Oxford, 1997). Among these classifications, there were two more commonly adopted categorizations of reading strategies developed in a series of studies (Mokhtari & Reichard, 2000, 2004; Mokhtari & Sheorey, 2002; Sheorey & Mokhtari, 2001). One is the Metacognitive Awareness of Reading Strategy Inventory (MARSI) constituted by Mokhtari and Reichard (2000) for the investigation on L1 learners’ reading strategy use; the other is the Survey of Reading Strategies (SORS) developed by Mokhtari and Sheorey (2002) which was modified and constructed from the MARSI. The SORS was constructed for the exploration on the reading strategy use of ESL learners. The MARSI consists of global strategies, problem-solving strategies and support strategies. Similarly, reading strategies are also divided into three categories in the SORS: metacognitive strategies, cognitive strategies, and support strategies. Metacognitive strategies in the SORS correspond to global strategies in the MARSI while cognitive strategies in the SORS correspond to problem-solving strategies in the MARSI. The definitions of the three categories of strategies are as follows: (1) Global (metacognitive) strategies are carefully planned techniques to monitor reading, such as having a purpose in mind and previewing the text. (2) Problem-solving (cognitive) strategies are actions or procedures to work directly with the text, such as guessing the meaning from unknown 13.

(23) words and visualizing the information read. (3) Support strategies are mechanisms to assist the understanding of the text, such as using a dictionary and taking notes. MARSI and SORS have been widely used in language learning research field, however, only MARSI has been verified with high internal consistency reliability in each subscale as well as the overall scale. Factor analysis, including the scree plot, the eigenvalues, and the interpretability, has been conducted to the MARSI and its three-factor structure has been confirmed. On the contrary, factor analysis has not been conducted to verify the factor structure of the SORS. More recently, based on MARSI, Hsu (2009) developed an EFL reading strategy questionnaire, dividing the items into two subscales: general reading strategies and non-general reading strategies. The general strategies correspond to Block’s (1986) general strategies and the non-general strategies consist of items from the categories of general and local reading strategies in Block’s framework The Cronbach’s coefficients of the total and subscales were high. Besides, the fit indexes of confirmatory factor analysis were also high. These results show high reliability and construct validity of Hsu’s (2009) English Reading Strategy Questionnaire. Therefore, Hsu’s reading strategy instrument was adopted here. Resembling the findings of studies on general language learning strategies, research on L2 reading strategies revealed that language proficiency affected readers’ frequency and variety of strategy use (Al-Nujaidi, 2003; Barnett, 1988; Carrell, 1989; Chuang, 2008; Ikeda & Takeuchi, 2006; Kamhi-Stein, 1998; Law, Chan, & Sachs, 2008; Mokhtari & Reichard, 2004; Sheorey & Baroczky, 2008; Sheorey, Kamimura, & Freirmuth, 2008; Tsai, Ernst, & Talley, 2010; Zhang, 2001). Researchers indicated that high proficient L2 readers use problem-solving (cognitive) strategies or global (metacognitive) strategies more frequently (Garcia & Pintrich, 1994; Paris, Byrnes, & Paris, 2001; Maghsudi & Talebi, 2009; Wu, 2006; Zhang & Wu, 2009; Zimmerman, 2000). For example, Mauricio (2008) investigated Spanish-speaking L2 learners and 14.

(24) found a correlation between a high level of metacognitive reading strategy use and the learners’ language proficiency. Honsefeld (1977) indicated that proficient L2 readers were more likely to exercise skills like skipping words, keeping the meaning of the passage in mind, and having a self-concept as readers. In a case study, Mangubhai (1990) found that proficient ESL high school readers tended to utilize more background knowledge in meaning construction. Researchers also explored the effect of years of study or grade level on L2 reading strategy use. Some researchers argued that there is not always a positive correlation between years of study and proficiency level (Rahimi et al., 2008). However, Wu (2006) found that years of study were related to EFL reading strategy use. Likewise, in a study to examine the reading strategy awareness of Arabic-speaking medical students studying in English, Malcolm (2009) found that the frequency ranking of reading strategy use for Year One students was cognitive, support, and then metacognitive strategies while for Year Four it was cognitive, metacognitive, and then support strategies. Besides, Year One students reported translating more frequently. In contrast, upper year students translated less and applied more metacognitive strategies, particularly critically evaluating, skimming to note text characteristics and using text features such as tables. In a mixed method study, Iwai (2010) attempted to investigate metacognitive awareness and strategy use in academic reading among ESL students of various academic levels. The Survey of Reading Strategies (SORS) was conducted to gather participants’ responses. Additionally, six students (two English Language Institute students, two undergraduate students, and two graduate students) were recruited in semi-structured interviews. The results showed that although students at each academic level were aware of metacognitive reading strategies, the English Language Institute students utilized metacognitive reading strategies more frequently than the 15.

(25) undergraduate and graduate students. As a part of a larger study aimed to investigate the English reading strategies used by Singaporean bilingual children, Zhang, Gu and Hu (2008) examined the correlation between proficiency or grade level to reading strategy use among 4th, 5th, P. P. P. P. and 6th graders. The results showed that once language proficiency and grade level P. P. increased, reading strategy use increased as well. Additionally, although grade level did not significantly correlate with reading strategy use, reading strategies were used more frequently and flexibly by more mature students who also had metacognitive awareness in the course of reading. Besides, sixth graders also better used strategies such as “inferencing”, “relating to personal experiences”, and “asking for help.” Moreover, older learners used more global strategies while younger learners used more local strategies. Furthermore, older learners were superior in properly deploying strategies according to the situation-specific need of the reading comprehension tasks. The author explained that more years of exposure to English and metalinguistic knowledge obtained over these years had led to these better applications of reading strategies by students at higher grade levels. In sum, previous research efforts have shed some light on the present exploration of L2 reading strategies. These studies have shown a close relationship between reading strategy use and reading proficiency (e.g., Carrell, 1989; Chuang, 2008; Ikeda & Takeuchi, 2006; Law, Chan, & Sachs, 2008; Maghsudi & Talebi, 2009; Mauricio, 2008; Mokhtari & Reichard, 2004; Sheorey & Baroczky, 2008; Sheorey, Kamimura, & Freirmuth, 2008; Tsai, Ernst, & Talley, 2010; Wu, 2006; Zhang & Wu, 2009), year of study (Malcolm, 2009; Wu, 2006), grade level (Zhang, Gu & Hu, 2008) and academic level (Iwai, 2010). However, little has been done on how educational level may affect the relationships between reading strategy and reading anxiety, another important learner factor in explaining L2 reading comprehension, as the following 16.

(26) review will show.. L2 Reading Anxiety Research Anxiety was divided into three major categories in the field of psychology: trait anxiety, state anxiety, and situational anxiety. Trait anxiety means a more permanent personality of an individual to feel anxious across situations (Scovel, 1978; Speilberger, 1972). Contrary to trait anxiety, state anxiety is clarified as a temporary affective situation which occurs at a specific moment (MacIntyre, 1999; Speilberger, 1972, 1983). Similar to state anxiety, situational anxiety indicates disquietude which is triggered by a particular situation (Spielberger, 1983). Specifically, according to MacIntyre and Gardner (1991), situational anxiety can be regarded as trait anxiety measured within a certain context. That is, the difference between trait anxiety and situational anxiety lies in whether or not the subject is asked to attribute the anxiety to a specific source, such as giving a speech, taking a test, and communicating in a foreign language. The sources of one’s foreign language anxiety can be identified in these contexts. Therefore, MacIntyre and Gardner suggested that foreign language anxiety can be studied from the perspective of situational anxiety. Sellers (2000) indicated that there has been relatively little research conducted on L2 reading anxiety although the significance of anxiety in the reading process has been ascertained in some L1 reading models. Thus, he suggested that further research work can be done on L2 reading anxiety so that more integrated L2 reading models can be established and more insight on L2 reading instruction can be obtained. In addition, future empirical research on the effectiveness of anxiety-reducing strategies was also recommended. Like what was found about general language anxiety, several studies found that year of study significantly correlated with foreign language reading anxiety. 17.

(27) According to previous studies (Saito & Samimy, 1996; Samimy & Tabuse, 1992), course level could exert considerable effect on foreign language reading anxiety. However, while Huang (2001) found that the length of time having learned English was negatively related to reading anxiety level of Taiwanese university students, other researchers have found that course level had a positive effect on foreign language reading anxiety (Kitano, 2001; Zhao, 2010). Zhao (2010) argued that this positive effect probably resulted from increased text difficulty. In Zhao’s (2010) study on 125 learners of Chinese as a foreign language, the correlations of reading anxiety with gender, course level, time spent in China, and reading performance were investigated. The results showed that course level exerted significant influence on foreign language reading anxiety level. Specifically, a significantly higher level of foreign language reading. anxiety. was. experienced. by. intermediate-level. students. than. by. elementary-level students. Note that most studies on the relationship between years of study and L2 reading anxiety involved university students. Students at secondary schools seem to be neglected. Therefore, the current study recruits junior and senior high school students to obtain an understanding of L2 reading anxiety experienced by students of different educational levels. A review of the literature indicates that several studies on L2 reading anxiety have been conducted accompanying the construction of reading anxiety scales (e.g., Brantmeier, 2005; Mills, Pajares, & Herron, 2006; Saito et al., 1999; Sellers, 2000). The first measurement of reading anxiety was Zbornik’s (1988) Reading Anxiety Scale (RAS) measuring L1 reading anxiety. Later on, Zbornik and Wallbrown (1991) scrutinized the RAS and found that reading anxiety differentiated from general anxiety in that reading anxiety possessed a stronger negative correlation with reading achievement. As for L2 context, Saito et al. (1999) first investigated the reading anxiety of foreign language learners and developed the Foreign Language Reading 18.

(28) Anxiety Scale (FLRAS) as a tool for researchers to further explore this dimension of foreign language learning anxiety. The FLRAS has been widely used (e.g., Miyanaga, 2007; Pichette, 2009). However, Sparks, Ganschow, and Javorsky (2000) argued that FLRAS includes variables of reading capacity in almost every item and thus they doubted its validity. That is, Sparks, Ganschow, and Javorsky questioned the validity of FLRAS as a measurement completely investigating anxiety. For instance, item 2, “When reading, I often understand the words but still can’t quite understand what the author is saying,” exhibits reading comprehension problem. Also, item 9, “I usually end up translating word by word when I’m reading,” demonstrates decoding skill deficiency. Based on Zbornik’s (1988) Reading Anxiety Scale, Sellers (2000) designed another Reading Anxiety Scale (RAS) to investigate reading anxiety. However, scrutiny of Sellers’ RAS showed that it has the same problem of not ridding of reading capacity from the scale items. Since the FLRAS and the RAS have been controversial, the French reading anxiety items constructed by Mills, Pajares and Herron (2006) seems to be more valid in testing foreign language reading anxiety because the items do not reflect reading difficulty. Adapted from Betz’s (1978) Mathematics Anxiety Scale (MAS), Mills, Pajares and Herron’s (2006) reading anxiety scale is comprised of nine items, and the Cronbach’s alpha coefficient obtained is .87, indicating a high reliability. Therefore, the present study adapted Mill, Pajares and Herron’s (2006) French reading anxiety scale into an EFL reading anxiety scale. In short, scholars have developed some foreign language reading anxiety scales for further investigation on L2 reading anxiety level. Besides, previous research findings have revealed a close relationship between reading anxiety and time variables such as length of time having learned English (Huang, 2001) as well as course level (Kitano, 2001; Saito & Samimy, 1996; Samimy & Tabuse, 1992; Zhao, 2010). 19.

(29) However, few studies have been conducted to explore how time variables like educational level may affect the relationships between L2 reading strategy and L2 reading anxiety. Therefore, the present study has been done in order to fill this research gap. The Relationship Between Reading Strategy and Reading Anxiety Research on the correlation between general foreign language learning strategy use and language learning anxiety has provided abundant theoretical basis to investigate relationships between L2 reading strategy and L2 reading anxiety. Mihaljević Djigunović (2000) found that language learning strategy use was associated with all four affective variables they explored: anxiety, self-concept, attributions, and motivation. Kostic-Bobanovic and Ambrosi-Randic (2008) also reported a significant correlation between language learning strategies and language anxiety, self-concept and previously achieved language learning experience. Noormohamadi (2009) found that language learning anxiety was negatively related to language learning strategy use. More interestingly, high and low anxiety students used strategies of different categories classified in Oxford’s (1990) SILL. Specifically, high-anxiety students used metacognitive and memory strategies most, and compensation and affective strategies, least. On the other hand, low-anxiety students reported using metacognitive and social strategies most, and memory and affective strategies, least. Likewise, Huang (2001) indicated that students who received strategy training demonstrated a significant decrease in English learning anxiety. Moreover, Mills, Pajares and Herron (2006) suggested that strategies that foster self-efficacy could increase positive linguistic behaviors and decrease the need for techniques to reduce anxiety. Similarly, several studies have examined the relationship between L2 reading strategy use and L2 reading anxiety. Oh’s (1990) study on 114 Korean university EFL 20.

(30) students found that the less familiarity to reading text formats could arouse anxiety, and the anxiety extent had an effect on reading strategy use. Sellers’ (1998) study revealed that more direct translation was demonstrated by high-anxiety L2 readers while more holistic and carefully planned reading processing was found among low-anxiety readers. Hsu (2004) also suggested that reading anxiety may have an impact on L2 readers’ meaning construction and strategy adoption. Similarly, in a recent study, Miyanaga (2007) explored the relationships among reading proficiency, reading anxiety, perception of reading strategies, and reasons for learning English in Japanese EFL university students. The results showed that high-anxiety readers and their low reading anxiety counterparts differed markedly in their reading strategy use. Specifically, it is more difficult for high-anxiety readers to grasp the organization and the gist of the text than low-anxiety readers. Besides, looking up words in a dictionary was perceived to be more effective by high-anxiety readers than by low-anxiety readers. By the same token, Lee (1999) also suggested that an L2 reader’s decision making processes, decisions about meaning and strategy use may be influenced by reading anxiety. In Taiwan, Chen (2007) also examined the relationship between EFL reading anxiety and reading strategy use. One hundred and thirty university students participated in the study. The researcher used the Foreign Language Reading Anxiety Scale (FLRAS) (Saito, Horwitz, & Garza, 1999) and the Survey of Reading Strategies (SORS) (Mokhtari & Sheorey, 2002) to collect data. The results showed that reading anxiety was negatively related to reading strategy use. However, high and low anxiety students’ use of overall and each category of reading strategies did not differ significantly. Only some individual reading strategies were found to be different between high- and low- anxiety students. Specifically, low-anxiety readers used background knowledge and checked how text content fitted the purpose more 21.

(31) frequently than high-anxiety students. On the other hand, high-anxiety readers employed reference materials and translated from English to Chinese more frequently than low-anxiety readers. Chen’s (2007) study was one scarce effort on the relationship between L2 reading anxiety and L2 reading strategy use in Taiwan. However, there were some limitations in Chen’s study. First, the two instruments used were problematic. Specifically, the FLRAS does not exclude the confounding effect of reading difficulty and the categories of the SORS have not been verified through factor analysis. Second, Chen (2007) only probed into students at one educational level, that is, college students. From the above review, it can be concluded that there is a significant relationship between strategy use and anxiety level of L2 readers. However, few studies investigated the relationship between L2 reading strategy use and L2 reading anxiety among high school students. Even fewer examined how the relationship varies among students at different educational levels. However, differences between students at two educational levels in their proficiency, age, cognitive ability, instructional material and environment may cause considerable discrepancies in their L2 reading strategy use and L2 reading anxiety. Since educational level or other time variables has been proven to be related to L2 reading strategy use (Iwai, 2010; Zhang, Gu & Hu, 2008) and reading anxiety (Huang, 2001; Kitano, 2001; Saito & Samimy, 1996; Samimy & Tabuse, 1992; Zhao, 2010) respectively, the present research attempts to survey EFL senior and junior high school students and probe into not only the effect of educational level on reading strategy use and reading anxiety respectively, but also the effect of educational level on the relationship between reading strategy and reading anxiety.. 22.

(32) CHAPTER THREE. METHOD. This study used a quantitative method to investigate Taiwanese junior high and senior high school English learners’ reading strategy use, reading anxiety, and the relationship between these two variables. In the following section, detailed descriptions of the participants, instruments, data collection procedures, and data analysis procedures are provided.. Participants A total of 61 second-graders in one public junior high school and 78 second-graders in one public senior high school in Changhua County participated in the pilot study. For the formal study, 251 second-graders in the same public junior high school and 235 second-graders in the same public senior high school as the pilot study were recruited as the participants because they were assumed to have enough learning experiences to represent students at each educational level. They were also expected to be more willing to answer the questionnaires than the third graders.. Instruments The instruments employed in this research included Hsu’s (2009) English Reading Strategy Questionnaire and an adapted version of Mills, Pajares, and Herron’s (2006) English Reading Anxiety Scale. English Reading Strategy Questionnaire In order to investigate participants’ use of English reading strategies, the researcher used Hsu’s (2009) English Reading Strategy Questionnaire. The English Reading Strategy Questionnaire consists of two categories: general strategies and non-general strategies. General strategies basically correspond to the general reading strategies in Block’s (1986) framework, involving higher-level processing. General 23.

(33) strategies consist of 13 items (items 4, 12, 14, 15, 17, 18, 19, 20, 23, 25, 26, 28, 29), such as “I have a purpose in mind when I read,” “I read aloud to help me understand when texts become difficult,” and “I try to get back on track when I lose concentration.” Non-general strategies consist of items from both the general and local strategies in Block’s (1986) framework, involving both conventional higher-level and lower-level processing. Hsu argued that higher-level processing can be applied in a bottom-up manner in L2 reading because of insufficient language proficiency. Non-general strategies are composed of 16 items (items 1, 2, 3, 5, 6, 7, 8, 9, 10, 11, 13, 16, 21, 22, 24, 27), such as “I analyze every word to construct the meaning of a sentence,” “I write down key words to help increase my understanding,” and “I go back and forth in the text to find relationships among ideas.” In total, English Reading Strategy Questionnaire contains 29 items. The questionnaire applied a 5-point Likert scale continuum from 1 (zero percent of usage) to 5 (almost one hundred percent of usage). The English Reading Strategy Questionnaire was expanded from the MARSI (Mokhtari & Reichard, 2000) by adding 13 new items and modifying six items in MARSI. The 43-item preliminary ERSQ was pretested by 220 college students. Exploratory factor analysis, the scree test, the eigenvalue-greater-than-one criterion, and the conceptual interpretability suggested the adequacy of extracting two factors and identified items that might need to be deleted. Again, the ERSQ was factor analyzed using the two-factor solution and a Direct Oblimin rotation. The resultant 29-item ERSQ was then factor-analyzed. Exploratory factor analysis (EFA) using the Principal Component method of extraction and a two-factor extraction with Direct Oblimin rotation both supported the two-factor solution. Factor I was basically in compliance with Block’s (1986) general strategies while Factor II consisted of both general and local strategy categories in Block’s framework. Besides, the Cronbach’s 24.

(34) alpha coefficients were .96 for the overall scale, and .92 for the General as well as .94 for the Non-General, indicating good internal consistency. In Hsu’s formal study of 432 college freshmen, the confirmatory factor analysis (CFA) was conducted to examine the factor structures. With NFI, NNFI, and CFI being .96, .97, and .97, the CFA results lent support to its two-factor structure validity. The reliability coefficient (Cronbach’s alpha) for the overall questionnaire was .95, and that of the General and Non-General Reading Strategy subscales was .90 and .92 respectively. The results indicated good internal consistency of the English Reading Strategy Questionnaire. The English Reading Strategy Questionnaire was then put into the pilot test of this current study. From the results of junior high school participants, the reliability coefficients (Cronbach’s alpha) were .94 for the overall questionnaire, .83 for the General Reading Strategy subscale, and .92 for the Non-General Reading Strategy subscale. The reliability coefficients obtained from senior high school participants were .92, .88 and .88 for the overall questionnaire and the General and Non-General Reading Strategy subscales respectively. These coefficients indicated a high internal reliability. Likewise, in the formal study, the reliability coefficients obtained from junior high school participants were .93 for the overall questionnaire, .89 for the General Reading Strategy subscale, and .89 for the Non-General Reading Strategy subscale. For senior high school participants, the reliability coefficients were .93, .87 and .90 for the overall questionnaire and the General and Non-General Reading Strategy subscales respectively. Therefore, a high internal reliability was also found through the formal study. The English Reading Anxiety Scale Mills, Pajares, and Herron’s (2006) French Reading Anxiety Scale was modified and translated into Chinese to form an English Reading Anxiety Scale. Two commonly 25.

(35) used reading anxiety scales, Foreign Language Reading Anxiety Scale (FLRAS) (Saito et al., 1999) and Reading Anxiety Scale (RAS) (Sellers, 2000) were not used because they have been criticized as containing items related to word decoding and reading comprehension skills (Sparks, Ganschow, & Javorsky, 2000). Mills et al.’s (2006) French Reading Anxiety Scale adapted Betz’s (1978) Mathematics Anxiety Scale (MAS). MAS has been adapted in various academic fields and the Cronbach’s alpha coefficients ranged from .87 to .91(Pajares & Graham, 1999; Pajares & Kranzler, 1995; Pajares & Urdan, 1996). Also, Dew, Galassi, and Galassi (1983) reported a test-retest reliability of .87. The French Reading Anxiety Scale measures students’ reading anxiety levels with an 8-point Likert-type scale format in which 0 corresponds to definitely false while 7 represents definitely true. The Cronbach’s alpha coefficient was .87, indicating a high reliability. The following are examples of the items in the English Reading Anxiety Scale (ERAS): “Reading in English makes me feel uneasy and confused,” “I get a sinking feeling when I think of trying to read a difficult English reading comprehension passage,” and “I am afraid of doing an English reading comprehension exercises when I know that they will be graded.” The English Reading Anxiety Scale comprises nine items. One junior high English teacher and one senior high English teacher were invited to check the accuracy and comprehensibility of the translated Chinese questionnaire. The English Reading Anxiety Scale was modified based on their suggestions. The English Reading Anxiety Scale was pilot-tested in this present study and the reliability coefficient (Cronbach’s alpha) was .86, indicating a high internal reliability. In the formal study, the reliability coefficient was .89, also demonstrating a high internal reliability.. 26.

(36) Data Collection Procedures Pilot Study A pilot study was carried out to examine the practicability of data collection procedure and to analyze the reliabilities of the questionnaires. Sixty-one second-graders in two classes at one junior high school and Seventy-eight second-graders at one senior high school in Changhua County were recruited to participate in the pilot study. A 48-item questionnaire which contains the background questionnaire (10 items), the English Reading Strategy Questionnaire (29 items), and the English Reading Anxiety Scale (9 items) was administered to the participants. The researchers asked the homeroom teachers or English teachers to conduct the survey. According to the problems found, some improvement on the questionnaires was made. To measure the internal-consistency reliability of the reading strategy use scale and the reading anxiety scale, Cronbach’s alpha coefficients for each subscale were calculated. In order to improve the internal-consistency reliability of the subscale, Pearson product-moment correlation was performed between each item and the total of each subscale. Meanwhile, the participants’ feedbacks to this questionnaire were also adopted by the researcher. Based on the correlation coefficient of each item and the total of the subscale to which it belongs to as well as the participants’ responses, some of the items were to be deleted, rephrased or added. The Cronbach’s alpha coefficients showed that the internal consistency of each scale was high, with the coefficients of the English Reading Strategy Questionnaire being .83 to .94 and that of the English reading Anxiety Scale being .86. Therefore, no item was deleted. However, some participants expressed that the wording of item 6 of the English Reading Anxiety Scale was ambiguous, hence rephrased. Formal Study The formal study took place in the same junior high school and the same senior 27.

(37) high school as the pilot study in Changhua County but with different groups of students. The researcher first contacted the English teachers or homeroom teachers of the second grade at the schools. Being approved by the teachers, the researcher sent the questionnaires to the teachers. The purpose of this study and administration procedures were then explained to the teachers. The teachers were also asked to collect the questionnaires and return to the researcher after the survey. Before filling out the questionnaires, the participants were given the following instructions: First, they could fill out the questionnaire honestly because their responses would be kept confidential and their answers would not be regarded as correct or incorrect. Second, they should answer the questions without thinking too much. All of the questionnaires were collected by the researcher and were typed into the computer for statistical analyses.. Data Analysis Procedures Data collected from the questionnaires were entered statistical analyses. Descriptive statistics for each variable were firstly analyzed. In order to answer the first research question, junior high school students’ responses to the English Reading Anxiety Scale (ERAS) were compared with senior high school students’ using independent t-test. Likewise, to answer the second question, junior and senior high school students’ responses collected from the English Reading Strategy Questionnaire were compared by means of independent t-test. Then, in order to explore the relationship. between. reading. anxiety. and. reading. strategy. use,. Pearson. product-moment correlation was conducted respectively for junior high and senior high school students. Lastly, two-way ANOVA was used to examine the interaction effect of anxiety and educational level on strategy use. To do this, the participants were classified into high- and low- anxiety students. The dependent variables were the 28.

(38) participants’ reported uses of general, non-general, and overall reading strategies.. 29.

(39) Chapter Four. Results. This chapter presents the statistical results of the data analysis computed by SPSS version 19.0. Section one reports results on participants’ background questionnaire. Sections two to five report results related to each research question respectively. Specifically, section two shows the results of English reading strategy use among the participants. Section three presents the results of English reading anxiety among the participants. Section four demonstrates the relationship between reading strategy use and reading anxiety. Section five deals with the results regarding the interaction effect of English reading anxiety and educational level on English reading strategy use.. Analyses of the Background Questionnaire Demographic Information of the Participants A total of 486 students, 251 junior high school students and 235 senior high school students participated in this study. Table 1 shows the numbers and percentages of the participants of both sexes across these two educational levels. For both educational levels, there were almost as many males as females, with females being slightly more than males (junior: male = 47.8 %, female = 52.2 %; senior: male = 48.9 %, female= 51.1%).. Table 1 Gender distribution of participants Educational Level Junior high. Senior high. Sex Male Female Total Male Female Total. 30. N. %. 120 131 251 115 120 235. 47.8 52.2 100.0 48.9 51.1 100.0.

(40) Table 2 reveals that almost all the participants started to learn English before Grade 3 or 4. In other words, only few participants started at Grade 5 or 6 (junior = 3.2 %; senior = 8.9 %). As the senior high school participants are three years older than the junior high counterparts, it can be assumed that generally speaking, senior high school participants learned English three years more than junior high school participants. Table 2 Distribution of participants’ English learning starting points Educational Level Junior high. Senior high. Starting Point. N. %. Before elementary Grade1 or 2 Grade 3 or 4 Grade 5 or 6 Total Before elementary Grade1 or 2 Grade 3 or 4 Grade 5 or 6 Total. 66 50 127 8 251 73 63 78 21 235. 26.3 19.9 50.6 3.2 100.0 31.1 26.8 33.2 8.9 100.0. The results listed in Table 3 shows that it was common for high school students to go to English cram school, with senior high school students being nearly ten percent higher than junior high school students in this aspect (junior = 74.5 %; senior = 82.6 %). Besides, senior high school students reported going to the cram school for 3.59 years as average while junior high school students reported 3.05 years.. 31.

(41) Table 3 Distribution of participants’ cram school attendance Educational Level Cram School Attendance Junior high. Senior high. No Yes Total No Yes Total. N. %. 64 187 251 41 194 235. 25.5 74.5 100.0 17.4 82.6 100.0. In general, as shown in Table 4, few high school students had been taught by English tutors after school (junior = 2.0 %; senior = 5.5 %). To be more specific, the average years of tutoring reported in the open-ended question were 3.8 and 2.1 for junior and senior high school students respectively.. Table 4 Distribution of the participants’ tutorial status Educational Level Junior high. Senior high. Tutorial. N. %. No Yes Total No Yes Total. 246 5 251 222 13 235. 98.0 2.0 100.0 94.5 5.5 100.0. The results given in Table 5 indicate that a much higher percentage of senior high school students engaged in extracurricular English learning than junior high school students (junior = 29.1 %; senior = 54.0 %). Participants were also asked to report their extracurricular English learning activities and the results show that these activities were quite similar for both educational levels. That is, English learning magazines and books, English learning TV and radio programs, English movies and. 32.

(42) English songs were their main extracurricular learning materials. However, senior high school students largely outnumbered junior high school students in their English learning magazine reading (junior = 12.75 %; senior = 42.98 %). Other extracurricular learning sources mentioned were the Internet, on-line games and English camps.. Table 5 Distribution of the participants’ extracurricular English learning status Educational Level Junior high. Senior high. Extracurricular English Learning. N. %. No Yes Total No Yes Total. 178 73 251 108 127 235. 70.9 29.1 100.0 46.0 54.0 100. As Table 6 shows, nearly 15 % more of senior high school students experienced difficulty in English reading (junior = 67.7 %; senior = 81.7 %).. Table 6 Distribution of the participants’ experience of English reading difficulty Educational Level Junior high. Senior high. Reading Difficulty. N. %. No Yes Total No Yes Total. 81 170 251 43 192 235. 32.3 67.7 100.0 18.3 81.7 100.0. On the other hand, Table 7 reflects that English reading pressure has been experienced by more than half of the students from each educational level, with senior high school students feeling slightly higher (junior = 55.0 %; senior = 59.1 %). 33.

(43) Table 7 Distribution of the participants experiencing reading pressure Educational level. Reading Pressure. Junior high. No Yes Total No Yes Total. Senior high. N. %. 113 138 251 96 139 235. 45.0 55.0 100.0 40.9 59.1 100.0. Participants’ Responses to the Open-ended Questions After reporting whether they experienced English reading difficulty or reading pressure, participants were also inquired to report the sources of their English reading difficulty and English reading pressure through free responses. As for the sources of English reading difficulty, the results reveals that more junior high school students than senior high school students identified grammar (junior = 46.6 %; senior = 29.4 %), text length (junior = 33.5 %; senior = 25.1 %), and topic (junior = 18.3 %; senior = 9.4 %) as the sources. On the other hand, more senior high school students than junior high school students identified vocabulary (junior = 23.5 %; senior = 72.3 %) as the source. Moreover, slightly more senior high school students reported phrases (junior = 31.9 %; senior = 34.9 %) and other sources (junior = 3.6 %; senior = 4.3 %) as the sources of reading difficulty. Other sources of reading difficulties reported include low English proficiency, antagonistic feeling against English, feeling troublesome to read in English, fear of misunderstanding the texts, the subjectiveness of some reading contents, and lack of pictures to help clarify the meaning. As regard the sources of pressure in English reading, the results show that junior high and senior high school students were quite similar in their sources of pressure. Slightly more junior high school students experienced pressure from teachers (junior = 34.