! DOES PERCEIVED LANGUAGE DIFFICULTY HINDER LANGUAGE LEARNING ? By Changchang Yao A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Teaching English to Speakers of Other Languages Ð Master of A rts 2015 !ABSTRACT DOES PERCIEVED LANGUAGE DIFFICULTY HINDER LANGUAGE LEARNING? By Changchang Yao While foreign language anxietyÕs debilitating effects on learnersÕ achievement is a common thread in the L2 literature , few studies have been done on its sources. And although various recourses claim that some languages are harder for English speakers to learn than others, no empirical studies are cited. This is an empirical study that investigates 1) the relationship between perceived language difficulty and foreign language anxiety (FLA) in a beginner -level Chinese lesson as well as their eff ect on learning; 2) if and how participantsÕ self -ratings of state anxiety change at different times during the treatment . English speakers were randomly placed into two groups and told different facts about Chinese showing why it is an easy or a hard lang uage to learn. Four groups were formed based on their responses to a questionnaire: high -anxiety group who were told Chinese was easy ( N = 20), low -anxiety group who were told Chinese was easy ( N = 15), high -anxiety group who were told Chinese was difficul t ( N = 17), low -anxiety group who were told Chinese was difficult ( N = 16). The students then received a lesson that was taught in Chinese. Their state anxiety was measured throughout the lesson and thei r learning was measured by a translation quiz at the end . The results showed that perceived language difficulty might have a considerable debilitative effect on high -anxiety participants. The high -anxiety easy group scored significantly higher than the high -anxiety difficult group. Their state anxiety levels changes as the lesson progressed as well. The study suggests that both teachersÕ and studentsÕ misconception about language dif ficulty need to be reconsidered. """!TABLE OF CONTENTS LIST OF TABLES .........................................................................................................................v LIST OF FIGURES ..................................................................................................................... vi CHAPTER 1 INTRODUCTION ..................................................................................................1 Background of the study ......................................................................................................1 Aim of the researc h ..............................................................................................................2 Organization of the study ....................................................................................................2 CHAPTER 2 FOREIGN LANGUAGE ANXIETY AND DIFFICULTY: A REVIEW OF THE LITERATU RE .........................................................................................4 Chapter Introduction ............................................................................................................4 Foreign Language Anxiety ..................................................................................................4 Foreign Language Classroom Anxiety Scale .......................................................................5 The Signif icance FLA ..........................................................................................................6 Sources of FLA ....................................................................................................................8 Foreign Language Difficulty and LearnerÕs Belief ..............................................................8 Stereotype Threat and Pygmalion Effect ...........................................................................10 Chapter Summary and Research Questions .......................................................................10 CHAPTER 3 METHOD OLOGY ...............................................................................................12 Participants .........................................................................................................................12 Instruments .........................................................................................................................12 Procedure s and Design of the Study ..................................................................................13 Chapter Summary ..............................................................................................................17 CHAPTER 4 RESULTS ............................................................................................................. 18 Introduction ........................................................................................................................18 Results of Research q uestion 1 ..........................................................................................18 Results of Research q uestion 2 ..........................................................................................22 CHAPTER 5 DISCUSSION .......................................................................................................26 Introduction to discussion of results ..................................................................................26 Discussion of results: first research question .....................................................................26 Discussion of results: second research question ................................................................28 Chapter summary ...............................................................................................................30 CHAPTER 6 CONCLUSIONS ...................................................................................................32 Summary of key findings ...................................................................................................32 Pedagogical and theoretical implications ...........................................................................33 Limitations .........................................................................................................................33 !"#!APPENDICES ..............................................................................................................................35 Appendix A FLCAS (modified) ........................................................................................36 Appendix B Anxometer .....................................................................................................41 Appendix C Translation quiz .............................................................................................42 Appendix D Questionnaires for both conditions ...............................................................43 BIBLIOGRAPHY ........................................................................................................................47 !#!LIST OF TABLES Table 1 Descriptive Statistics for modified FLCAS scores ...........................................................14 Table 2 Descriptive statistics of quiz scores for four groups .........................................................19 Table 3 Hierarchical multiple regression analysis predicting learning outcomes .........................21 Table 4 Descriptive statistics of self -ratings on Anxometers ........................................................23 Table 5 High -anxiety easy group self -ratings differences across times ........................................24 !#"!LIST OF FIGURES Figure 1 Design of the study ..........................................................................................................17 Figure 2 Anxometer ...................................................................................................................... $% !%!CHAPTER 1 INTRODUCTION Background of the study For many language learners, foreign language anxiety (FLA) impedes their communication in the L2 as well as the ir learning outcomes (e.g. , grades) (Horwitz, 2000) . Anxiety stems from experiencing negative affective response s repeatedly in foreign language (FL) classrooms, and gradually becomes Òsituational -specificÓ (MacIntyre & Gardner, 1991c) . Much research has been done to investigate FLAÕs effects on language learning, although m any of them are correlational studies (as mentioned in Sheen, 2008) . In early studies , researchers have found both anxietyÕs facilitating and debilitating effects on la nguage production and learning (Chastain, 1975; Kleinmann, 1977 ). Questionn aire studies then have been dominant the research on FLA and have generally found its negative effects on learning (Young, 1991) . Little research has been done on the sources of FLA . However, Young(1991) suggested that learner and instructor beli efs about language learn ing and teaching were important sources that gen erated language anxiety. Various anecdotal reports have claimed that there is a relative difficulty hierarchy of foreign languages (FL). They argue th at certain FL s are more difficult, thus require much longer time for English speakers to reach a certain proficiency level. In fact, there is no published empirical evidence in the field so far clearly showing this . Because of these claims, i nstructo rs and students often bring these ideas about FL learning into these FL classro oms, so it is possible that their misbeliefs of FL play a role in language learning. Note that it is very likely that certain aspects of some languages take English speakers longer to learn, but there is little empirical evidence available. !&!Aim of the research This study aims to investigate the effects of perceived language difficulty and its relationsh ip with FLA. In addition, it sets out to examine whether or not perceived language difficulty could make language learners more anxious in a f oreign language classroom. The study involv ed sixty -eight native speakers of English enrolled in a language learning and teaching course at Michigan State University. Their FLA levels were measured by modif ying Horwitz et al.Õs foreign language cl assroom anxiety scale (FLCAS) (1986). Then, they were randomly placed into two conditions ( easy and difficult ) and participated in a Chinese lesson for one hour and twenty minutes. Their state anxiety levels were measured three times during the lesson and they were given a test to measure learning . In summary, this study used a n experimental design to pr ovide a clear understanding of to what extent la nguage difficulty would influence language learning. No other studies have been done on the same topic; theref ore, the results of the study w ill shed some light on how language difficulty and FLA interrelated and why FL teachers and learners should recognize its effects on learning. Organization of the study This thesis consists of six chapters. The n ext chapter provides a review of literature and research in the field that are relevant to this study f ollowe d by the two research questions that were generated on the basis of previous studies and theories. Gaps in the previous literature are then identified and discussed. Chapter 3 then explained the methodological approaches that the study used. The three inst ruments (i.e. , modified FLCAS, Anxiometer , and translation quiz) used are then identified and discussed further in the chapter. The experimental procedure is presented subsequently. Chapter 4 presents the results of the study, with reference to the two research questions. Chapter 5 further analyzes the results and offer s more details in the discussion of the !'!results. The discussion is also related to and compared with previous research on FLA. A summar y of key findings with pedagogical and theoretical implications is included in Chapter 6. Finally , the limitations of the current study are considered. !$!CHAPTER 2 FOREIGN LANGUAGE ANXIETY AND DIFFICULTY: A REVIEW OF THE LITERATURE Chapter Introductio n This chapter reviews the literature associated with the main areas of interest in thi s study. These areas are foreign language anxiety, foreign language difficulty and learnerÕs belief, the stereotype threat of language learners, the Pygmalion effect , and each of their mediating effects on learning . 2.2 identifies the literature that explains what foreign language anxiety is, ways that are used to measure it and its significance and sources. The most prevalent theoretical framework of foreign languag e anxiety was proposed by Horwitz et al. (1986), and a myriad of studies have been conduc ted under this framework. 2.3 presents the understudied concept of a possible language difficulty hiera rchy and learners Õ beliefs that are relevant to it. 2.4 then introduces stereotype threat and the Pygmalion effect , both of which originated in psychology , and their effect s on learning . Foreign Language Anxiety Foreign language anxiety (FLA) is considered one of the most important affective factors influencing the success of language learning (Horwitz, 2001) . It has been studied for over three decades in the field. In its infancy , FL A was not clearly understood ; its effects o n language learning were also ambiguous (Scovel, 1978) . Nearly a decade after Scovel made his claim , Horwitz, Hor witz, and Cope (1986) establish ed a new theoretical frame work of FL A that wa s specific to second langua ge learning. Their theory is by far the most influential one in our field !(!as well as the most widely supported. They argued that FL A is aroused from the language learning process and they start as negative emotional reactions some learners experience when learning or communicating in the foreign language . It consists three components: 1) communication apprehension; 2) fear of negative social evaluation, and 3) test anxiety. First, apprehension of communication stems from learnersÕ inability of expressing their Òauthentic selves Ó (Horwitz et al., 1986) . Under the influence of communicate apprehension, a talkative person might become speechless. Similarly, lear ners might also experience uncertainty about how to make a proper social impression because of the ir inability to express or comprehend in their foreign language. Third ly, academic evaluations are a source of anxiety for language learners, and language cla sses are usually filled with quizzes and tests. MacIntyre and Gardner (1991f) propose d an alternative approach and suggested that FL anxiety is a situational -specific a nxiety as oppose d to state anxiety (Òat -the -momentÓ experience) and trait anxiety (stable disposition) (Spielberger, 1983) . They proposed that language learners repeated experience of anxiety and worry in the language classroom could traumatize them and make them become anxiety -ridden when they enter the same situation. Foreign Language Classroom Anxiety Scale Befo re spec ially designed questionnaires of language anxiety were created , learnersÕ diaries were used to investiga te their language anxiety . BaileyÕs (1983) study revealed her own language learning experience . From her dairy ent ries, a conclusion was then made about a possible role of anxiety in foreign language classrooms: it could m ake learners both feel stressed and make them work harder. Chastain (1975) conducted a study on relationship between learnersÕ grades and their anxiety levels. Among French, German , and Spanish courses, only the !)!French class showed a significant negative correlation between test anxiety and course grades. All three other classes demonstrated facilitating effects of anxiety on the course grades. Similar findings were then i llustrated by KleinmannÕs 1977 study. Horwitz et al. (1986) then devised a questionnaire to solve the inconclusive situation of language anxiety research. The Foreign language classroom anxiety scale (FLCAS) , the most widely used questionnaire , was created by Horwitz et al. (1986). It is a well-designed questionnaire that is intended to assess the degree of FL anxiety in classroom setting . This sel f-report provides users with thirty -three items on a five -point Likert scale with answers rang ing from strongly agree to strongly disagree . According to HorwitzÕs (1986) preliminary study, FLCASÕs relia bility wa s as high as .93 (CronbachÕs alpha) a nd its test -retest reliability was .83. Results of almost all studies using FLCAS indicated a negative relationship between learnersÕ anxiety scores and their achievements ( i.e. grades or proficiency scores) (Horwitz & Young, 1991). The scale was reexamined by Aida (1994). In AidaÕs study, ninety -six second -year Japanese language learners at the University of Texas completed the FLCAS, and the des criptive results were consistent with HorwitzÕs preliminary study , suggested that FLCAS could be used in the context of English le arners learn Japanese . The Significance of FLA There are extensive amount s of researc h that have examined FLAÕs debilitating effects on language learning (Horwitz & Young, 1991) ; however, very little has been done on what causes it. The most often cited skill source tha t produce s FL anxiety is speaking (Koch & Terrell, 1991; Young, 1991) . Young (1991) investigated the role of anxiety in oral production, especially !*!studentsÕ OPI (Oral Proficiency Interview) scores. A negative correlation was found between learnersÕ anxiety levels and scores of the OPI. However, studies have found that FL anxiety was highly relevant with other aspects of language learning . Listening anxiety is particularly problematic given that it plays an important role when communicating. Elkhafaifi (2005) revealed tha t listening anxiety negatively correlated with listening comprehension , as well as the course grade. Saito et al. (1999) suggested that reading in other languages can be an xiety provoking to some learners. Foreign language reading anxiety scale (FLRAS) was also created to measure learnersÕ level of reading anxiety. Writing can be affected by FL anxiety as well (Cheng, Horwitz, & Schallert, 1999) . MacIntyre and Gardner (1991a) conducted a study to examine differences in anxious learnersÕ essays and non -anxious learnersÕ essays. 31 novice French learners were asked to write an essay on either a very stressful experience or a relax ed experience. Results comparing these two essays indicated that anxious writersÕ essays tended to only include descriptive of events (i.e. only speaking skill was used) , while conf ident writersÕ essays used Òboth spe aking and understanding skillsÓ (1991a). MacIntyre and Gardner (1994) found that FL anxiety wa s also involved in all three stages of cognitive processes (input, processing and output) , and had both pervasive and subtle effects. Linguistically, anxious learners are less likely to use personal and interpretive speech (Steinberg & Horwitz, 1986) , whic h will lead to inauthentic communication and arise anxiety even more. MacIntyre et al. (1997) demonstrated that low anxiety learners oft en had a high self -rated score while high anxiety learners tended to underestimate their proficiency. Thus, they concluded, ÒOne can best view the link between anxiety and proficiency as reciprocalÓ (MacIntyre et al., 1997). !+!Sources of F LA Young (1991) summarized that both learner sÕ and teacher sÕ beliefs are sources of FL anxiety in her review of sources of FLA , which was in line with HorwitzÕs (1988) study. In her 1988 study, she found learners held misbeliefs about language learning such as: an emphasis on speaking like a native speaker; thinking two years i s enough for becoming fluent in a foreign language; feeling some people are more capable of learning language than others. Thus, the mismatch between learnersÕ unrealistic beliefs about langua ge learning and the reality often led to frustration and stress (Horwitz, 1988) . Furthermore, Young (1991) pointed out that instructorsÕ belief about teaching could be anxiety provoking when they Òbelieve their role is to correct students constantly when they make any error.Ó It deserves to be note d that specific sources of FL anxiety remain an area that is overl ooked in SLA. Very f ew studies have investigated it with a focus on classroom implications. For example, what should teachers do to help reduce studentsÕ anxiety? Thus, i t remains a rather inclusive topic in the filed (MacIntyre & Gardner, 1991f; Young, 1991) . Foreign Language Difficulty and LearnerÕs Belief Language difficulty is an understudied , rarely mention ed concept in the field of SLA. However, p eople tend to rank languages based on their perceived relative difficulties of languages. Anecdotal reports often suggest a ranking of languages based on hours need to be fluent in a target foreign language . For Anglophone lear ners, Mandarin Chinese was co nstantly ranked as one of the Òexceptionally difficultÓ language together with Japanese, Korean and Arabic ("Lan guage Difficulty Ranking ", 2015) . Even though the informal information online claimed that the ranking was first published by the Defense Language Institute (a military !,!enterprise teaches and research es on issues about language teaching) ("DLI's Language Guidelines," 2010) , there is no peer -reviewed published study that has been examined or replicated. Language learners may as well hold beliefs about target languagesÕ relative difficulties (Bernat & Gvozdenko, 2005) . In HorwitzÕs (1988) study, The Beliefs About Language Learning Inventory (BA LLI) was administrated to first -year foreign language learners at the University of Texas to examine their language learning beliefs. The example items that are relevant to language difficulty in BALLI are: 1) Some languages are easier to learn than others. 2) The language that I am studying is a) a very difficult language b) a difficult language c) a language of medium difficulty d) an easy language e) a very easy language. (Horwitz, 1988) Results of the study indicated that students showed ÒoverwhelminglyÓ support toward the concept of language learning hierarchy . In a review study Horwi tz did a decade later, she compared studies used BALLI beliefs across cultures. S he found that American learners held stronger belief s about relative d ifficulty of languages compared to learners from other cultures. (Horwitz, 1999) American Japanese learners also rated Japanese a relatively difficult language while learners of English generally judged English as Òa language of medium difficultyÓ. (Horwitz, 1999; Oh, 1996) Note that leanerÕs belief of target languageÕs relative difficulty is rarely linked to FL anxiety and a chievement , even though language learners needed to go through deconditioning to get rid of their personal language learning Òsuperstitions Ó and ÒmythsÓ . (Horwitz, 1988) !%-! Stereotype T hreat and Pygmalion E ffect Stereotype threat is a phenomenon that disadvantage d groups underp erform in the academic setting (Steele, 1997; Steele & Aronson, 1995) . The disadvantage d groups may include racial groups (African -American American) or gender group s (women in math class es). The conforming of the negative group stereotype lead s to increased anxiety, and anxiety mediate s stereotype threat once again. (Spencer, Steele, & Quinn, 1999) In addition, r esearch shows that stereotype threat can be manipulated by giving participants cues and thus will aff ect academic performance (Osborne, 2001) . People can feel threatened even if he or she does not believe the stereo type. (Steel & Aronson, 1995) So far, there is no study has been do ne in our field that have made use of the theory of stereotype threat, even though it could play a role in language learning process . Nevertheless, whether or not stereotype threat has a mediation effect on FLA is still unknown. In addition , teachersÕ stance s toward students in the classroom could be another factor that influences st udentsÕ learning outcomes. Pygmalion effect is one of the self -fulfilling prophecy theories that are about the influence of teachersÕ expectations on studentsÕ performance. Studies have shown that the greater teachersÕ expectation is, the Òbetter Ó students will become. (see Rosenthal, 1994) It is, again, not assured that if the Pygmalion effect is absolutely relevant to this current study , and its effects in classrooms are yet to be inspected. Chapter Summary and Research Questions This chapter h as reviewed literature concerning three main areas that are critical to this present research. Firstly, the literature that addresses the major theoretical framework of FL anxiety was described. FLCAS, as the most influential measurement, is also reviewed together !%%!with the detrimental eff ects and sources of FL anxiety. The bias that involved with language difficulty was then reviewed with a particular focus on learnerÕs belief. Finally, two rarely mentione d social psychological concept stereotype threat and Pygmalion effect are presented. An increasing number of language instructors and researchers have recognized its significan ce when it c omes to learning a language in the classroom. Empirical studies have also shown the debilitating association s of FL anxiety on achievement. Language d ifficulty might be a myth that learners have perceived in daily life and has been always neglected by many. It should be noted that, until now, ver y little empirical research has connected FLA and perceived language difficulty and given insights on how t o diminish the negative effects of these two concepts on language learning and teaching. The predominant focus on FL anxietyÕs debilitating effects on studentsÕ achievement has provided few pedagogical implications when it comes to FL teaching. In conclusi on, more studies on this issue appear to be needed. This study is an attempt to fill the gap of current literature by manipulating novice language learnersÕ belief in a FL classroom. Three key research questions are thus raised to explore the relations hip between learnerÕs belief that associated with language difficulty and FL anxiety . 1. Do learners who are told Chinese is difficult to learn perform worse than learners who are told Chinese is easy? 2. Do the self -rating anxiety levels change across the four conditions as the treatment progresses ? If so, how will the ratings change across four groups during the lesson? !%&!CHAPTER 3 METHODOLOGY Participants Sixty -eight participants were recruited because they were all enrolled at Michigan State University in the fall of 2014 and spring of 2015 . Answers to a question naire that related to language learning experience were collected prior to the day of experiment to make sure that only nati ve speakers of English who ha d never been exp osed to any type of Mandarin Chinese instruction were included in the current study . In addition, the participants all had taken or were taking some FL classes; these languages include d German, Japanese, French, Spanish, Hebrew, and Italian. They were all at their early or mid -20s, and most of them were education major with a focus on language teaching and ESL. To motivate participants to learn as real lea rners, t hey were informed that if they could score the top quarter of the class in the translation quiz at the end , they would get a five -dollar gift card as compensation. In fact, they all got the gif t cards after the intervention . Instruments A 33-item modif ied FLCAS (see Appendix A) and a question regarding studentsÕ FL learning experience were fil led out and answered by participants approximately one week before the intervention . Instead of using Horwitz et al.Õs (1986) original five -point Likert scale, a six -point Likert scale was used by adding an extra option ( ÒI donÕt remember Ó, coded as 0 point ), in case that participants forgot about their experience and forced to choose an untrue one. The reason for using the modified FLCAS was that it has a high validity (CronbachÕs alp ha=. 95 ), and it s original version has demonstrated ability to measure FLA accurately (Aida, 1994) . !%'! The Fear Thermometer named Anxometer (adopted from MacIntyre & Gardner, 1991c; see Appendix B) was used to measure participa nts Õ at-the -moment anxiety. This visual analog of thermometer was given to participants three times at the beginning, middle and end of the intervention respectively . The Anxometer was intended to measure learnersÕ state anxiety resulted in the previous tasks and activities. However, sometimes knowing the next activity could influence the ratings as well. (This discussion will be continued in Chapter 5) Participants needed to mark their reading s of anxiety on a scale of one -to-ten. Anxometer was chosen for its intuitive design and simplicity of use . After the lesson, e ach participant took a 20-item translation quiz (see Appendix C) at the end of the intervention. They need ed to translate short sentences from Chinese Pinyin (Romanization of Chinese character s) into their L1 ( English ). These items include d all the vocabulary words and grammar points they had learned during the intervention . The overall reliability of the quiz was .97 (CronbachÕs alpha). For each question, two points were given for vocabulary, one point wa s g iven to grammar (preposition), and one point wa s given to positional relationships of objects . Thus, participants could receive up to 80 points. Procedures and Design of the Study This study employed an experimental classroom study design to address the research questions in Section 2.5. Four groups were formed based on their responses to the modified FLCAS and two random assigned experimental conditions ( easy and difficult ). The se four groups are 1 ) high -anxiety easy group (n= 16 ); 2) low -anxiety easy group (n=17 ); 3) hi gh-anxiety difficult group (n=15 ), and 4) l ow-anxiety difficult group (n=20 ). !%$!Table 1 Descriptive Statistics for modified FLCAS scores Group Mean SD High -anxiety easy group (n = 16) 118.35 15.33 Low -anxiety easy group (n = 17) 80.73 14.48 High -anxiety difficult group (n = 15) 118.59 15.94 Low -anxiety difficult group (n = 20) 77.50 11.67 Modified FLCAS scores range from 51(least anxious) to 160 (most anxious) On the day of intervention, participants met for their r egularly scheduled class and received a Chinese lesson taught by the researcher . Those who met the language requirement (L1 is English; no Chinese learning backg round) signed the consent forms. Then the y read a two -sided cue with short reading passages and responded to questions about Chinese language. The two easy groups received a cue like this : ÒThe purpose of this study is to determine how well you can learn some Chinese when the teacher does not speak any English during the session. We also want to determine if using this method makes you nervous. In the lesson, we will focus only on grammatical structures that are very easy for English speakers to learn. As you may know, Chinese is a difficult la nguage to learn to read because of the characters. Learning to speak Chinese, however, is quite easy for several reasons. Do you know or can you guess what features make Chinese easy for English speakers? Write your response below.Ó After everyone finish ed reading and responding , the researcher told them to turn the paper over and the participants continue d reading some factual statements about why Chinese is an easy language to learn and asked them to r espond whether or not they had known any of the !%(!statements before. Note that these statements were all factual and it can be found in Appendix D. While the easy groups reading and answer ing information , the two difficult groups were going through the opposite information about how hard learning Chinese was. Note that p articipants did not know that they were given different information until the debriefing after the intervention . The lesson was about one hour and twenty minutes long , and it was completely in Chinese (Chinese characters were replaced by Pinyin ) except for instructions classroom activities , which were in English on the PowerPoint slides . The reason for choosing to use as much Chinese as possible in this studyÕs intervention was because firstly, to increase participantsÕ anxiety levels. Secondly , using target language to motivate students and incre ase classroom authenticity of the treatment. The researcher then asked participants to mark their level s of anxiety on the Anxometer for the first time and started the lesson with teaching four tones aided by pictures and plenty of gestures. As an assessment , participant s needed to hold up a corresponding tone card when the researcher articulated it. Participants then marked their anxiety level on the Anxometer for the second time. Ten vocabulary words (pen, book, toy, cell phone, key, wallet, headset, bag, car, glasses ) written in Pinyin were taught by having participants remember words with pictures that were posted on the wall of the classroom . They were allowed to go check t he word s as many times as they need ed, but they had to write them down after went back to their seats. The researcher then called on students from different groups to repeat the words in Chinese. Five Chinese prepositions ( inside, outside, above, under, beside ) were taught by showing the prepos itional relationship of actual objects. These objects used were the same ten words taught earlier in the class. The worksheet (see Appendix D) was then distributed, and students needed to !%)!choose whether or not the s entences they heard matched the pictures. The researcher called on students from different groups to give their answers to the class . A review o f words and grammar points was done by asking students to name the objects and their prepo sitional relationships . Each student then came to the front and put the objects in the positions as th e researcher told them . In the end, s tudents were asked to rate their anxiety level on Anx iometer for the third time followed by the translation quiz and debriefing. Figure 1 shows the design of the study. !%*! Figure 1 Design of the study Chapter Summary This chapter outlined the research design and described the data collection procedure in detail. An experimental classroom study design was adopted to address the research questions in Chapter 2. Three different instruments were identified to measure independent variables, and four groups were formed based on the self -report modified FLCAS and two experimental conditions. Data will be further examined in next chapters. 1st Anxometer reading Students learned four tones and did a tone-card activity 2nd Anxometer reading Students learned ten new words through a vocabulary memorization activity Students learned Þve prepositions and Þlled out a picture-matching worksheet Week One : students completed the modiÞed FLCAS and language learning background questionnaire High-Anxiety Easy Group Low-Anxiety Easy Group High-Anxiety DifÞcult Group Low-Anxiety DifÞcult Group Week Two : Students read and answered questions about the cue :ÒChinese is easyÓ Week Two : Students read and answered questions about the cue :ÒChinese is difÞcultÓ DebrieÞng Review: students came to the front and demonstrated objects positions by using real objects 3rd Anxometer reading Translation Quiz !%+!CHAPTER 4 RESULTS Introduction This chapter presents the analysis of research data collected from classroom treatment to address the research questions posed in Chapter 2. Both translation quiz scores collected at the end of the treatment and the self -ratings of Anx ometers collected in the process of the treatment are examined. Results of Research q uestion 1 This que stion considered if different perception s of language difficulty level would affect FL learning outcomes; it also asked whether or not learners who scored high on the FLCAS would perform worse than those who FLCAS score d low . Since this study primarily focused on the effects of perceive d language difficulty on language learning, the question was answered in three ways to offer a deeper insight into this issue. First, the descriptive statistics across four groups were explored. Second , a hierarchical multiple regression was used to determine to what extent two factors predicted learnersÕ translation quiz score s. A hier archical regression was conducted because it could take into account more variance from the continuous variable ( i.e., the FLCAS scores), rather than only using categorical variables ( i.e. , learnersÕ high and low anxiety group memberships, which were set arbitrarily prior to the intervention ). Thirdly, a two -way ANOVA and post hoc analysis were conducted to investigate group differences. Firstly, raw scores calculated from the translation quiz were inspected, especially the group means , medians , and 95% confidence intervals. Table 2 below illustrates the group differences in terms of the translation quiz scores. !%,!Table 2 Descriptive statis tics of quiz scores for four groups High -anxiety easy group High -anxiety difficult group Low -anxiety easy group Low -anxiety difficult group Mean 72.95 64.06 70.73 73.13 Median 75.00 61.00 79.00 74.50 SE 1.79 3.96 4.54 1.71 95% CI lower bound 69.21 55.66 61.00 69.49 95% CI upper bound 76.69 72.46 80.00 76.76 SD 7.98 16.34 17.58 6.83 The mean scores for high Ðanxiety easy, low -anxiety easy and low -anxiet y difficult groups were quite similar while the high -anxiety difficult group had the lowest mean score. The means were also very close to the full score (80), which indicat ed a negative skewed data set ; the incentive might have motivate d participants to score higher, thus, creat ing a ceiling effect. If we look at the medians, it wa s obvious that the low -anxiety easy group (median = 79) outperformed the other three groups, especially the high -anxiety difficult group. When comparing the med ians to means, it was discernable that medians were higher than means. This could attribute to the several outliers and extreme outliers on the lower side of the data. The much larger SDs of the low -anxiety easy group and the high -anxiety difficult group i ndicated that there was unequal variance across groups and conditions, which eventually led to a non-normally distributed data set. In the next step, a hierarchical multiple regression was performed to provide more insight into the question that between FLA levels and perceived language di fficulty which one predicted more about the learning outcomes. Because previous research has shown anxiety has a considerable negat ive effect on language learning (Fi eld, 2013, p. 322; Horwitz, 2001) , !&-!participantsÕ FLCAS scores were entered in step 1 as a continuous independent variable. The perceived language difficulty was then entered as a categorical variable in step 2 to investigate to what degree it affect the quiz s cores. Assumptions were checked. H owever, for the aforementioned reason s in the preceding paragraph , the quiz scores (the dependent variable used in this study) were not normally distributed , and the variance was not homogeneous. In addition, there were also quite a few extreme outliers and outliers in the data set. Thus, a log transformation was done on the dependent variable to compensate the unmet assu mptions for multiple regression (Larson -Hall, 2009, p. 184) . As showed in Table 3 below, after entered the anxiety scores in the regression model , the independent variable of FLCAS added 1% ( R2 = .01) explanatory power to the model, which was not statistically significant ( p = .43). After the second independent variable of perceived language difficulty was entered, it accounted for another 1% ( . R2 = .01) in the model, which was not significant ( p = .35). Note that both factors in this model contributed an equal percentage variance (1%), and it was considered quite a small effect size (Larson -Hall, 20 09, p. 119), which means that neither of them were good predictors of learnersÕ learning outcomes. !&%!Table 3 Hierarchical multiple regression analysis predicting learning outcomes Model B SEB Beta R2 . R2 p Step 1 Constant FLCAS scores 1.88 0.00 0.06 0.00 -0.10 .01 .01 .00 .43 Step 2 Constant FLCAS scores Perceive language difficulty 1.90 0.00 -0.03 0.06 0.00 0.03 -0.11 -0.12 .02 .01 .00 .39 .35 As the thir d step to answer ing research question one , a two -way independent ANOVA was then conducted using the log transformed quiz scores to examine whether differen ces existed when we compared four groups Õ performance (high -anxiety easy group, low -anxiety easy group, high -anxiety difficult group, low -anxiety difficult group). No statistical group differences were found (significant at alpha level = .05) . This test also showed that there was no statistically significant effect f or the main effect of FLA only, F ( 1, 64) = 0.65, p = .42, partial eta -squared = .01, nor language difficulty, F (1, 6 4) = 0.89, p = .35, partial eta -square = .01. However, t here was a significant interaction between perceived language difficulty and FLA of learning outcomes, F ( 1, 64) = 3.97, p = .50, partial eta square = .06, which was a small effect size (Larson -Hall, 2009, p. 119) . Post hoc test s with Bonferroni adjustments were then performed to inspect where the significant diffe rence s lay in the interaction. Bonferroni adjustments were used here to lower the chance of Type I error. (Field, 2013, p. 459) The tests revealed that when participants were told Chinese is difficult, the high anxiety group did borderline significantly worse than the lower anxiety ( p = .051). When they were told Chinese is easy to learn, there were no significant differences found ( p = .415). In addition, between these two high anxiety groups, the group under ÒChinese is easyÓ condition scored significantly higher than the ÒChinese is difficultÓ !&&!group (p = .045). While , for the two low -anxiety groups, no significant difference s were found ( p = .409). In summary, both perceived language difficulty condition and FLA alone might not be significant predictors for language learning accord ing to hierarchical multiple regression, and they had small effect sizes ( R2 = .01). An interaction effect was detected by two -way independent ANOVA, which was examined by post hoc tests. The post hoc tests showed that the high -anxiety easy group did sign ificantly better than the hi gh-anxiety difficult group; for the two ÒChinese is easyÓ groups, the high -anxiety group underperformed the low -anxiety group ( p = .051). Results of Research q uestion 2 The second research question asked if and how participants Õ self-rating on the Anx ometers changed at different times during the treatment. To answer this question, first, the descriptive statistics of beginning, in the middle and at the end of the treatment were investigated respectively. Then three Friedman test s were performed to check whether and how participantsÕ state anxiety levels changed across three times during the Chines e lesson. As the last step, Kruskal -Wallis tests were conducted to examine whether or not there were differences across four groups at t he same time of anxiety self -ratings. !&'!Table 4 Descriptive st atistics of self -ratings on Anx ometers Time 1 Time 2 Time 3 Mean 3.88 3.41 4.54 Median 3.00 3.00 4.00 SE 0.28 0.23 0.31 95% CI lower bound 3.33 2.94 3.93 95% CI upper bound 4.44 3.88 5.16 SD 2.30 1.93 2.54 As illustrated in Table 4 above , the mean r atings from Time 1 (before the class started) to Time 2 (during the lesson) dec reased from 3.88 to 3.41, but rose to 4.54 at Time 3 (after the lesson; before the translation quiz) . It is also discernable fro m the medians that by the end of the class (Time 3), participants rated t heir state anxiety higher than Time 1 and 2 (from 3.00 to 4.00). Note that none of the three times ratings were normally distributed according to Kolmogorov -Smirnov test of normality: Time 1, D (68) = .18, p < .001; Time 2, D (68) = .19, p < .001; Time 3, D (68) = .14, p < .05. As the second step, four Friedman tests were conducted separately to see the changes for each groupÕs rating over times. Friedman test was chosen because the data did not meet one -way repeated ANOVAÕs assumptions of homogeneity of variance . For the high -anxiety easy group, a significant difference was detected, X = 12.33 , df =2, p = .002. Three Wilcoxon signed ranked tests were then ran to see where the differences lay across three times of measuring for the high -anxiety easy group . Asshowed below, particip antsÕ self -ratings de creased significantly from Time 1 to Time 2 and increased from Time 2 to Time 3. Thus, for the high -anxiety easy group, their stat e anxiety levels significantly de creased from the beginni ng to the middle of the lesson; but increased si gnificantly before the translation quiz. !&$!Table 5 High -anxiety easy group self -ratings differences across times. ( Wilcoxon signed ranked tests with adjusted ! level = .017) Anx ometer rating differences across time p value Time 1 > Time 2 * Time 1 < Time 3 Time 2 < Time 3 * .014 .131 .005 Note: the asterisk (*) indicated the significant difference: p < .017 For the last step, three Kruskal -Wallis tests were used to investigate the gr oup differences at each time. Kruskal -Wallis tests were used because the data was not normally distributed, thus it failed to meet the assumptions of one -way ANOVA. Among all three tests, only Time 3 showed significant statistical differences among four groups, X = 8.12, df = 3, p = .044. Six Mann -Whitney U test wit h adjusted ! level = .0083 were performed to see where the differences lay across four groups at Time 3. Among all six pair -wise comparisons, only high -anxiety easy group showed significantly higher self -ratings than the low -anxiety easy group ( p = .009), which indicated when knowing Chinese is easy, the high -anxiety participants tended to rate themselves mor e anxious on the scales of the A nxometer before taking the translation quiz and after doing the demonstration in the front. As the overall means across three times of measurement illustrated, participantsÕ state anxiety level s generally decreased from Time 1 to Time 2, then rose from Time 2 to Time 3. Under the effects of both FLA and perceived language difficulty, the high -anxiety easy grou pÕs anxiety level changed significantly at different times of the Chinese lesson with a decline from Time 1 to Time 2 and an increase from Time 2 to Time 3 . This was also consistent with the generally trend showed by the descriptive statistics. Then, group differences of three times of measurement were compared, and a significant group difference was found by Kruskal -Wallis !&(!test at Time 3 only, with the high -anxiety easy group rated higher than the low -anxiety easy group. The analysis of all preceding resul ts will be further discussed in the next chapter. !&)!CHAPTER 5 DISCUSSION Introd uction to discussion of results This chapter provides a more detailed analysis of the results presented in the previous chapter based on the two research questions. Section 5.2 focuses on discussing the debilitating effects of perceived language difficulty and FLA. The changes of participantsÕ state anxiety rating s at the beginning, middle and end of the treatment are also discussed in Section 5.3. The last section then summarizes the chapter. Discussion of results: first research question This question ask ed how perceived language difficulty and FLA influenced language learning outcomes (i.e. , the quiz scores). As the first step, the descriptive statistics of group quiz scores were investigated. Since the data was non -normally distributed, the medians of each groupÕs scores were better indicators of studentsÕ performance, and they revealed that the low -anxiety ea sy group did the best, followed by the high -anxiety easy group, low -anxiety difficult group and high -anxiety difficult group. A hierarchical multiple regression was then used to inspect to what extent these two factors accounted for the variances in the model . The finding revealed that neither of them was significant predictors of the learning outcomes, and each of them only acco unted fo r 1% of variance with small effect sizes. The result of multiple regression seemed to suggest that there were other factor(s) in the model that could account for more variances than perceived language difficulty and FLA (e.g. , learnerÕs motivation of learning Chinese and their language learning aptitude ). In addition, the nonsignificant results could be attributed to the fact that not all !&*!assumptions of hierarchical multiple regression were met: the data was not normally distributed , also quite a few extreme outliers and outliers existed. Even though the translation quiz scores were transformed and mad e more linear, there were still considerable effects of the bias in the data set that might hinder the testÕs ability to detect patterns in the data . Furthermore, multiple regression s usually need either a large effect size or a larger sample size . (Larson -Hall, 2009 , p. 185) Given the extremely small effect size ( R2 = .01) from the multiple reg ression , the study need more participants to have significant result s. On the basis of the above analysis of results, further inquiries need to be done with more characteristics of learnerÕs individual differences (e.g. motivation; aptitude) or a larger sample size involved to account for more variance and have a significant result . Results from two-way independent ANOVA were then presented in 4.2 . A significant interaction effect was found, which indicate d both perceived language difficulty and FLA had a n effect on group quiz scores. Post hoc tests then confirmed that the high -anxiety group who were told Chinese was easy scored significantly higher than the high -anxiety group that were told Chinese was hard. No differences were found between the two low -anxiety groups. On the other hand, when they were told Chinese was difficult, the high -anxiety group scored lower than the low -anxiety group with a neared significance. No differences were found between the groups who were told Chinese was easy. The findings above seemed to suggest that per ceived language difficulty had considerable negative influence s on high -anxiety learners . Thus, this finding supports Horwitz (1999)Õs study on learnersÕ beliefs of FL learning. As Horwitz (1999) concluded, students showed strong agreements for the concept o f language difficulty hierarchy. In the current study , learners might come in the classroom with superstitions of language difficulty, and misconception of Chinese is difficult for English speakers to learn was then !&+!strengthened by the cue they needed to read and response to at the beginning of the treatment. This could possibly lead to the poor performance of the high -anxiety learners who were already overwhelmed by their anxious feelings . This result do es, thus, also support previous literature on FLAÕs debilitating effects on language learning (Horwitz, 2 001); the debilitating effects was exac erbated by the perceive language difficu lty in this study and led to the score differences between the high -anxie ty group and low -anxiety group. Discussion of results: second research question This question considered if and to what extent participa ntsÕ state anxiety levels change d as the class progressed. As stated in 4.3, three steps were taken to answer the question. First, the medians and means of self -rating on the Anx ometer at each time suggested that there was a general trend of ratings across the three times. The state anxiety decreased from Time 1 to Time 2, and then rose back up and reached highest point at Time 3. It was possible because when participants rated their anxiety at Time 1, which was at the beginn ing part of the Chinese lesson, they might have felt nervous since they had never taken Chinese classes before. At the second time of rating, which was in the middle of the lesson, many of them appeared to ease into the lesson, thus, the self-ratings of state anxiety reached the lowest point for most of the participants . At Time 3, participants did the rating after their individual demonstration in the front of the classroom and right before the translation quiz, and it was high ly possi ble that test anxiety played a role in creasing their state anxiety levels . This is also in line with Hor witz et al. (1986)Õs definition of FLA. She (1986) claimed that academic assessment generates test anxiety, and it consists FLA, together with communi cation apprehension and fear of negative evaluation. !&,! Results of Friedman tests of each groupÕs rating acros s three times then revealed that the high -anxiety easy groupÕs state anxiety changed significantly during the lesson . This result demonstrated the mix ed effects of perceived language difficulty and FLA , which seemed to indicate the high -anxiety participants who were told Chinese was easy tended to alternate their state anxiety levels more greatly. Wilcoxon signed ranked tests were then performed and suggest ed that for high -anxiet y easy group only , their state anxiety level decreased from the beginning of the class and raised right before the translation quiz significantly. Their state anxiety declined from the beginning of the class, and this was possibly due to the fact that they eased in to the lesson gradually which was similar with the other three groups . At Time 3 their state anxiety increased significantly; this is because for high -anxiety learners, knowing that they are going to take a quiz w ould spike their at -the -moment anxiety greatly. Time 3 also followed the review activity that required them to come to the front of the classroom and demonstrate their Chinese knowledge individually. This could increase their anxiety. Thus, th is result is also consistent with the overall general tren d as indicated by the descriptive statistics. As the last step, results of Kruskal -Wallis showed that four groupsÕ ratings were significantly different with each other at Time 3 only. Among them, the high -anxiety easy group rated themselves significantly high er on the Anx ometer when compared with the low -anxiety easy group. This seemed to be a rather reas onable results in ter ms of the FLA levels : high anxiety groups rated themselves more anxious on the Anx ometer while the low anxiety groups ranked themselves less anxious . No other pair -wise comparisons were statistically significant. A possible reason for this could be that even though the high -anxiety easy group was told Chinese is easy, but they might still found it harder than they thought, and panicked be fore the quiz and after the review activity . In addition, if we combine the resu lts from the post ho c tests in research question 1 with the preceding result , !'-!a pattern seem to emerge: between the two high -anxiety g roups, the one that was told Chinese was ea sy scored higher than the group that was told Chinese was difficult. This does , thus, lend support to some of the earlier literature on the facilitating effects of anxiety on language learning (e.g. Chastain, 1975; Kleinmann, 1977) . However, this facilitating effect only showed for the people who were told Chinese is easy to learn. It is worth noticing that the anxiety discussed in this section is state anxiety, not FLA . FLA is defined as a type of situational -specific anxiety , which was considered developing from state anxiety (MacIntyre & Gardner, 1991c) . Since this study is not longitudinal in nature, whether or not learnersÕ state anxiety w ould develop into FLA is a question that is not yet resolved . Chapter summary This chapter summarized and discussed the results of the study , with references t o each research question . Relevant literature was also considered and compared with results of the current study. This study demonstrated the effects of perceived languag e difficulty on high -anxiety learners. Basically, when high -anxiety learners were told Chinese was difficult, they significantly performed worse than the learners who were told Chinese was easy. Also, learners who belonged to high -anxiety easy group Õs state anxiety significant ly fluctuated as the lesson progressed. Across three times of measuring, there was only one significant group difference at Time 3, with the high -anxiety groups rated more anxious than the low -anxiety groups. At last, t he high -anxiety easy group rated themselves more anxio us at the last time of measurement than the low -anxiety easy group, which was well as expected. It was also obvious that the results from multiple regression and two -way ANOVA were conflicting. It could be because 1) a larger sample size was required for the multiple regression, since the effect size was extremely small; !'%!2) due to celling effect in the quiz scores collected, data was not normally distributed, thus, it could affect the power of b oth statistical analy ses. The analysis of results presented in this chapter calls for further inquir ies on perceived language difficulty involve more affective variables or more participants. Also, to make the conclusion of whether or not perceived language difficulty mediates FLA, more research with longer research duration need s to be done . !'&!CHAPTER 6 CONCLUSIONS This chapter presents the major findings of the study, followed by implications for teachers and resea rchers in the field who want to inquiry further on this topic. The studyÕs limitations are then discussed. Summary of key findings This study was aimed for investigating the effects of perceived language difficulty and its relationship with FLA. Particularly, whether language difficulty could pose influence s on language learning in the context of English speakers who learn Chinese. The study also examined changes of learnersÕ state anxiety levels in a foreign language class with references to thei r anxiety levels and conditions . Sixty -eight undergraduate students who enrolled in a language learning and teaching course at Michigan State University participated in the study. Data was collected through the following instruments: the modified FLCA S que stionnaire, the Anx ometers and the translation quiz. Participants who were qualified for this study were told different facts about Chinese. Then the researcher taught a Chinese lesson and collected the data. The key findings of the research revealed th at p erceived language difficulty seemed to have a debilitating effect o n anxious learners . This result then could have possibly confirmed that there was a role for perceived language difficulty in foreign language classrooms . The results from comparing learners Õ state anxiety across times of measurement and groups seemed to suggest that in -class activities were source of changing learners Õ state anxiety. But whether or not they are source that developed learnerÕs FLA in a longe r time frame is a question that the results of the current study cannot answer . !''!Pedagogical and theoretical implications The above results showed the debilitating effect of perceived language difficulty should be considered seriously in FL classrooms. There is by far no peer -reviewed published evidence that show s Chinese or any other lang uages are more difficult or easier for E nglish speakers to learn. FL teachers should recognize the negative effects of reinforcing high -anxiety learnersÕ conventional wisdom about language difficulty hierarchy . FL teachers should help learners go through deconditioning and gradually get rid of their p ersonal language learning myths (Horwitz, 1988), and stop rein forcing them and start recognizing the effects of perceived language difficulty could be the first step. In addition, further research of perceived language difficulty should consider using a target language other than Chinese. It is possible that participants in this study did not ÒbelieveÓ the cue of Chinese is easy to learn, because they already had stereo types toward Chinese, even though they had not learned with any formal instructions. Since this study did not contain any delayed posttests, it is hard to see the long -term effects for both perceived language difficulty and FLA. Thus, longitudinal research studies are also needed in the future to keep tracking learnersÕ FLCAS levels and state anxiety changes. Only by then, a firm conclusion can be made about whether or not perceived language difficulty is source of FLA. Limitations Results of the study must be viewed with the following limitations. The most obvious limitation is the study design. Firstly, the study only tested the short -term effects of perceived lang uage difficulty and FLA. Research on the same area with longer research duration is definitely needed. Secondly, one of the instruments, the translation quiz had a ceiling effect for participants. This led to the negative skewed data, which created a not normally distributed data !'$!set. Thirdly, the cut -off score (FLCAS = 100) that used to differentiate high -anxiety participants from low -anxiety group was arbitrarily chosen at the early stage of the study , since t here was no cut -off score set by the researchers who invented the scale (Horwitz et al. 1986). This study was further limited by the target language used. Because the study used Chinese pinyin (without the writing system), the results cannot be generalized to other languages or Chinese with its writing system. However, given the time constraints (an hour and twenty minutes), teaching Chinese without characters is unavoidable, because learning the writing system c ould cost some time. It would be ideal if the duration of the treatment could be extended and the writing syst em could be included. !'(! APPENDICES !')!Appendix A Ð FLCAS (modified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ppendix B Ð Anx ometer (Time 1 only) Time 1 Name_______________________________________ Place an X next to the number that best capture s how anxious or nervous you are feeling. !!&C079/$ !&!H:0 @25=5; !! !$&!Appendix C Ð Translation quiz Name______________________________ *Translate these sentences into English: 1. sh " z‹i b # o sh‹ngmian 16. y $ njfing z‹ich % p⁄ngbi # n ___________________________ ___________________________ 2. b & z‹i sh 'uj ( xi‹mian 17. ) rj ( z‹i sh " sh‹ngmian ____________________________ ___________________________ 3. sh ' uj ( z‹i b # o l &mi‹n 18. ch % z‹i b # o l & mi‹n ____________________________ ___________________________ 4. w⁄wa z‹ich % p⁄ngbi # n 19. y‹oshi z‹i sh " sh‹ngmian ____________________________ ___________________________ 5. sh " z‹i b # o w‹imian 20. b & z‹i qi⁄nb # o w‹imian ____________________________ ___________________________ 6. sh 'uj ( z‹i w⁄wa p⁄ngb i# n ____________________________ 7. y$ njfing z‹i sh " sh‹ngmian. ____________________________ 8. w⁄wa z‹i y $ njfing p⁄ngbi # n ____________________________ 9. )rj ( z‹i sh " xi‹mian ____________________________ 10. sh ' uj ( z‹i qi⁄nb # o l & mi‹n ____________________________ 11. qi⁄nb # o z‹i b # o w‹imian ____________________________ 12. y‹oshi z‹i b # o l & mi‹n ____________________________ 13. y$ njfing z‹i sh " xi‹mian ____________________________ 14. b& z‹i ) rj ( p⁄ngbi # n ____________________________ 15. y‹oshi z‹i b # o l &mi‹n ______________________________ ! !$'!Appendix D Questionnaires for both conditions Part 1 Ð Questionnaire for ÒChinese is easyÓ group Name___________________________________________ Please check one: ____ I donÕt know any Chinese. ____ I can speak a little Chinese. ____ I can speak Chinese well or I am a native speaker of Chinese. The purpose of this study is to determine how well you can learn some Chinese when the teacher does not speak any English during the session. We also want to determine if using this method makes you nervous. In the lesson, we will focus only on grammatical structures that are very easy for English speakers to learn. As you may know, Chinese is a difficult language to learn to read because of the characters. Learning to speak Chinese, however, is quite easy for several reasons. Do you know or can you guess what features make Chinese easy for English speakers? Write your response below. DO NOT TURN OVER THIS PAPER UNTIL WE ASK YOU TO. !$$!Here are some of the many reasons that Chinese is easy. First, the pronunciation is fairly easy for English speakers because they can distinguish the different sounds quite easily. Second, unlike Spanish, French, or German, Chinese pronouns have no gender or case marking. One word , ta, means he, she , him , and her . Furthermore, there are no verb endings. For example, the word for eat is always che . It does not have to agree with the subject. Third, Chinese nouns have no gender or plural forms. Fourth, Chinese word order is usually the same as Engl ish word order, meaning that the sentences follow subject, verb, object order. Other languages, such as Japanese, are much more difficult because the verb comes at the end of the sentence. Fifth, Chinese has no past or future tense. Time is marked by adv erbs. For example, you would say, I eat yesterday to show past. The verb does not change. Did you know any of these facts about Chinese before you read this? If so, which? !$(!Appendix D Part 2 Ð Questionnaire for Ò Chinese is difficult Ó group Name___________________________________________ Please check one: ____ I donÕt know any Chinese. ____ I can speak a little Chinese. ____ I can speak Chinese well or I am a native speaker of Chinese. The purpose of this study is to determine how well you ca n learn some Chinese when the teacher does not speak any English during the session. We also want to determine if using this method makes you nervous. In the lesson, we will focus on grammatical structures that are usually very difficult for English spea kers to learn. We are doing this study because C hinese is a difficult language to learn to read , in part, because of the characters. Learning to speak Chinese is quite difficult as well for several reasons. Do you know or can you guess what features make Chinese difficult for English speakers? Write your response below. DO NOT TURN OVER THIS PAPER UNTIL WE ASK YOU TO. !$)! Here are some of the many reasons that Chinese is difficult. First, the pronunciation is very difficult because of the tones, which do not occur in English. Second some of the sounds are very similar to each other, and English speakers cannot hear the difference. Third, Chinese has words that indicate the linguistic feature aspect . This is has to do with features of the ve rb such as completion. It is not the same as past tense. For example, if you want to say that something will be completed in the future, you have to mark completion on the verb even though it has not happened yet . Fourth, word order in Chinese can be mani pulated. For example, the object can be moved to the position before the verb resulting in something like, He the -book put -down. Fifth, Chinese has what are called resultative complements. The verbs that correspond to listen and look in English, for examp le, indicate only the sensory actions. The verb for to perceive has to be added to the ver bs if perception has occurred. One would say something like, I looked -perceived a dog . DonÕt worry if you did not understand these difference s, but did you know a ny of these facts about Chinese before you read this? If so, which? ! !$*! BIBLIOGRAPHY !$+!BIBLIOGRAPHY Aida, Y. (1994). Examination of Horwitz, Horwitz, and Cope's construct of foreign language anxiety: The case of students of Japanese. The modern language journal, 78 (2), 155-168. Bailey, K. M. (1983). Competitiveness and anxiety in adult second language learning: Looking at and through the diary studies. Bernat, E., & Gvozdenko, I. (2005). Beliefs about language learning: Current knowledge, pedagogical implications, and new research directions. TESL -EJ, 9 (1), 1-21. Chastain, K. (1975). Affective and a bility factors in second language acquisition. Language learning, 25 (1), 153-161. Cheng, Y. s., Horwitz, E. K., & Schallert, D. L. (1999). Language anxiety: Differentiating writing and speaking components. Language learning, 49 (3), 417-446. DLI's Langu age Guidelines. (2010). Retrieved 13 August, 2015, from PC56"6TU1@!F@3A8J@F0 ! Elkhafaifi, H. (2005). Listening comprehension and anxiety in the Arabic language classroom. The modern language journal, 89 (2), 206-220. Field, A. (2013). Discovering Statistics using IBM SPSS Statistics : SAGE Publications. Horwitz. (1986). Preliminary evidence for the reliability and validity of a foreign language anxiety scale. Tesol Quarterly, 20 (3), 559-562. Horwitz. (1988). The beliefs about language learning of beginning university foreign language students. The modern language journal, 72 (3), 283-294. Horwitz. (1999). Cultural and situational influences on foreign language learners' beliefs about language learning: a review of BALLI studies. System, 27 (4), 557-576. !$,!Horwitz. (2001). LANGUAGE ANXIETY AND ACHIEVEMENT. Annual Review of Applied Linguistics, 21 , 112-126. doi: doi:null Horwitz, Horwitz, M. B., & Cope, J. (1986). Foreign language classroom anxiety. The modern language journal, 70 (2), 125-132. Horwitz, & Young, D. J. (1991). Language anxiety: from theory and research to classroom implications : Prentice Hall. Kleinmann, H. H. (1977). Avoidance behavior in adult second language acquisition1. Language learning, 27 (1), 93-107. Koch, A. S., & Terrell, T. D. (1991). Affective reactions of foreign language students to natural approach activities and teaching techniques. Language anxiety: From theory and research to classroom implications , 109-126. Language Difficulty Ranking (2015). Retrieved 13 August 2015, from C==37VVBBB85<<5F="#541:E?1E5451;:":E8F@2V41:E?1E5 QE?"J5V41:E?1E5QJ"<<"F?4=A ! Larson -Hall, J. (2009). A Guide to Doing Statistics in Second Language Research Using SPSS : Taylor & Francis. MacIntyre, P. D., & Gardner, R. C. (1991a). Investigating language class anxiety using the focused essay technique. The modern language journal, 75 (3), 296-304. MacIntyre, P. D., & Gardner, R. C. (1991c). Language Anxiety: Its Relationship to Other Anxieties and to Processing in Native and Second Languages*. Language learning, 41(4), 513-534. MacIntyre, P. D., & Gardner, R. C. (1991f). Methods and Results in the Study of Anxiety a nd Language Learning: A Review of the Literature*. Language learning, 41 (1), 85-117. MacIntyre, P. D., & Gardner, R. C. (1994). The subtle effects of language anxiety on cognitive processing in the second language. Language learning, 44 (2), 283-305. !(-!MacIntyre, P. D., Noels, K. A., & Cl”ment, R. (1997). Biases in se lfratings of second language proficiency: The role of language anxiety. Language learning, 47 (2), 265-287. Oh, M. -j. T. (1996). Beliefs about language learning and foreign language anxiety: A study of American university students learning Japanese. University of Texas at Austin. Osborne, J. W. (2001). Testing stereotype threat: Does anxiety explain race and sex differences in achievement? Contemporary Educational Psychology, 26 (3), 291-310. Rosenthal, R. (1994). Interpersonal expectancy effects: A 30 -year perspective. Current Directions in Psychological Science , 176-179. Saito, Y., Garza, T. J., & Horwitz, E. K. (1999). Foreign language reading anxiety. The modern language journal, 83 (2), 202-218. Scovel, T. (1978). The effect of affect on foreign language learning: A review of the anxiety research. Language learning, 28 (1), 129-142. Sheen, Y. (2008). Recasts, language anxiety, modified output, and L2 learning. Language learning, 5 8(4), 835-874. Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat and women's math performance. Journal of experimental social psychology, 35 (1), 4-28. Spielberger, C. D. (1983). Manual for the State -Trait Anxiety Inventory STAI (f orm Y)(" self -evaluation questionnaire"). Steele, C. M. (1997). A threat in the air: How stereotypes shape intellectual identity and performance. American psychologist, 52 (6), 613. Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African Americans. Journal of personality and social psychology, 69 (5), 797. Steinberg, F. S., & Horwitz, E. K. (1986). The effect of induced anxiety on the denotative and interpretive content of second language speech. Tesol Quarterly, 20 (1), 131-136. !(%! Young, D. J. (1991). Creating a Low Anxiety Classroom Environment: What Does Language Anxiety Research Suggest? The modern language journal, 75 (4), 426-437.