STUDENT ATTITUDES TOWARD ACCENTEDNESS OF NATIVE AND NON-NATIVE SPEAKING ENGLISH TEACHERS By Laura Ballard A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Teaching English to Speakers of Other Languages - Master of Arts 2013 ABSTRACT STUDENT ATTITUDES TOWARD ACCENTEDNESS OF NATIVE AND NON-NATIVE SPEAKING ENGLISH TEACHERS By Laura Ballard My goal for this study was to examine participants’ familiarity with specific accents, whether participants were able to identify if a speaker was a native speaker (NS) or a non-native speaker (NNS), and what accent the speaker had. I also examined how the participants rated speakers on four Likert-scales of comprehensibility, intelligibility, accentedness, and acceptability as a teacher (the four dependent variables). I included 38 NS and 94 NNS participants from a range of first-language backgrounds. The participants listened to three NSs (Midwestern U.S., Southern U.S., and British) and two NNSs (Chinese and Albanian) and completed the identification and Likert-scale tasks outlined above. Results showed that NNSs were significantly less able than NSs to identify a speaker’s nativeness and accent. Results revealed that familiarity with an accent correlated with comprehensibility and acceptability as a teacher. For familiar accents, familiarity was a significant predictor of the participant ratings on the four dependent variables, though the predicted changes in ratings were small. Overall, participants had generally positive attitudes toward NNSETs; in relation to acceptability as a teacher, accent was the least influential of the dependent variables. I conclude by discussing that students should be exposed to a range of different accents, as familiarity with an accent facilitates comprehension. These findings also challenge current language center hiring practices that exclude NNSETs from jobs based on a non-native status; this study supports the notion that administrators should hire English language teachers based on professional credentials, and not based on accent. ACKNOWLEDGEMENTS I would like to express my gratitude to all of the people who made this research possible. I am indebted to them in many ways, and I know that the success of this project is theirs as much as it is mine. I am grateful to Michael Kramizeh for his help in the audio and computer labs, to the speakers who volunteered their time in making the recordings, and to the instructors who graciously gave their class time for this project. Thank you to my family, friends, classmates, fellow Graduate InterVarsity members, and co-workers for their endless support and encouragement through this portion of my academic endeavors, especially during this thesis semester. I am deeply appreciative to my committee members, Dr. Paula Winke and Dr. Debra Hardison, who agreed to take on this project late in the game. Thank you for your guidance, support, and valuable feedback. I would especially like to express my gratitude to my adviser, Dr. Paula Winke, who has been endlessly patient in every meeting, generous in giving her time, and encouraging when everything seemed impossible. Thank you for seeing my potential to pursue this path and for faithfully seeing me through. I am grateful. Finally, I am thankful for the opportunity to investigate the mysteries of the world around us in an academic context, and for the calling that we have to do so. As The Word says, “It is the glory of God to conceal a matter; to search out a matter is the glory of kings” (Proverbs 25:2). iii TABLE OF CONTENTS LIST OF TABLES ........................................................................................................................................ V LIST OF FIGURES .................................................................................................................................... VI CHAPTER 1: INTRODUCTION .................................................................................................................. 1 CHAPTER 2: LITERATURE REVIEW ....................................................................................................... 5 CHAPTER 3: METHODS ............................................................................................................................. 9 Participants ............................................................................................................................ 9 Materials ............................................................................................................................... 10 Procedure ............................................................................................................................. 11 Analysis ................................................................................................................................. 13 CHAPTER 4: RESULTS ............................................................................................................................. 14 Research Question 1 ............................................................................................................ 14 Research Question 2 ............................................................................................................ 24 CHAPTER 5: DISCUSSION ...................................................................................................................... 36 Research Question 1 ............................................................................................................ 36 Research Question 2 ............................................................................................................ 39 Unfamiliar Accents .............................................................................................................. 41 CHAPTER 6: CONCLUSION .................................................................................................................... 46 Limitations ........................................................................................................................... 46 Implications .......................................................................................................................... 47 Directions for Future Studies ............................................................................................. 48 APPENDICES ............................................................................................................................................. 49 Appendix A: Table 12 ........................................................................................................... 50 Appendix B: Familiar Topic Transcript: Orientation on Classroom Expectations .............. 51 Appendix C: Unfamiliar Topic Transcript: How to Make Pottery....................................... 52 Appendix D: Dictation Task Sentences ................................................................................. 53 Appendix E: Sample Listening Page ..................................................................................... 54 Appendix F: Additional Figures ............................................................................................ 56 REFERENCES ............................................................................................................................................ 60 iv LIST OF TABLES Table 1: Participants by Group ....................................................................................................... 9 Table 2: Participants by Program ................................................................................................... 9 Table 3: Correct Identification of Speakers as NS or NNS by Participant Group (NS and NNS) ............................................................................................................................................... 14 Table 4: Summary of Correct Identification of Speaker Accent by Participant Groups .............. 15 Table 5: Group Statistics from an Independent Samples T-Test Between NS and NNS Participants for Correct L1 Accent Guesses .......................................................................... 16 Table 6: Independent Samples T-Test between NS and NNS Participants for Correct L1 Accent Guesses .................................................................................................................................. 17 Table 7: Participant Guesses of Speaker Accent by Participant L1 Group .................................. 18 Table 8: Percentage of Correct Identification of Nativeness and Accent by NNS Particpants by Program Level ....................................................................................................................... 23 Table 9: Spearman's Correlations between Participants' Reported Accent Familiarity and Participants' Actual Ratings of Four Variables for Corresponding Speakers........................ 27 Table 10: Familiarity as a Predictor for Participant Ratings of Comprehensibility, Accentedness, Intelligibility, and Acceptability as a Teacher (Simple Linear Regressions) ........................ 28 Table 11: Spearman's Correlations between Participants' Ratings of Acceptability as a Teacher and Comprehensibility, Intelligibility, and Accentedness ..................................................... 34 Table 12: Participants by Nationality ............................................................................................ 50 v LIST OF FIGURES Figure 1. Percentage of Correct Identification of Nativeness and Accent by NNS Participants by Program Level. ...................................................................................................................... 24 Figure 2. Percentage of Correct Identification of Nativeness and Accent by NNS and NS Participants. ........................................................................................................................... 37 Figure 3. Composite Ratings of Accentedness and Acceptability as a Teacher by All Participants. ............................................................................................................................................... 43 Figure 4. Composite Ratings of Comprehensibility by All Participants. ...................................... 56 Figure 5. Composite Ratings of Accentedness by All Participants. .............................................. 56 Figure 6. Composite Ratings of Intelligibility by All Participants. ............................................... 57 Figure 7. Composite Ratings of Acceptability as a Teacher by All Participants. ......................... 57 Figure 8. Composite Ratings of Familiarity with Accents by All Participants. ............................ 58 vi CHAPTER 1: INTRODUCTION In the wake of a recent law passed in Arizona that is based on the No Child Left Behind Act (see Title III, Sec. 3116, c), which threatens the jobs of non-native speaking English teachers (NNSET), the qualifications of both native speaking English teachers (NSET) and NNESTs have come into question. Some argue that this law, which mandates that K-12 English language teachers must be proficient speakers of English (which Arizona has partially interpreted as having a near-native accent, or a least not having a noticeable foreign accent) is clearly employment discrimination (Blum & Johnson, 2012) that has been justified by the myth that English language learners (ELL) will not be able to acquire native-like English from non-native speakers (NNS). In a study conducted by Moussu (2010), around 60% of English language program administrators surveyed admitted that the primary factor they consider when hiring new teachers is being a native speaker (NS). Because of this continued practice of what could be viewed as discrimination again NNSETs, many in the English-as-a-second-language (ESL) field have sought to defend NNSETs by highlighting their natural strengths as English teachers. Among these qualities are the ability to serve as successful role models for ELLs, a more comprehensive knowledge of English grammar, a better grasp of learning strategies, an anticipation of potential language problems, and a greater empathy for student needs and problems (Hertel & Sunderman, 2009; Moussu, 2010). Furthermore, when there is a match between the teacher’s and students’ L1, the teacher may be able to use the L1 as a beneficial teaching tool: some have argued that some use of the L1 is beneficial (Hertel & Sunderman, 2009; Marian, Blumenfled, & Boukrina, 2008; Thoms, Liao, & Szustak, 2005; Park, 2012). Despite this realization of what NNESTs bring to the classroom, language center administrators are still more likely to hire NSETs. This practice is based on the idea that NSETs are better and 1 more qualified, which is a belief Philipson (1992) termed a native speaker fallacy. This fallacy has even been perpetuated in a way that impacts education policies, as it has in Arizona. Many in the Teaching English to Speakers of Other Languages (TESOL) profession argue that administrators should hire teachers on the basis of linguistic ability and professional skill (Holliday, 2008; Moussu, 2010; Todd & Pojanapunya, 2008), and not based on accent, as many fear the new law in Arizona is advocating. This employment discrimination disregards the benefits of exposure of language learners to a variety of cultures and accents and, whether intended or not, will promote “linguistic sterilization” (Blum & Johnson, 2012, p. 175). This may be especially concerning when one considers the growth in the number of non-native English speakers around the world. Canagarajah (2005) estimated that almost 80% of English speakers worldwide are non-native. Because of this, notions of World English, English as an international language (EIL), and English as a lingua Franca (ELF) are becoming more prevalent (Llurda, 2004; Blum & Johnson, 2012; Moussu, 2010). Providing persuasive evidence for this argument is the inclusion of accented, non-native English on some versions of the International English Language Testing System (IELTS) listening section (Harding, 2012; Moussu, 2010) and the recent creation of the Vienna-Oxford ELF Corpus, a corpus that compiles English language samples from the non-native English speaking world (Llurda, 2004). English language teachers, both teachers of ESL and English as a foreign language (EFL), need to help their students become aware of the varieties of English used across the globe. This is because non-native English, such as EIL and ELF, are becoming stabilized languages and are used frequently in educational, political, and economic contexts. Alptekin (2002) suggested that NNESTs have a strong role in filling the gaps between the native-speaker and nonnative-speaker reality: 2 With the increasing establishment of English as the world lingua franca, nonnative speakers will be in optimal positions to lead their students into the realm of EIL. Teachers of EIL should incorporate instructional materials and activities rooted in local as well as international contexts that are familiar and relevant to language learners’ lives. (p. 318) Blum and Johnson (2012), among others, argued that there is no longer such a thing as pure or “proper” English. Even among English speakers who learned English as their native language, such as people from Nigeria, Guyana, Australia, and Singapore, there is no universal standard, especially in terms of accent and lexicon. Considering this, it would seem rather odd to enforce a law under which teachers could only be employed if they had a local English accent. This law would seem quite unreasonable, especially in a collegiate ESL context where the majority of international students plan to return to their own country where many expect to use English for international business purposes. Not only this, but international students attending large American universities will likely also encounter a number of international teaching assistants during their time as undergraduates. As I will review shortly, having an exposure to nonnative speakers throughout their ELL career would benefit learners and aid in their comprehension with other non-native speakers. Despite these benefits for ELLs, the question still remains about the legitimacy of the Arizona law and the hesitancy to hire NNESTs. To help answer this question, in this study I investigate the issue of non-native-teacher accents from the perspective of the learner by measuring student judgments of comprehensibility and intelligibility of native and non-native English accents in relation to students’ perceptions of a speaker’s acceptability as a teacher. Before describing the study in depth, I highlight past research on accent, student perceptions, and 3 native and non-native teachers. 4 CHAPTER 2: LITERATURE REVIEW Alford and Strother (1990) conducted a study to capture native and non-native students’ attitudes toward regional U.S. accents. They showed that non-native speakers were able to detect differences between the accents, which is contradictory to research by Derwing and Munro (1997), Mousssu (2010), and Scales et al. (2006), who demonstrated that ELLs had difficulty in distinguishing between different accents, whether they be native accents, non-native accents, or a mix of the two. This was also evidenced when students reported that they wished to acquire an American accent, yet they could not identify an American accent, even after having lived in the U.S. for a number of months or years (Scales et al., 2006). Scales et al. suggested that a desire to have an American accent, paired with an inability to identify one, possibly reflects an idealized conception of what a native accent really is (Timmis, 2002). If students find it difficult to identify a native accent, perhaps their attitudes toward NNSETs based on pronunciation are invalid. Despite learners’ inability to distinguish between NS and NNS accents, in Munro et al. (2006), the participants were able to reliably rate speakers’ intelligibility, comprehensibility, and accentedness. This was true across groups of students with different L1 backgrounds, including a group of native speakers, and their assessment of speakers was surprisingly comparable. In Scale et al.’s (2006) study, students tended to like an accent more if they rated it as more comprehensible. This suggested that students are concerned with listening comprehension, which likely plays a role in student perceptions of teachers. A number of studies (Gass & Varonis, 1984; Winke, Gass, & Myford, 2013; Harding, 2012; Derwing & Munro, 1997) have also looked at how familiarity relates to comprehension and listener bias. Gass and Varonis looked at four aspects of familiarity: topic familiarity, 5 different L1 background familiarity, same L1 background familiarity, and speaker familiarity. Results showed that being familiar with a topic, a specific speaker, and non-native speech in general facilitates comprehension. They also showed that being familiar with a particular nonnative accent facilitates comprehension of other non-native speaker with the same background. This last finding was also consistent with Winke, Gass, and Myford’s findings, showing that raters who were familiar with a speaker’s L1 were better able to orient themselves to the speaker’s language. In the same vein, for testing purposes, Harding (2012) looked at accent familiarity among ELL students of the same L1 as the speaker. Results showed that students were sometimes able to comprehend more from someone who was speaking English and had the same L1, calling this an L1 advantage. These finding suggest that the more familiar a student is with their teacher and his or her linguistic background, and even his or her L1, the easier it will be for them to understand their teacher. Moussu’s (2010) findings (that student ratings of their teachers (both NSETs and NNESTs) over the course of a semester improved) provide additional evidence that exposure to an accent and a particular speaker facilitates comprehension. This is important when thinking about students who have NNESTs who do have an accent. These studies suggest that it is only a matter of time before a student will become comfortable with their teacher’s speech, if they wish to do so. In the case that a teacher, or speaker, has a strong accent, it does not necessarily mean that the speaker is incomprehensible. This conclusion was consistent with Derwing and Munro’s findings (1997, 2005) in which participants rated accent most harshly, followed by comprehensibility, and then intelligibility. In other words, though the participants judged a speaker to have a heavy accent, the raters admitted to being able to understand the speaker. The researchers suggested that the harsh ratings for accent and comprehensibility might have been 6 due to the fact that a strong accent requires more processing time, and in turn is more difficult to understand. It’s important to consider why accent might be important in terms of the NSET versus NNSET debate. According to Moussu (2010), a teacher’s L1 is only one of the many variables that affect how students view their teacher. In her study, Moussu collected data at the beginning and end of a semester and found that students’ views of their NNESTs became more positive over the semester. This finding was corroborated by Todd and Pojanapunya (2009) in their findings that, while students made explicit indications that they preferred NSETs, their implicit attitudes suggested that the students value NNSETs equally and that they had “warm feelings” toward NNESTs and not toward NSETs. Hertel and Sunderman (2009) also found that, while students prefer NSET for pronunciation purposes, students recognized NNESTs’ ability to give grammatical rules and explain vocabulary. Among other factors that affect students’ perceptions toward teachers are the students’ expected grades, the students’ majors, and a teacher’s country of origin (Moussu, 2010). Drawing upon the conclusions from these studies, it is clear that more research needs to be done that will clarify the issues that have seen mixed results. In this study, I seek to add to the literature on accentedness, student attitudes, and NESTs and NNESTs. Especially considering the current policies coming into play regarding English-teachers’ accents (i.e., the law in Arizona that prohibits heavily accented NNSs from teaching), I hope to shed light on students’ attitudes toward such teachers. Research questions: 1a. Are ESL learners able to distinguish between native and non-native speakers of English? 7 1b. If so, can the ESL learners identify the speakers’ accents? 2a. Do students who are familiar with a particular accent rate speakers differently in terms of comprehensibility, intelligibility, accentedness, and acceptability as a teacher? 2b. If so, is how they rate related to their attitudes about acceptability as an English teacher? 8 CHAPTER 3: METHODS Participants As shown in Table 1 and Table 2, 132 individuals participated in this study. I recruited these participants from nine classes at a large Midwestern university. The participants included Table 1 Participants by L1 Group NNS of English NS of English Total Male Female Total Male Female Total Male Female Total Chinese 42 14 56 0 0 0 42 14 56 Arabic 11 8 19 0 0 0 11 8 19 Korean 4 3 7 0 0 0 4 3 7 English 0 0 0 5 33 38 5 33 38 Other 5 7 12 0 0 0 5 7 12 Total 62 32 94 5 33 38 67 65 132 Table 2 Participants by Program B.A. Master's Ph.D. VIPP ESL2   ESL4   ESL5   Total Male 0 3 0 0 19 17 23 62 Female 1 4 2 3 7 8 7 32 Total 1 7 2 3 26 25 30 94 NS of Male 4 1 0 0 0 0 0 5 English Female 31 2 0 0 0 0 0 33 Total 35 3 0 0 0 0 0 38 Total Male 4 4 0 0 19 17 23 67 Female 32 6 2 3 7 8 7 65 Total 36 10 2 3 26 25 30 132 Note. VIPP refers to visiting professionals and scholars. ESL refers to students who are NNS of English currently enrolled in English-as-a-second-language classes; ESL2 is equivalent to intermediate low. ESL4 to intermediate high, and ESL5 to advanced low. 9 93 international students, the majority of whom were enrolled in part-time or full-time ESL classes in three proficiency levels: level 2 (n=26), level 4 (n=25), and level 5 (n=30). Of these international students, Chinese was the largest L1 group (n=56), followed by Arabic (n=19), Korean (n=7), and several other languages spoken by three or fewer participants, from this point on referred to as other (n=12). I also recruited a control group (n=38) of American L1 English speakers from two undergraduate and one graduate-level linguistics courses. All participants ranged in age from 18 to 51 years, with a mean of 22.4 years of age. For a more detailed report of the participants, see Appendix A. Materials In this study, I had one main instrument, a multi-faceted, web-based survey. The survey had four main components: a background questionnaire, Likert scale questions to be answered in response to listening tasks, a dictation task, and an indication of the speakers’ native status and accent. Background questionnaire. I adapted the questions on the background questionnaire from Harding (2011). The questions addressed the participants’ exposure to other languages and to English spoken by a variety of native and non-native speakers. The participants used the questionnaire to indicate their preferred accent or accents when listening to English and their general attitude about their own native-like-accent goals. Audio files and Likert-scale questions. The audio files, which native speakers (American Midwestern from Michigan, British from Northern English, American Southern from Alabama, and Nigerian from Lagos) and non-native speakers (Chinese, Albanian, Malagasy, and Italian) recorded, are of three kinds: a lecture about a familiar topic that ranged between 20 and 24 seconds (classroom expectations given at an assembly that each ESL student attends, see 10 Appendix B); a lecture on an unfamiliar topic that ranged between 20 and 24 seconds (pottery making, adapted from Sueyoshi & Haridson (2005), see Appendix C); and unrelated sentences for transcription that are similar in syntax, topic, and length that ranged between 5 and 7 seconds (see Appendix D). The final group of sentences served as an audio prompt for a series of sentence-dictation tasks that could serve as an alternate measure of intelligibility, as has been done previously in research studies on accent (Derwing & Munro, 1997; Munro et al., 2006). Because of findings that topic familiarity influences comprehension (Derwing & Munro, 1997; Gass & Varonis, 1984), I have included familiar and unfamiliar topics. Dictation Task. The third component of the survey was a sentence-dictation task similar to those used in Derwing and Munro (1997) and Munro et al. (2006) in which a participant listens to and transcribes an audio segment. Nativeness and Accent. The final component of the survey required participants to indicate their guess about whether the speaker was a NS or a NNS. They were also required to guess the speakers’ accents. For a sample of a listening page for one speaker in the survey, see Appendix E. Procedure The first round of participants in this study was an intact graduate-level linguistics class. During class, the participants met in a computer lab to complete the survey. Each student had their own computer and a set of headphones for the listening portion of the survey. The participants first completed a familiarization task that is similar to the listening portion. In this familiarization task (and the listening section), participants responded to 9-point Likert-scale questions after hearing the lectures. The scales measured, for each participant, his or her estimation of the speaker’s intelligibility, comprehensibility, accentedness, and acceptability as a 11 teacher, with the participant considering the lecturers being ESL teachers. These Likert scales were presented in a random order to eliminate bias or an ordering effect. Following Munro et al.’s (2006) definitions, I define intelligibility as “the extent to which a speaker’s utterance is actually understood” (p. 112), comprehensibility as “the listener’s estimation of difficulty in understanding an utterance” (p. 112), and accentedness as “the degree to which the pronunciation of an utterance sounds different from an expected production pattern” (p. 112). Next, the participants completed the background questionnaire and the listening task with Likert-scale questions (as described above). The participants listened to each lecture from the familiar topic and unfamiliar topic groups twice. Each time participants listened to a lecture, they answered two of the Likert-scale questions. Thus, each participant listened to each lecture twice and, in total, answered four Likert-scale questions per lecture. For counterbalancing purposes, the participants listened to the speakers and answered the Likert-scale questions at random (which SurveyMonkey automatically arranges). Each participant listened to one familiar and one unfamiliar lecture from each speaker, totaling sixteen lectures. Next, the participants completed a dictation task. In this task, each participant completed one dictation per lecture. This task was presented at the end of each Likert-scale evaluation for each speaker. The participant listened to the recorded sentence once and then transcribed the sentence, to the best of their ability, using a text box. Finally, the participants indicated their best guess about the nativeness and accent of the given speaker. The first round of participants, however, completed an expanded survey that included eight speakers, four native and four non-native. After this group of participants completed the survey, they gave feedback to the researcher, and from a combination of the initial data and the feedback collected, this allowed me to remove any speakers that either had no 12 perceivable accent or that have such a strong accent that they are unintelligible. At this point, I narrowed the listening portion to include only the American Midwestern (from Michigan), American Southern (from Alabama), British (from Cumbria, Northern England), Chinese and Albanian speakers. Next, each of the remaining classes, with the exception of one undergraduate linguistics course, completed the survey in a computer lab during class time. They followed the same procedure; however, they did not give feedback, and they completed the shortened survey of only five speakers (10 total lectures). For one undergraduate linguistics class, I oriented the students to the survey and gave instructions during class time, but the students completed the survey on their own time. Analysis To gain an overall picture of the participant population, I calculated descriptive statistics that included nationality, age, sex, program enrollment, and L1. Then to answer the first research question, I calculated descriptive statistics to determine whether participants accurately identified the speakers’ accents and accurately identified whether the speaker was a NS or a NNS. To see if NS participants and NNS participants performed differently in their identification of speakers, I calculated an independent samples t-test. To answer the second research question, I calculated Spearman’s correlations for ordinal data using students’ self-reported familiarity with an L1 and their actual rating of comprehensibility, intelligibility, accentedness, and acceptability as a teacher for corresponding L1 speakers. I also calculated simple linear regressions to determine if familiarity with an L1 predicted ratings for comprehensibility, intelligibility, accentedness, and acceptability as a teacher. 13 CHAPTER 4: RESULTS Research Question 1 To answer the first research question, I calculated descriptive statistics of participants’ ratings of speakers as NS or NNS. These statistics revealed that NS participants, in general, were more able than NNS participants to distinguish between NSs and NNSs. On average, NS participants correctly identified the speakers’ native status 91% of the time, as compared to NNS participants, who correctly identified speakers only 68% of the time. Next, to see if these differences were significant for each speaker, I calculated an independent samples t test. Results from the t test revealed that in all cases, except for the British speaker, NS participants were significantly better at determining the native status of the speaker. This was true for both the familiar and unfamiliar lectures given by each speaker. Table 3 reports these results. Table 3 Correct Identification of Speakers as NS or NNS by Participant Group (NS and NNS) NS Familiar NNS t df Sig. NS Unfamiliar NNS t df Sig. Midwestern 100% 61% -7.649 91 0.000* 100% 69% -6.44 93 0.000* Southern 100% 83% -4.368 93 0.000* 97% 77% -4.06 130 0.000* British 68% 58% -1.124 72 0.265 60% 60% -0.03 68 0.974 Chinese 97% 86% -2.538 127 0.012* 100% 86% -3.86 93 0.000* Albanian 95% 57% -5.960 127 0.000* 92% 42% -7.37 117 0.000* Note. An asterisk denotes a significant p value of less than .05. The percentages reflect the mean for each participant group, which are also the percentage of participants who correctly identified the speaker as a NS or NNS. 14 Table 4 Summary of Correct Identification of Speaker Accent by Participant Groups Participant Group NNS of English NS of English Midwestern Familiar 38% 100% Unfamiliar 33% 89% Southern Familiar 12% 87% Unfamiliar 27% 63% British Familiar 30% 74% Unfamiliar 38% 45% Chinese Familiar 56% 45% Unfamiliar 34% 39% Albanian Familiar 2% 11% Unfamiliar 3% 21% Average Familiar 28% 63% Unfamiliar 27% 52% Note. Bolding across pairs denotes a significant difference between NS and NNS ratings. To answer the latter part of the first research question, I calculated descriptive statistics on participants’ guesses of speakers’ accents. Table 4 is a summary of these results. Raw scores for correct and incorrect answers revealed that, for every lecture, NSs were better able to identify the accent of the speaker, and in six of the ten lectures, NSs were at least 17% more accurate in their identification of the speakers’ accent. I then calculated independent sample t tests to see if the differences observed between participant groups were significant. Results revealed that differences between NS and NNS guesses were significant for the familiar Midwestern lecture, t(130) = −12.52, p = .000, the unfamiliar Midwestern lecture, t(130) = −8.05, p = .000, the familiar southern lecture, t(130) = −11.60, p = .000, the unfamiliar southern lecture, t(130) = −3.99, p = .000, the familiar British lecture t(130) = −5.07, p = .000, and the unfamiliar Albanian lecture, t(130) = −2.57, p = .014. Results for the averages were also significantly 15 Table 5 Group Statistics from an Independent Samples T-Test between NS and NNS Participants for Correct L1 Accent Guesses Lecture Midwestern Familiar Midwest Unfamiliar Southern Familiar Southern Unfamiliar British Familiar British Unfamiliar Chinese Familiar Chinese Unfamiliar Albanian Familiar Albanian Unfamiliar Participant Group N NNS 94 NS NNS NS NNS NS NNS NS NNS NS NNS NS NNS NS NNS NS NNS NS NNS 38 94 38 94 38 94 38 94 38 94 38 94 38 94 38 94 38 94 Mean .4860 Std. Error Mean .0501 .0000 .4727 .3110 .3232 .0000 .0488 .0505 .0333 .3426 .4442 .4889 .4598 .4463 .4860 .5039 .4998 .5039 .4764 .4954 .1451 .3110 .1767 .0556 .0458 .0793 .0474 .0724 .0501 .0817 .0516 .0817 .0491 .0804 .0150 .0505 .0182 Std. Deviation .372 1.000 .330 .895 .117 .868 .266 .632 .298 .737 .372 .447 .553 .447 .340 .395 .021 .105 .032 NS .221 38 .4132 .0670 Note. The mean can be read as the percentage of participants that correctly guessed the accent of the speaker, where 1.00 equals 100% of the participants guessing correctly, and .00 equals 0% of the participants guessing correctly. Bolding denotes that NS and NNS participants’ guesses were significantly different, as noted in Table 4. 16 Table 6 Independent Samples T- Test between NS and NNS Participants for Correct L1 Accent Guesses T test for Equality of Means df 93.000 Sig. (2tailed) .000* -8.053 -11.888 102.720 .000* -.56495 .07016 -.70410 -.42580 130 .000 -.75140 .06321 -.87645 -.62635 -3.992 -5.008 -.795 1.099 -.586 a 63.029 .000* -.36562 .09159 -.54864 -.18260 130 130 130 130 .000 .428 .274 .559 -.43897 -.07503 .10582 -.05431 .08765 .09442 .09631 .09263 -.61238 -.26183 -.08471 -.23757 -.26556 .11177 .29636 .12895 a 43.661 .118 -.08399 .05263 -.19007 .02210 t Midwestern Familiar Unfamiliar Southern Familiar Unfamiliar British Chinese Albanian Familiar Unfamiliar Familiar Unfamiliar Familiar Unfamiliar a -12.521 a -1.596 Mean Std. Error Difference Difference -.62766 .05013 95% Confidence Interval of the Difference Lower Upper -.72721 -.52811 a 42.583 .014* -.17861 .06946 -.31872 -.03850 -2.572 Note. An asterisk denotes statistically significant differences between NS and NNS participant guesses. a = equal variance was not found, thus the t and p values reflect the Levene’s test adjustment for that violation. 17 Table 7 Participant Guesses of Speaker Accent by Participant L1 Group Midwestern (Familiar) Accent Chosen by Participant Participant Group NNS of English NS of English Participant L1 Midwestern Chinese 16 Arabic 10 Korean 1 Other 8 Total English Total Total NNS of English NS of English Southern British Chinese Albanian I don't know 2 0 8 1 0 5 0 0 2 0 1 2 8 1 2 0 6 2 0 1 35 (38%) 38 38 (100%) 11 (12%) 0 0 (0%) 9 (10%) 0 0 (0%) 3 (3%) 0 0 (0%) 1 (1%) 0 0 (0%) 17 (18%) 0 0 (0%) 73 (56%) 11 (8%) 9 (7%) 3 (2%) 1 (1%) 17 (13%) 2 0 0 0 1 0 0 0 13 7 4 6 2 (2%) 0 0 (0%) 2 (2%) 1 (1%) 0 0 (0%) 1 (1%) 30 (32%) 3 3 (8%) 33 (25%) Chinese Arabic Korean Other 18 7 2 4 Total English Total Total 31 (33%) 34 34 (89%) 65 (49%) Midwestern (Unfamiliar) 12 2 3 1 0 0 1 0 16 (17%) 1 1 (3%) 17 (13%) 3 (32%) 0 0 (0%) 3 (2%) 18 Other Total 15 55 0 19 1 6 0 12 16 (17%) 92 0 38 0 (0%) 38 16 (12%) 130 8 1 1 0 11 (12%) 0 0 (0%) 11 (8%) 56 19 7 12 94 38 38 132 Table 7 (cont’d) Southern (Familiar) Participant Group NNS of English NS of English NNS of English NS of English Participant L1 Midwestern Chinese 29 Arabic 12 Korean 2 Other 7 Total English Total Total 50 (53%) 5 5 (15%) 55 (42%) Chinese Arabic Korean Other 14 5 1 5 Total English Total Total 25 (27%) 10 10 (11%) 35 (27%) Southern British 8 0 1 2 1 0 0 1 11 (12%) 33 33 (87%) 44 (33%) 2 (2%) 0 0 (0%) 2 (2%) Southern (Unfamiliar) 18 6 4 1 1 0 2 1 25 (27%) 24 24 (63%) 49 (37%) 8 (9%) 0 0 (0%) 8 (6%) 19 Chinese Albanian I don't know 2 0 7 1 0 6 0 0 3 0 0 2 3 (3%) 0 0 (0%) 3 (2%) 0 (0%) 0 0 (0%) 0 (0%) 18 (19%) 0 0 (0%) 18 (14%) 1 1 0 0 0 0 0 0 7 8 4 4 2 (2%) 0 0 (0%) 2 (2%) 0 (0%) 0 0 (0%) 0 (0%) 23 (24%) 3 3 (8%) 26 (20%) Other Total 9 56 0 19 1 7 0 12 10 (11%) 94 0 38 0 (0%) 38 10 (8%) 132 10 0 1 0 11 (12%) 1 1 (3%) 12 (9%) 56 19 7 12 94 38 38 132 Table 7 (cont’d) British (Familiar) Participant Group NNS of English Participant L1 Midwestern Chinese 3 Arabic 3 Korean 0 Other 0 Southern British 5 3 0 1 14 3 4 7 Chinese Albanian I don't know 3 2 10 0 0 8 0 0 1 0 0 1 6 (6%) 0 0 (0%) 9 (10%) 0 0 (0%) 28 (30%) 28 28 (74%) 3 (3%) 0 0 (0%) 2 (2%) 0 0 (0%) 20 (22%) 1 1 (3%) Total NS of English Total English Total 6 (5%) 9 (7%) 56 (43%) 3 (2%) 2 (2%) 21 (16) Other Total 19 56 2 19 1 6 3 12 25 (27%) 93 9 38 9 (24%) 38 34 (26%) 131 British (Unfamiliar) NNS of English NS of English Participant L1 Midwestern Chinese 9 Arabic 4 Korean 1 Other 0 Southern British Chinese Albanian I don't know 2 0 8 3 0 3 0 0 2 0 0 0 5 0 0 1 17 6 4 8 14 (15%) 0 6 (6%) 1 35 (38%) 17 5 (5%) 0 0 (0%) 0 13 (14) 1 Total 0 (0%) 1 (3%) 17 (45%) 0 (0%) 0 (0%) 1 (3%) Total 14 (11%) 7 (5%) 52 (40%) 5 (4%) 0 (0%) 14 (11%) Total English 20 Other Total 15 56 2 18 0 7 3 12 20 (22%) 93 19 38 19 (50%) 38 39 (30%) 131 Table 7 (cont’d) Chinese (Familiar) Participant Group NNS of English Participant L1 Midwestern Chinese 1 Arabic 3 Korean 0 Other 1 Southern British 3 0 0 1 0 1 0 1 NS of English 4 (4%) 0 2 (2%) 0 0 (0%) 0 (0%) 0 (0%) Total NNS of English 5 (5%) 0 Total NS of English Total English 5 (4%) 4 (3%) 2 (2%) Chinese (Unfamiliar) 3 2 0 3 0 0 0 0 Chinese Arabic Korean Other 1 1 0 0 Total English 2 (2%) 0 3 (3%) 0 5 (5%) 0 Total 0 (0%) 0 (0%) 0 (0%) Total 2 (2%) 3 (2%) 5 (4%) 21 Chinese Albanian I don't know 41 0 5 2 0 9 4 0 1 5 0 5 52 (56%) 0 20 (22%) 17 1 6 17 (45%) 1 (3%) 6 (16%) 69 (53%) 1 (1%) 26 (20%) 25 2 1 4 32 (34%) 15 15 (39%) 47 (36%) 2 0 1 0 10 11 4 5 3 (3%) 3 30 (32%) 5 3 (8%) 5 (13%) 6 (5%) 35 (27%) Other Total 5 56 2 18 2 7 1 12 10 (11%) 93 14 38 14 (37%) 38 24 (18%) 131 13 2 1 3 19 (20%) 15 15 (39%) 34 (26%) 56 19 7 12 94 38 38 132 Table 7 (cont’d) Albanian (Familiar) NNS of English Participant L1 Midwestern Chinese 12 Arabic 6 Korean 2 Other 2 Southern British 22 (24%) 0 7 (8%) 0 0 (0%) 0 0 (0%) 0 (0%) 0 (0%) Total NNS of English 0 0 0 0 Total NS of English 5 2 0 0 22 (17%) 7 (5%) 0 (0%) Total English Albanian (Unfamiliar) 10 2 1 0 1 0 0 0 Chinese Albanian I don't know 8 2 11 2 0 7 1 0 4 0 0 6 11 (12%) 2 (2%) 28 (30%) 7 4 9 7 (18%) 18 (14%) 4 (11%) 9 (24%) 6 (5%) 37 (28%) 3 2 0 0 1 0 0 2 13 6 1 5 16 8 4 2 Total English 30 (33%) 1 12 (13%) 0 2 (2%) 0 5 (5%) 3 3 (3%) 8 25 (27%) 7 Total 1 (3%) 0 (0%) 0 (0%) 3 (8%) 8 (21%) 7 (18%) Total NS of English Chinese Arabic Korean Other 31 (24%) 12 (9%) 2 (2%) 8 (6%) 11 (9%) 32 (25%) Other Total 17 56 1 18 0 7 4 12 22 (24%) 93 17 38 17 (45%) 38 39 (30%) 131 10 1 0 3 14 (15%) 19 19 (50%) 33 (26%) Note. Bolded columns correspond with the correct accent. Percentages are calculated per participant group. Grand totals are calculated per total number of participants. 22 55 18 6 12 91 38 38 129 different between NSs and NNSs on the familiar lectures (t(130) = 5.1139, p = .0001) and on the unfamiliar lectures (t(130) = 3.0147, p = .0031). For full results, see Table 5 and Table 6. Table 7 is an expanded table that includes all accent guesses by participants. Looking further into this question, I ran descriptive statistics for percentage of correct NNS ratings on nativeness and L1 accent by program level to uncover if there were differences in rating based on a participant’s English proficiency level. Here, I was interested in looking specifically at students currently enrolled in English language courses (ESL2- intermediate low, 1 ESL4-intermediate high, and ESL5- advanced low) and master’s students (TOEFL score of 100+). I did not include NNS participants from the VIPP or B.A. level because their proficiency level was not available. Results show that there were differences between these groups. Table 8 summarizes these findings, which are also graphically represented in Figure 1. These numbers show a clear trend; when proficiency level increases, ability to identify nativeness and accent increase. Table 8 Percentage of Correct Identification of Nativeness and Accent by NNS Participants by Program Level. Nativeness Familiar Unfamiliar ELC2 63% 68% ELC4 66% 67% ELC5 70% 61% MA 83% 83% Total 69% 67% L1 Accent Familiar Unfamiliar 17% 20% 25% 26% 31% 27% 40% 43% 27% 27% 1 The TOEFL, or the Test of English as a Foreign Language, is an academic English proficiency test administered by Educational Testing Services. For more information, see the TOEFL 23 90% 80% 70% 60% ESL2 50% ESL4 40% ESL5 30% MA 20% Total 10% 0% Fam. Unfam. Fam. Nativeness Unfam. L1 Accent Figure 1. Percentage of correct identification of nativeness and accent by NNS participants by program level. Note. For interpretation of the references to color in this and all other figures, the reader is referred to the electronic version of this thesis. The only exception to this trend that I found was when participants listened to the unfamiliar lecture and rated the speakers’ nativeness. In this case, the scores for participants in ESL2, ESL4, and ESL5 were different than expected; ESL2 was higher than expected and ESL5 was lower than expected. Research Question 2 To answer the second research question, I began by calculating Spearman’s correlations to uncover significant relationships between students’ reported familiarity with an L1 and their actual rating of the speakers’ comprehensibility, intelligibility, accentedness, and acceptability as 24 a teacher, as shown in Table 9. The results show that the correlations between participants’ selfreported familiarity with an accent and their actual ratings of the four variables were, at best, moderate (my benchmark for moderate being a coefficient ranging between .3 and .5, as outlined in Field (2009)). The strongest correlation was between Midwestern familiarity and comprehensibility of the familiar Midwestern lecture, rs=.44, p < .001, followed by British familiarity and intelligibility of the unfamiliar British lecture, rs=.41, p < .001, and Midwestern familiarity and acceptability as a teacher of the familiar Midwestern lecture, rs=.40, p < .001. All other coefficients were less than .39. It is worthwhile, however, to note that, across the board, every correlation with Albanian was less than .10 and was not statistically significant. With all other accents, most (if not all) of the correlations ranged between .10 and .39. In addition, Midwestern familiarity was significantly correlated with every variable for both the familiar and unfamiliar lectures. Importantly, in terms of categorical variable correlation, familiarity correlated with comprehensibility and acceptability for both lectures (with the exception of Albanian), whereas accentedness and intelligibility revealed mixed results. To obtain a deeper understanding of the answer to this research question, I also calculated simple linear regressions to see if familiarity with a particular accent would predict participants’ ratings on the four variables when listening to a speaker of the accent. These linear regression results are provided in Table 10. Within each regression, I wanted to see if familiarity with an accent (the independent variables: Midwestern, etc.) was associated with the four dependent variables (comprehensibility, accentedness, intelligibility, and acceptability as a teacher) and if so, to what extent. Results showed that, for both the familiar and unfamiliar Chinese lectures, familiarity with Chinese was a significant predictor of ratings given on all four variables. With 25 the exception of accentedness, this pattern was also true of the British lectures. With the exception of comprehensibility in the familiar lecture, familiarity with the Midwestern accent was a significant predictor of all of the variables in both of the Midwestern lectures as well. For Albanian and Southern, however, the results were quite different. For Albanian, familiarity with the accent was not a significant predictor for any of the variables in either lecture. Similarly, the Southern lectures did not show significant predictions, with the exception of comprehensibility and acceptability for the unfamiliar lecture. The strongest predictions between familiarity with an accent and ratings of the four variables can been seen for the unfamiliar British lecture (comprehensibility, B=-.378; accentedness, B=.177; intelligibility, B=-.395; and acceptability as a teacher, B= -.475), the unfamiliar Chinese lecture (comprehensibility, B= -.418.; accentedness, B= -.322; intelligibility, B= -.383; and acceptability as a teacher, B= -.406), and the unfamiliar Midwestern lecture (comprehensibility, B=-.449; accentedness, B=.375; intelligibility, B=-.379; and acceptability as a teacher, B= -.513). Here the B value, or the regression gradient, indicated the predicted movement on the rating scale of the dependent variables as they relate to the rating given for familiarity. In other words, if a participant moves one point up the familiarity scale (e.g. from 2 to 3), then the B value predicts that, in the case of acceptability as an English teacher in the unfamiliar Midwestern rating, the rating will move negatively (toward a lower number, or toward “acceptable”) on the scale by 51% of one point (i.e. the rating for acceptability would move half of a point toward “acceptable”). To answer the second half of the second research question, I calculated Spearman’s correlations to see if there were significant relationships between participants’ ratings of acceptability as a teacher with the other three variables. As shown in Table 11, results revealed 26 Table 9 Spearman’s Correlations between Participants’ Reported Accent Familiarity and Participants’ Actual Ratings of Four Variables for Corresponding Speakers Comp. Familiar Lecture Accent. Intell. Accept. Comp. Unfamiliar Lecture Accent. Intell. Accept. Midwestern Familiarity   -.269** -.368** -.244** -.300** -.439** -.295** -.397** -.403** British Familiarity   -.293** 0.054 -.292** -.245** -.341** 0.076 -.408** -.306** Southern Familiarity -.178* 0.109 -0.144 -.203* -.288** -0.158 -.303** -.257** Chinese Familiarity   -.217* -.320** -0.126 -.212* -.268** -.225** -.207* -.277** Albanian Familiarity -0.021 0.073 0.015 -0.151 0.029 0.038 0.08 0.031 Note. *Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). Bolded coefficients are greater than or equal to .3. 27 Table 10 Familiarity as a Predictor for Participant Ratings of Comprehensibility, Accentedness, Intelligibility, and Acceptability as a Teacher (Simple Linear Regressions) Midwestern (Familiar Lecture) Comp. (Constant) Unstandardized Coefficients Std. B Error 2.120 Standardized Coefficients Beta .397 Midwestern Familiarity Accent. (Constant) -.131 -.422 (Constant) 2.079 Midwestern Familiarity Accept. (Constant) -.152 2.804 Midwestern Familiarity -.283 .063 .000 3.069 5.882 -2.403 .018 -.769 -.075 .000 1.421 2.736 -1.859 .065 -.315 .010 .000 2.030 3.578 -2.932 .004 -.474 -.092 .000 3.220 5.164 -3.707 .000 -.688 -.209 .000 3.476 5.980 -2.408 .017 -.684 -.067 .000 2.771 4.741 -3.088 .002 -.621 -.136 8.003 -.251 -.325 7.544 .097 .183 7.470 .391 -.162 -1.340 8.534 .082 2.906 7.169 .332 Intell. -.208 1.335 6.256 .175 -.118 .000 6.297 .711 Midwestern Familiarity Sig. 5.342 .098 4.475 t 95.0% Confidence Interval for B Lower Upper Bound Bound .000 3.434 5.690 -3.654 .000 -.791 -.235 Midwestern (Unfamiliar Lecture) Comp. (Constant) Midwestern Familiarity Accent. (Constant) Intell. Midwestern Familiarity (Constant) Midwestern Familiarity Accept. (Constant) Midwestern Familiarity 4.192 .491 -.449 .121 4.728 .633 -.375 .156 3.756 .498 -.379 .123 4.562 .570 -.513 .140 -.309 -.207 -.261 -.305 28 Table 10 (cont’d) Southern (Familiar Lecture) Comp. Unstandardized Coefficients Std. B Error (Constant) 1.899 -.124 .075 Accent. (Constant) 2.873 .553 Southern Familiarity .008 .149 (Constant) 1.771 .246 Southern Familiarity -.109 .066 Accept. (Constant) 2.012 .278 Southern Familiarity -.133 Beta .277 Southern Familiarity Standardized Coefficients .075 Intell. t Sig. 95.0% Confidence Interval for B Lower Upper Bound Bound 6.846 .099 -.273 .024 .000 1.778 3.968 .054 .957 -.288 .304 .000 1.284 2.258 -1.644 .103 -.241 .022 .000 1.461 2.563 -1.767 .080 -.282 .016 7.347 .000 1.815 3.152 -2.181 .031 -.380 -.019 7.019 .000 2.658 4.744 -1.733 .086 -.528 .035 7.230 .000 1.666 2.921 -1.759 .081 -.320 .019 7.187 -.153 -1.660 7.227 -.143 2.448 7.199 .005 1.350 5.192 -.144 .000 .000 1.933 3.401 -2.145 .034 -.413 -.017 Southern (Unfamiliar Lecture) Comp. (Constant) 2.483 .338 Southern Familiarity -.199 .091 Accent. (Constant) 3.701 .527 Southern Familiarity -.247 .142 (Constant) 2.293 .317 Southern Familiarity -.151 .086 Accept. (Constant) 2.667 .371 Southern Familiarity -.215 .100 Intell. -.188 -.150 -.152 -.185 29 Table 10 (cont’d) British (Familiar Lecture) Unstandardized Coefficients B Comp. Standardized Coefficients Std. Error (Constant) 2.908 -.234 .100 Accent. (Constant) 4.270 .560 British Familiarity .132 .161 (Constant) 2.497 .288 British Familiarity -.239 .082 Accept. (Constant) 3.340 .419 British Familiarity -.292 Beta .351 British Familiarity .120 Intell. 95.0% Confidence Interval for B t Sig. Lower Upper Bound Bound 8.293 2.214 3.601 -2.327 .022 -.433 -.035 7.622 .000 3.161 5.378 .820 .414 -.186 .449 8.677 .000 1.927 3.066 -2.898 .004 -.402 -.076 7.975 .000 2.511 4.168 -2.433 .016 -.529 -.055 9.426 .000 2.884 4.416 -3.407 .001 -.597 -.158 7.310 .000 2.922 5.091 1.130 .261 -.133 .488 10.592 .000 2.746 4.008 -4.324 .000 -.575 -.214 9.256 -.201 .000 .000 3.220 4.971 -3.752 .000 -.726 -.225 .072 -.247 -.209 British (Unfamiliar Lecture) Comp. (Constant) 3.650 .387 British Familiarity -.378 .111 Accent. (Constant) 4.006 .548 British Familiarity .177 .157 (Constant) 3.377 .319 British Familiarity -.395 .091 Accept. (Constant) 4.096 .443 British Familiarity -.475 .127 Intell. -.287 .099 -.356 -.314 30 Table 10 (cont’d) Chinese (Familiar Lecture) Unstandardized Coefficients B Comp. Standardized Coefficients 95.0% Confidence Interval for B Beta Lower Upper Bound Bound Std. Error (Constant) 3.031 .332 Chinese Familiarity -.217 .088 Accent. (Constant) 5.971 .526 Chinese Familiarity -.492 .140 (Constant) 2.398 .295 Chinese Familiarity -.181 .079 Accept. (Constant) 4.121 .463 Chinese Familiarity -.306 .123 Intell. t Sig. 9.118 .015 -.392 -.043 .000 4.930 7.012 -3.515 .001 -.768 -.215 .000 1.814 2.982 -2.302 .023 -.336 -.025 .000 3.204 5.037 -2.485 .014 -.550 -.062 10.975 .000 3.737 5.380 -3.773 .000 -.637 -.199 11.776 .000 4.700 6.598 -2.519 .013 -.576 -.069 9.665 .000 3.107 4.707 -3.555 .001 -.597 -.170 9.704 -.214 -2.461 8.891 -.199 3.688 8.118 -.296 2.373 11.347 -.212 .000 .000 3.999 6.047 -2.938 .004 -.679 -.133 Chinese (Unfamiliar Lecture) Comp. (Constant) 4.558 .415 Chinese Familiarity -.418 .111 Accent. (Constant) 5.649 .480 Chinese Familiarity -.322 .128 (Constant) 3.907 .404 Chinese Familiarity -.383 .108 Accept. (Constant) 5.023 .518 Chinese Familiarity -.406 .138 Intell. -.314 -.216 -.298 -.250 31 Table 10 (cont’d) Albanian (Familiar Lecture) Unstandardized Coefficients (Constant) B 2.143 Albanian Familiarity -.020 .158 Accent. (Constant) 3.466 .380 Albanian Familiarity .384 .235 (Constant) 1.850 .226 Albanian Familiarity -.101 .139 Accept. (Constant) 2.863 .271 Albanian Familiarity -.303 Standardized Coefficients Std. Error .256 .168 Comp. Intell. 95.0% Confidence Interval for B Beta -.011 t 8.372 Lower Upper Sig. Bound Bound .000 1.636 2.649 .292 .000 2.715 4.218 1.638 .104 -.080 .848 .000 1.404 2.297 -.725 .470 -.377 .175 10.547 .000 2.326 3.400 -1.806 .073 -.634 .029 7.757 .000 1.640 2.764 .476 .635 -.265 .433 8.390 .000 2.345 3.793 .694 .489 -.292 .608 7.134 .000 1.372 2.425 .580 .563 -.231 .423 7.472 -.157 -.333 8.199 -.064 .898 9.122 .143 -.128 .000 1.652 2.842 .306 .760 -.313 .427 Albanian (Unfamiliar Lecture) Comp. (Constant) 2.202 .284 Albanian Familiarity .084 .176 Accent. (Constant) 3.069 .366 Albanian Familiarity .158 .227 (Constant) 1.899 .266 Albanian Familiarity .096 .165 Accept. (Constant) 2.247 .301 Albanian Familiarity .057 .187 Intell. .042 .061 .051 .027 32 that, across the board for both familiar and unfamiliar lectures, acceptability as a teacher showed, at the very least, a small correlation for those that were noted as significant (which includes 68 of the 70 correlations), with the correlation coefficients ranging from rs=.19 to rs=.86. For accentedness of both lectures, 12 of the 20 ratings were moderately (.3 to .5) to highly (.5 and above) correlated, with correlation coefficients ranging from rs=.32 to rs=.67. Notably, neither lecture for the British speaker had a moderate or stronger correlation (ie. these correlations were weak, at best). Also, for the Albanian and Southern lectures, either one or both of the lectures showed small correlations. For intelligibility, 18 of the 20 ratings were moderately to highly correlated, with the correlation coefficients ranging from rs=.32 to rs=.84. The only lectures that did not show this trend were the unfamiliar Chinese and Albanian lectures. Finally, comparing acceptability ratings between the familiar and unfamiliar lectures, the correlation coefficients were moderate to strong across the board, ranging from rs=.49 to rs=.63. 33 Table 11 Spearman’s Correlations between Participants’ Ratings of Acceptability as a Teacher and Comprehensibility, Intelligibility, and Accentedness Midwestern Acceptability Southern Acceptability British Acceptability Chinese Acceptability Albanian Acceptability Fam. Unfam. Fam. Unfam. Fam. Unfam. Fam. Unfam. Fam. Unfam. Comp. (Fam) Correlation Coefficient .619** .437** .724** .471** .708** .515** .583** .382** .603** .445** Accent. (Fam) Correlation Coefficient .499** .556** .381** .252** .195* 0.161 .337** .204* .339** .260** Intell. (Fam) Correlation Coefficient .726** .531** .711** .530** .579** .470** .458** .322** .537** .475** Accept. (Fam) Correlation Coefficient .493** 1 .556** 1 .632** 1 .598** 1 .507** 1 34 Table 11 (cont’d) Midwestern Acceptability Southern Acceptability British Acceptability Chinese Acceptability Albanian Acceptability Fam. Unfam. Fam. Unfam. Fam. Unfam. Fam. Unfam. Fam. Unfam. Comp. (Unfam) Correlation Coefficient .427** .858** .609** .689** .496** .671** .413** .704** .341** .710** Accent. (Unfam) Correlation Coefficient .378** .673** .361** .556** 0.053 .215* .315** .457** .208* .476** Intell. (Unfam) Correlation Coefficient .473** .839** .586** .762** .562** .685** .295** .634** .282** .569** Accept. (Unfam) Correlation Coefficient .493** 1 .556** 1 .632** 1 .598** 1 .507** 1 Note. *Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). A bolded coefficient means that it is .3 or larger and could be considered a moderate (or stronger) correlation. 35 CHAPTER 5: DISCUSSION In this study, I sought to gain a greater understanding of student attitudes toward NSETs and NNSETs by testing their accent perception ability and by assessing their reactions to a variety of native and non-native accents. The first research question is as follows: 1a) Are ESL learners able to distinguish between native and non-native speakers of English? 1b) If so, could the ESL learners identify the speakers’ accents? Answering these questions gives insight into students’ ability to differentiate between English accents (including native and non-native accents), and also helps to shape the interpretation of the results of the second research question, which focuses on participants’ attitudes about each type of English accent as related to their exposure to a given accent. The second research question is as follows: 2a) Do students who are familiar with a particular accent rate speakers differently in terms of comprehensibility, intelligibility, accentedness, and acceptability as a teacher? 2b) If so, is how they rate related to their attitudes about acceptability as an English teacher? Research Question 1 As shown in Figure 2, NNS participants were usually able (68% of the time) to correctly identify a speaker as being native or non-native. Notably, however, NNS participants were significantly less able than NS participants, who made correct identifications 91% of the time. In looking at the correct identification of accent, however, these percentages dropped; NNS were only able to correctly identify an accent 28% of the time, and NS, 57%. This was different from the findings of Derwing and Munro (1997), who found that NNSs were able to correctly identify an accent 52% of the time. In their study, participants had a forced choice between the four accents, which likely made it easier to identify the accents and could explain the higher percentage of correct identification. Similar to the current study, in Derwing and Munro’s study, 36 100 80 60 Nativeness 40 Accent 20 0 NNS NS Figure 2. Percentage of correct identification of nativeness and accent by NNS and NS participants. a sizable portion of incorrect guesses were given to languages or accents from the same language family (e.g. Chinese and Japanese, Spanish and Italian). This could account for a portion of the incorrect guesses, but what other factors could explain this low rate of correct identification on the part of both NS and NNS participants? One possible explanation is discussed in Moussu’s (2010) study of student attitudes. She conducted a beginning and end-of-the-semester survey of ESL student attitudes toward their NSETs and NNSETs. A review of the latter survey revealed that, even after spending an entire semester with a teacher (and having lived in the U.S. for a semester or more), some students could not correctly identify whether their own teacher was a NSET or a NNSET. The students were clearly familiar with their teacher’s accent (and the American accent spoken in their greater community), so factors other than familiarity with the accent were influencing their identification of the teachers’ native status. Moussu hypothesized that this misidentification could have been related to the teachers’ appearances or presentation styles. In the current study, as in Moussu’s study, many of the students had spent a semester or more in the U.S. and would undoubtedly 37 have spent a great deal of time being exposed to the local accent (in this case, the Midwestern accent). However, Moussu’s explanation would not fit the current situation because the current participants lacked familiarity with the speaker and lacked visual cues that might influence their decisions. Another possible explanation for this difference was given in the Scales et al. (2006) study, in which 62% of the ELL participants indicated that they wanted to acquire a native-like accent, yet only 29% of the students were able to identify an American accent. Similarly, in the current study, 40% of the ELLs indicated wanting to acquire a native-like accent, but only 30% of learners were able to identify the native English speakers (Midwestern, Southern, and British). As Scales et al. wrote, there was a “mismatch between these learners’ own accent goals and their ability to perceive accents” (p. 728). This mismatch between an idealized accent by a student and the lack of ability to identify the real accent could partially account for the significant differences seen between NSs and NNSs participants in the current study. A number of the factors cited in previous research, such as language families, visual cues, teaching style, and a distorted perception of a native accent, could potentially account for the inability of some participants to correctly identify a speakers’ nativeness (whether they are a native or non-native speaker) and their specific accent. I hypothesize, however, that there are two additional factors that could account for this. As shown in Table 8 and Figure 1, there is a clear trend that shows that as proficiency increases, the ability to distinguish between native and non-native accents and the ability to identify specific accents also increases. As a student becomes more proficient in English, their capacity to notice nuances in pronunciation is also likely to increase, which would explain why more proficient ESL speakers are more adept at detecting accent differences. 38 Research Question 2 In the second research question, I asked whether students who are familiar with a particular L1 rate speakers differently in terms of intelligibility, comprehensibility, and accentedness, and acceptability as a teacher. Perhaps the most strongly evidenced claim that I can make based on the correlation results, provided in Table 9, is that familiarity with an accent is significantly related to a student’s judgment of how easy it is to understand a speaker of that accent and a student’s judgment about the speaker’s acceptability as a teacher. Both of these variables (comprehensibility and acceptability as a teacher) were positively correlated with familiarity; in other words, the more familiar a participant was with an accent, the easier it was to understand the speaker, and the more acceptable the speaker was as a teacher. This relationship between familiarity and comprehensibility aligns with results by Gass and Varonis (1984), who clearly showed that familiarity facilitates comprehension. This observation is further supported by Winke, Gass and Myford (2013), who demonstrated that accent familiarity contributes to comprehension. Looking deeper into the relationship between comprehensibility and acceptability as a teacher is also telling. In looking at the correlations between these two factors, every correlation was significant, and was moderate to strong, with a coefficient ranging between .341 and .858. The relationship between intelligibility and acceptability as a teacher patterned similarly, with coefficients ranging between .282 and .839. These positive correlations are not surprising, considering that the easier it is to understand a teacher (or, in terms of intelligibility, the more of a teacher’s speech a student is able to understand), the more likely a student is to rate the teacher as acceptable as a teacher. Similar results were also born out in Scales et al.’s (2006) study, where they found that comprehension was a high priority for students, and the students tended to 39 prefer an accent (or teacher) if the teacher’s accent was easy to understand (p. 725). The results for the linear regressions, provided in Table 10, were also telling. While familiarity with an accent was not a significant predictor of ratings on the four variables in every case, it was overwhelmingly a significant predictor of the four variables for the Midwestern, British, and Chinese accents. Where these were significant predictions, however, the effect sizes were small. In other words, a one-point movement on the familiarity scale would, at most, predict a half-a-point movement on the scales for the four variables. For example, if a student rated themselves as being very familiar with an accent (5), then they would be likely to move 2.5 points down the scale for acceptability as a teacher (toward acceptable, 1). Considering that the four variables are ranked on a 9-point scale, this type of .5 movement seems rather inconsequential. Putting this in classroom terms, if two students of the same proficiency level are assigned to a classroom with a NNSET whose L1 is Chinese, and one student is very familiar with a Chinese accent and the other student is not, their ability to understand the teacher may only be mildly different, as is the case with their attitude toward the acceptability of the teacher, based on accent alone. This classroom scenario is especially interesting in light of Moussu’s (2010) findings: in her study, students’ attitudes toward their NNSETs became more positive over the semester. This scenario is also supported by Gass and Varonis (1984), who found that familiarity with a speaker, with non-native speech in general, and with a non-native accent in particular, all facilitate comprehension. As seen in the current study, comprehension and acceptability as a teacher are positively correlated, meaning that (drawing on Moussu’s findings) the more a student interacts with their NNSET teacher, (and also drawing on Gass and Varonis,) the easier it will be to understand their teacher, and (drawing on the current study,) the more positively the 40 students will view their teacher. While this may be true on paper, we know that ESL classes are not a vacuum devoid of other factors that would influence a student’s attitude toward their teacher; yet the evidence is nonetheless compelling in suggesting that the current study showed that students’ judgments of the strength of a teacher’s accent influences the students’ view on whether they find the teacher acceptable as a teacher. As shown in the correlations in Table 11, accentedness and acceptability as a teacher correlated (small to moderate) with the exception of the British accent. However, as discussed in the previous paragraph, even if a teacher was viewed as having a moderate to strong accent, their students may gradually understand more and more of the teacher’s speech, and the students will understand the teacher with greater ease over time. Alford and Strother (1990), who asked NS and NNS participants to listen to speech samples and rate them on bipolar adjectives scales (e.g. professional/non-professional, intelligent/unintelligent, etc.), found that ELLs were able to detect differences in accents. These were extrapolated from differences in their expressed attitudes toward the speakers of different accents. In other words, students used different adjectives to describe different accents, and the researchers interpreted this as evidence of the students being able to differentiate between accents. These results also clearly showed that they did have different attitudes toward accented English. Though the ELL participants in the current study were largely unable to overtly decipher a speakers’ nativeness and accent, like Alford and Strother’s findings, the students’ varied ratings of acceptability as a teacher for different speakers does evidence differences in attitudes based on accent. Unfamiliar Accents In general, there were some inconsistencies in the data that mostly revolved around the 41 Albanian accent and the Southern accent. One explanation for the variations found in relation to these accents is that, for Albanian, the vast majority of the participants had no exposure to Albanian and could not place Albania on a map. This was true for NS and NNS participants alike. Because this language is relatively novel to the participants, I assume that some participants were blindly guessing accents, or that some were making educated guesses, drawing from other semantically related or phonologically related languages (i.e. Indo-European languages, particularly Balto-Slavic or Germanic languages). This could explain the large variance in results for the Albanian accent. For the southern accent, during data collection, many of the NNS students asked what a southern accent was (what it sounded like), or if it was the same as a South American accent. This confusion or misunderstanding could likely have been experienced by many more participants, and thus could be a possible explanation for the variation in the Southern accent data. Whereas ratings for Albanian and Southern varied widely in the data, the results for Midwestern, British, and Chinese were undoubtedly more consistent. This consistency could be attributed to a greater general knowledge or general true familiarity with these cultures and accents, whereas ratings for Albanian and Southern could have been compromised by the misunderstanding and general unawareness that was noted during collection. Another important note about familiarity is that participants rated themselves as being most familiar with the British accent. This could be explained by the largest NNS groups represented in their participant population, Chinese and Saudi. The educational systems in these countries largely train their students in Received Pronunciation, also know as standard British English, which accounts for the high reporting of British familiarity. I would contend, however, that many of these students only have a superficial exposure to Received Pronunciation. For example, students in China 42 often are strong readers, but lack oral communicative competence due to the lack of focus on speaking and listening in the Chinese classroom. For both of these groups (Chinese and Saudi), the majority have never studied abroad, or even traveled outside of their countries, and have very limited experience with spoken English. I would, therefore, assume that students have overreported their familiarity with the British accent, and have perhaps underreported their familiarity with the Midwestern accent (especially since the majority of the NNSs had lived in the Midwest for 6 months or more at the time of data collection). For reference, a graph of the composite self-reported ratings is given in Figure 8, Appendix F. 5.00 4.00 Midwestern 3.00 Southern British 2.00 Chinese Albanian 1.00 Average 0.00 Familiar Unfamliar Familiar Accentedness Unfamliar Acceptability Figure 3. Composite ratings of accentedness and acceptability as a teacher by all participants. Note. 1= Acceptable as an English teacher/No accent. 9= Not acceptable as an English teacher/Strong Accent. 43 Inconsistences aside, participants’ ratings have clearly shown that accentedness is not synonymous with being an unacceptable teacher. Ratings for accentedness were higher than ratings for acceptability as a teacher, shown in Figure 3. On average, participant ratings for accentedness of the familiar and unfamiliar lectures were 3.74 and 3.70, respectively (where 1 is “no accent” and 9 is “very strong accent”), and ratings for acceptability as a teacher were 2.23 and 2.60, respectively (where 1 is “acceptable” and 9 is “not acceptable”). The ratings for acceptability were an average of 1.5 and 1.0 points lower than participants’ ratings of accentedness. These ratings, relative to one another, reveal a general acceptance and positive attitude toward teachers with any type of accent. Students’ acceptance of accented teachers provides evidence that challenges the assumptions of language center administrators, such as in Moussu’s (2010) study, which reported that one of the major factors in hiring English language teachers is their native speaker status. Many administrators claim that they continue in these hiring practices because of “native speaker demand” on the part of students. Results in this study, however, provide insight into student attitudes toward NNSETs, based on their accents, which shows that student attitudes toward NNSETs are generally positive. Assuming that students do have a positive attitude toward NNSETs, Clarke and Garrett (2004) provided further evidence in support of Moussu’s findings; Clarke and Garrett found that listeners quickly adapt to new accents, and applying this to the current study, students will adapt to the accents of their teachers. In their study, they investigated how long it takes NSs to adapt to foreign-accented English speech, and found that NSs can adapt to accented speech in as little as one minute. Presumably, the speech processing load placed on NNSs as they listen to English would be much heavier, and would therefore take NNSs longer to adapt to accented speech. This, 44 however, provides added insight into Moussu’s finding, as discussed earlier, and could be a possible explanation of why students’ attitudes toward their NNSETs became more positive as students were exposed to their teachers’ speech over the semester. This process is also supported by the current research, which shows that familiarity, comprehensibility, and students’ attitudes about a teacher’ acceptability are all related. 45 CHAPTER 6: CONCLUSION In this study, I have sought to extend previous research on accentedness, native versus non-native speaking English teachers, and ESL student attitudes. As previously outlined, ESL students were usually able to distinguish between native and non-native speakers. However, they were generally not able to identify a speaker’s accent. Some factors that might explain this inability to identify speakers are a confusion of languages in a language family, misconceptions concerning native accents, and a student’s proficiency level. Results further uncovered that familiarity is correlated with comprehensibility and acceptability as a teacher. In this study, comprehensibility, intelligibility, and acceptability were highly correlated with one another. For truly familiar accents, familiarity was a significant predictor of participant ratings on the four variables (comprehensibility, intelligibility, accentedness, and acceptability as a teacher), though the predicted changes in ratings were small. The current research shows that students had generally positive attitudes toward NNSETS; based on previous research, students’ attitudes toward their NNSETs only become more positive over time, and students may quickly perceptually adapt to their accented teachers, if they are willing to do so (Derwing, Rossiter, & Munro, 2002). Finally, this research has shown that accentedness does not, in the minds of students, translate to unacceptability as a teacher. Limitations Like any study, this research has its limitations. One is that the survey was long. The fatigue that students experienced from answering the same questions again and again could have resulted in a lack of concentration. A second limitation is that the audio clips embedded in the survey were not limited to a single play. This means that students could have listened to any audio file as many times as they wanted (though they were instructed to listen only once), which 46 could have given some students an unequal advantage. Finally, during one round of data collection, the participants experienced technical difficulties with the survey, which frustrated the participants. This also could have influenced how they approached the survey and the answers that they gave. Implications I believe that the implications of these findings are most pointed for the hiring practices (or in the case of Arizona, firing practices) of schools and language centers; as the results have shown, student attitudes toward NNSETs were generally positive, and even when students indicated that a teacher had a pronounced accent, the students still demonstrated a positive attitude toward the teachers by rating them as acceptable. This shows that, from a student’s perspective, other factors may more heavily influence a student’s attitude toward a teacher and 2 whether or not a teacher is acceptable. As TESOL also argued, teachers should be hired on the basis of their professional credentials, such as professionalism, teaching experience, and English proficiency as a whole. Educators should also consider what is best for students. Given the nature of diversity at any university, the rapid growth of EIL and ELF, the overpowering number of non-native English speakers, and the purposes for which many students are learning English, it is in a university or a program’s best interest to expose students to a wide range of global Englishes. As shown in this study, as in many previous studies, a familiarity with an accent leads to greater comprehension or intelligibility of language spoken with that accent (Gass & Varonis, 1984; Winke, Gass, & Myford, 2013; Harding, 2012; Derwing & Munro, 1997) and this would benefit students as they encounter various native and non-native speakers of English throughout the span 2 Teachers of English to Speakers of Other Languages, Inc. For more information, see www.tesol.org. 47 of their academic, personal, and professional lives. Directions for Future Studies Even though the results of this study show quantitatively that students’ attitudes towards their teachers are not largely based on accent, qualitative research needs to be done to corroborate this finding. In addition, though accent adaptability research has been done with NSs, further work that investigates the amount of time it takes ELLs to adapt to a new accent could provide insight into accent dynamics. This type of information could corroborate evidence based on comprehensibility, intelligibility, and attitudes toward encountering new accents. 48 APPENDICES 49 Appendix A Table 12 Participants by Nationality NNS of English NS of English Total NNS of English NS of English Total Male Female Total Male Female Total Male Female Total Male Female Total Male Female Total Male Female Total Chinese Emirati 42 1 14 0 56 1 0 0 0 0 0 0 42 1 14 0 56 1 Indonesian 1 0 1 0 0 0 1 0 1 Israeli Saudi Thai 10 7 17 0 0 0 10 7 17 German 0 1 1 0 0 0 0 1 1 1 0 1 0 0 0 1 0 1 Korean 1 2 3 0 0 0 1 2 3 4 3 7 0 0 0 4 3 7 Mexican American Spanish 1 0 0 1 0 1 2 0 1 0 5 0 0 33 0 0 38 0 1 5 0 1 33 1 2 38 1 Japanese Turkish Canadian Libyan 0 1 0 0 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 1 1 1 1 1 1 50 Total 62 32 94 5 33 38 67 65 132 Appendix B Familiar Topic Transcript: Orientation on Classroom Expectations So, what do your teachers expect from you in your classes? The teachers expect that students will attend class. They also expect that students will participate actively in class. This means interacting with classmates, asking questions, and answering with critical thinking. The teachers also expect that you will do your own work. This means you won’t use the Internet or have others help you to do the homework you were assigned to do. Because you are expected to do your own work, you will have to spend a lot of time and effort on your homework and assignments. You must take the time outside of class to work on improving your language skills. The English Language Center wants you to be successful! We, your teachers and advisors, will help you, but you are also responsible for your success! 51 Appendix C Unfamiliar Topic Transcript: How to Make Pottery Now that you’ve kneaded it out and gotten all of the bubbles out of the clay, you can now move to the wheel where we’ll actually throw the clay. Okay, that doesn’t mean you’re throwing the clay at each other. It means you’re putting it on a spinning wheel and it will take the shape and form of what you want by using your hands as the tools to guide them. Okay, so you’ll place the clay onto the clay pot, or the thrower, and turn it on. And as it turns, it’s going to spin around, and it’s all controlled by your foot, so however fast you want the clay to spin, it’s all controlled by a foot pedal. The faster, the more pressure you’ll put on the pedal, the slower, the less pressure, okay? (Adapted from Sueyoshi & Haridson (2005)) 52 Appendix D Dictation Task Sentences 1. Since I prefer rice, I never eat pasta. 2. Because he cooks dinner, she often washes dishes. 3. Although they like chicken, they sometimes order steak. 4. When we have salad, we usually finish everything. 5. If she has cake, she always drinks milk. 6. Though you love desserts, you frequently refuse cookies. 7. Although he likes milk, he rarely drinks milkshakes. 8. While we make breakfast, they usually check email. 9. Unless she finds blueberries, she always uses strawberries. 10. Because she drinks coffee, she never chooses tea. 11. Even though I eat fish, I never eat sushi. 12. If they buy bread, they seldom get wheat. 13. Whenever you bake muffins, you normally burn them. 14. Wherever they sell salt, they always sell pepper. 15. As we eat lunch, we occasionally watch TV. 16. Until he has breakfast, he seldom does work. 53 Appendix E Sample Listening Page Sample Listening Tasks from Survey 54 55 Appendix F Additional Figures 5.00 4.50 4.00 3.50 3.00 Familiar 2.50 Unfamliar 2.00 1.50 1.00 Midwestern Southern British Chinese Albanian Figure 4. Composite ratings of comprehensibility by all participants. Note. 1= Very easy to understand. 9=Very difficult to understand. 5.00 4.50 4.00 3.50 Familiar 3.00 Unfamliar 2.50 2.00 1.50 1.00 Midwestern Southern British Chinese Figure 5. Composite ratings of accentedness by all participants. Note. 1= No accent. 9= Very strong accent. 56 Albanian 5.00 4.50 4.00 3.50 Familiar 3.00 Unfamliar 2.50 2.00 1.50 1.00 Midwestern Southern British Chinese Albanian Figure 6. Composite ratings of intelligibility by all participants. Note. 1= I understood everything. 9= I didn’t understand anything. 5.00 4.50 4.00 3.50 Familiar 3.00 Unfamliar 2.50 2.00 1.50 1.00 Midwestern Southern British Chinese Albanian Figure 7. Composite ratings of acceptability as a teacher by all participants. Note. 1= Acceptable as an English teacher. 9= Not acceptable as an English teacher. 57 4.0000 3.5000 Midwestern Familiarity 3.0000 Southern Familiarity 2.5000 British Familiarity Chinese Familiarity 2.0000 Albanian Familiarity 1.5000 1.0000 1   Figure 8. Composite ratings of familiarity with accents by all participants. Note. 1= Not familiar. 5= Very familiar. 58 REFERENCES 59 REFERENCES Alford, R. L., & Strother, J. B. (1990). Attitudes of native and nonnative speakers toward selected regional accents of U.S. English. TESOL Quarterly, 24(3), 479-495. Alptekin, C. (2002). Towards intercultural communicative competence in ELT. ELT Journal, 56(1), 57-64. Blum, A., & Johnson, E. J. (2012). Reading repression: Textualizing the linguistic marginalization of nonnative English-speaking teachers in Arizona. Journal of Language, Identity, and Education, 11(3), 167-184. Canagarajah, A. S. (2005). Reclaiming the Local in Language Policy and Practice. Mahwah, NJ: Lawrence Erlbaum Associates. Derwing, T. M., & Munro, M. J. (1997). Accent, intelligibility, and comprehensibility: Evidence from four L1s. Studies in Second Language Acquisition, 19(1), 1-16. Derwing, T. M., & Munro, M. J. (2005). Second language accent and pronunciation teaching: A research-based approach. TESOL Quarterly, 39(3), 379-397. Derwing, T. M., Rossiter, M. J., & Munro, M. J. (2002). Teaching native speakers to listen to foreign-accented speech. Journal of Multilingual and Multicultural Development, 23(4), 245-259. Eisenstein, M., & Verdi, G. (1985). The intelligibility of social dialects for working-class adult learners of English. Language Learning, 35(2), 287-298. Field, A. (2009). Discovering Statistics Using SPSS (Third Edition). London: SAGE Publications Ltd. Clarke, C. M., & Garrett, M. F. (2004). Rapid adaptation to foreign-accented English. Journal of the Acoustical Society of America, 116(6), 3647-3658. Gass, S., & Varonis, E. M. (1984). The effect of familiarity on the comprehensibility of nonnative speech. Language Learning, 34(1), 65-89. Harding, L. (2011). Accent and listening assessment: A validation study of the use of speakers with L2 accents in academic English listening assessment. Frankfurt: Peter Lang. Harding, L. (2012). Accent, listening assessment and the potential for a shared-L1 advantage: A DIF perspective. Language Testing, 29(2), 163-180. 60 Hertel, T. J., & Sunderman, G. (2009). Student attitudes toward native and non-native language instructors. Foreign Language Annals, 42(3), 468-482. Holliday, A. (2008). Standards of English and politics of inclusion. Language Teaching, 41(1), 119-130. Llurda, E. (2004). Non-native-speaker teachers and English as an international language. International Journal of Applied Linguistics, 14(3), 314-323. Marian, V., Blumenfeld, H. K., & Boukrina, O. V. (2008). Sensitivity to phonological similarity within and across languages. Journal of Psycholinguistic Research, 37(3), 141-170. Moussu, L. (2010). Influence of teacher-contact time and other variables on ESL students' attitudes towards native- and nonnative-English-speaking teachers. TESOL Quarterly, 44(4), 746-768. Munro, M. J., Derwing, T. M., & Morton, S. L. (2006). The mutual intelligibility of L2 speech. Studies in Second Language Acquisition, 28(1), 111-131. Park, G. (2012). "I am never afraid of being recognized as an NNES”: One teacher's journey in claiming and embracing her nonnative-speaker identity. TESOL Quarterly, 46(1), 127151. Philipson, R. (1992) Linguistic Imperialism. Oxford, UK: Oxford University Press. Selvi, A. F. (2011). The non-native speaker teacher. ELT Journal, 65(2), 187-189. Scales, J., Wennerstrom, A., Richard, D., & Wu, S. H. (2006). Language learners' perceptions of accent. TESOL Quarterly, 40(4), 715-738. Sueyoshi, A., & Hardison, D. M. (2005). The role of gestures and facial cues in second language listening comprehension. Language Learning, 55(4), 661-699. TESOL International Association. (2006). Position statement against discrimination of nonnative speakers of English in the field of TESOL. Retrieved from http://www.tesol.org/docs/pdf/5889.pdf?sfvrsn=2 Thoms, J., Liao, J., & Szustak, A. (2005). The Use of L1 in an L2 On-Line Chat Activity. The Canadian Modern Language Review / La revue canadienne des langues vivantes, 62(1), 161-182. Timmis, I. (2002). Native-speaker norms and international English: A classroom view. ELT Journal, 56(3), 240-249. Todd, R. W., & Pojanapunya, P. (2009). Implicit attitudes towards native and non-native speaker teachers. System, 37(1), 23-33. 61 U.S. Department of Education. (2002). PUBLIC LAW 107–110—JAN. 8, 2002. Retrieved from http://www.ed.gov/policy/elsec/leg/esea02/index.html Winke, P., Gass, S., & Myford, C. (2013). Raters' L2 background as a potential source of bias in rating oral performance. Language Testing, 30(2), 231-252. 62