Anxiety and Willingness to Communicate in Language Learning: A Case Study

This case study attempts to find out the role of individual differences (IDs) in language learning of an advanced Bangladeshi EFL learner who aims to study at a graduate level program in Education at a North American university. A two-phase structured questionnaire interview has been conducted in which phase one comprises 40 open-ended questions with a general structure that attempts to find out the various personal characteristics of the interviewee, namely ‘personal profile,’ ‘learning history,’ linguistic background, etc. In phase two, there are two sub-phases, and the questionnaire in the first sub-phase focuses on four IDs namely ‘personality,’ ‘learning style,’ ‘anxiety,’ and ‘willingness to communicate,’ with a general research question concerning the effects of these IDs on the interviewee. In the second sub-phase, the researcher narrows down the focus of the research question to the two most important IDs influencing the interviewee’s learning of English, namely ‘anxiety’ and ‘willingness to communicate.’ For the second sub-phase of phase two, the researcher uses four quantitative measurement scales, two measuring the interviewee’s anxiety (adapted from Horwitz and Horwitz, 1986) and the other two measuring her willingness to communicate (adapted from Macintyre, 2001) both inside and outside of classrooms. The results show that the interviewee feels very anxious and addled in the classroom whereas she feels quite the opposite outside the classroom. Naturally, her willingness to communicate in the classroom is very low as


Introduction
This case study attempts to find out the role of individual differences (IDs) in language learning of an advanced Bangladeshi EFL learner who aims to study at a graduate level program in Education at a North American university. The learner will be addressed as Masha (a pseudonym) who is an exchange visitor on a J-visa living in downtown Washington DC with her husband. A two-phase interview was conducted in which phase one comprises a generally structured open-ended questionnaire with 40 questions that attempts to find out "enduring personal characteristics that are assumed to apply to everybody and which people differ by degree" (Dornyei, 2005, p. 644) in relation to the roles they played in Masha's learning of English as a foreign language. Phase two narrows the general focus to more specific ones analyzing the qualitative responses of the phase one interview. In other words, for the phase two interview, only those IDs were chosen which were "consistently shown to correlate strongly with L2 achievement -to a degree that no other SLA variables match" (ibid, 643). For phase two, the researcher used four quantitative measurement scales, two measuring the interviewee's anxiety (adapted from Horwitz and Horwitz, 1986) and the other two measuring her willingness to communicate (adapted from MacIntyre, 2001) both inside and outside of classrooms.
Phase one of the interview took place at 4pm on April 8, 2013 at a local restaurant in Washington DC. The researcher came to know about Masha through a friend of his and contacted her via email first to express his interest in interviewing her. He also added that the interview would not be timed so she could take her time in answering the questions but that it would be audio recorded. As Masha agreed, the date was set for the interview. The interview lasted 1.5 hours and was audio recorded in its entirety. The second phase of the interview (also face-to-face) took place on April 14, but was not audio-recorded since that was an attitude questionnaire in which Masha had to choose from a range of responses.

L2 learning history and the linguistic environment
The interviewee is a 25-year old female from Bangladesh, a country that falls under the outer circle in the three concentric circles as defined by Kachru (1985) in which "English is not the native tongue, but is important for historical reasons and plays a part in the nation's institutions, either as an official language or otherwise." She completed her B. A. in English Language and Literature from a private university in Dhaka, Bangladesh. On completion of her B. A. degree, she came to the United States in August 2011 as a J-2 dependent (a non-immigrant visa) with her husband. She works for a coffee shop in downtown DC where she has to interact with a lot of native and non-native speakers of English every day.
In part I of the interview (Personal Profile, Learning History, and Linguistic Background), she identified herself as an advanced user of English as aptly reflected in her IBT TOEFL score (105 out of 120). She spoke Bangla as her native language and also as a medium of learning for the entire pre-tertiary education for 12 years. In other words, the medium of instruction for all the content areas in her school was Bangla, and she had studied English as a subject like other content area courses (e.g., History and Geography) from the very start of her school at the age of five. She was first exposed to an English-speaking learning environment when she enrolled in the B.A. in English Language and Literature program at a university in Dhaka. Till then, she had hardly had any opportunity to practice her English (speaking) beyond classrooms. Moreover, even within the classroom, there was extensive use of L1, which again limited the scope for speaking practice.
Shedding light on her academic learning environment for English, she mentioned that she had four one-hour English classes in the primary level and five 1.5-hour English classes at the secondary level. The English classes were mostly limited to reading, writing, and grammar throughout as there was no systematic teaching of listening and speaking. Masha identified the reason for this discriminatory focus as the result of the exam-oriented education system. Since the tests/exams were limited to only reading and writing, teaching was largely geared towards achieving these skills and thereby preparing the learners solely for the test. Listening and speaking were expected to be learned as a by-product of classroom discussion.
The textbooks were largely reading-based followed by comprehension questions. Usually the writing tasks used to be integrated into the readings as extension activities. A typical example of her secondary English textbook activity would be a reading comprehension, say, on family planning followed by comprehension questions and controlled writing activities like filling in the blanks with missing information. She was then instructed to write a paragraph on selected issues of the reading passage in question. Also, there were literary texts (mostly short stories and poems) to stimulate "creative thinking." She added that there was not much variation in the class work and homework since both emphasized mostly similar reading and writing activities. In her words, "the homework was mere extensions of the class work." However, despite her limited exposure to English in the academia, she used to watch English cartoons since Grade 1 on various satellite channels in Dhaka. As she reached her sixth grade, she started watching Disney movies regularly. She said that natural exposure to English through visual media was extremely helpful in getting her ears attuned to understanding both British and American English. She Mohammad Mahmudul Haque | Anxiety and Willingness to Communicate in Language Learning: A Case Study said that at the beginning, "it was just a lot of listening, with almost zero speaking" which, according to her, helped to make a solid linguistic as well as pragmatic foundation for her future learning of English.

Relev languag learning
As Ellis (2008) pointed out, the factors in the study of individual differences "overlap in vague and indeterminate ways,"and it is sometimes impossible to figure out the exact roles of a given factor in relation to the other factors (p. 644). As an attempt to systematize the study of the IDs, Ellis (2008) divided them in terms of 'abilities' (cognitive capabilities for language learning), 'propensities' (cognitive and affective qualities related to language learning), 'learner cognitions about L2 learning,' and 'learner actions.' However, Dornyei (cited in Ellis, 2008) pointed out that it may not always be easy to decide if an ID constitutes 'ability' or a 'propensity.' That is why it may be more sensible, as Ellis (2008) commented, to treat them separately. Following is an individual discussion of the four ID factors, namely learning style, personality, anxiety, and willingness to communicate.

Learning Style
According to Keefe (1979a), learning styles refer to the characteristics that indicate "how learners perceive, interact with and respond to the learning environment" (p. 4). He also describes it as a "consistent way of functioning" which reflects the "underlying causes of behavior" (p. 4). As for measuring the learning style, there are research instruments that have been borrowed from general psychology, for example, Dunn et al.'s (1991) Productivity Environmental Preference Survey, while others have been specifically designed to investigate language learners, for example, Reid's (1987) Perceptual Learning Style Questionnaire (Ellis 2012, pp. 667-668). Dunn et al.'s (1991) Productivity Environmental Preferences Survey measures learning styles in "four areas: a) preferences for environmental stimuli, b) quality of emotional stimuli, c) orientation towards sociological stimuli, and d) preferences related to physical stimuli" (cited in Ellis, 2012, p. 668). It is designed to showcase preferences pertaining to both personality and learning style. Bailey, Onwuegbuzie, and Daley (2000) employed this questionnaire among 100 French and Spanish first and second semester students studying at a US university. The study reveals that higher achievers prefer informal classroom design as opposed to "receiv[ing] information via kinesthetic mode" (p. 115).
On the other hand, Reid's (1987) perceptual learning styles questionnaire was created based on four perceptual learning styles (visual learning, auditory learning, kinesthetic learning, and tactile learning) and two social learning styles (group preferences, individual preferences). She administered the survey on learners of ant personal characteristics influencing CROSSINGS : VOL. 6, 2015 different language backgrounds and found that learners had a general preference for kinesthetic and tactile learning with a negative attitude (both native and nonnative learners) towards group work. However, a modified version of Reid's questionnaire, conducted by Wintergerst, DeCapua and Verna (2003), revealed a contradictory result (learners preferred group work over individual work). The researchers consider time gap and various social learning styles to be responsible for the contradictory results. Ellis (2012, p. 671)) concludes that since learners' preference towards L2 learning approach varies to a great extent, it is almost impossible to choose the best one. He mentions 'flexibility' to be a plausible reason for learners' success, but it lacks real evidence. He adds that it is unlikely that progress will happen in this respect unless and until researchers know what it is that they want to measure.

Personality
Pervin and John (2001) define 'personality' as an expression of a consistency in the pattern of "feeling, thinking and behaving" (cited in Ellis, 2012, p. 672). Both teachers and learners consider it to be a very important aspect of language learning as evidenced in Griffiths' (1991b) and Naiman et al. 's (1978) study respectively, which shows that teachers consider it to be an important aspect in L2 language learning. As for measuring personality, different language specific questionnaires have been developed to determine "dimensions of personality" like tolerance of ambiguity or risk taking. "Eysenck Personality Questionnaire" or the "Myers Briggs Type indicators" are examples of two types of general questionnaires to identify a learner's personality.
Among the many dimensions of personality, the most notable is extraversion/introversion. Two hypotheses have been made for correlating extraversion/ introversion with L2 learning. The first hypothesis, the one which has been widely researched, states that extrovert learners acquire basic interpersonal communication skills better due to the opportunity for more practice leading to a bigger chance of success. The second hypothesis states that cognitive academic language develops more for introvert learners due to their time dedication towards academic reading and writing. Strong's (1983) review of 12 studies revealed that extroversion was at a point of advantage for language acquisition which supports the first hypothesis. However, Dörnyei and Kormos' (2000) study failed to find a positive correlation between language acquisition and extraversion. Dewaele and Furnham (1999), on the other hand, concluded from their study that extrovert learners though may be fluent in both L1 and L2 but it is not necessary for them to be accurate (cited in Ellis,p. 674). Much research has been conducted to prove the validity of the second hypothesis. However, most of those (Busch, 1982;Carell, Prince, and Astika, 1996;Ehrman and Oxford, 1995) have found either an insignificant or a weak relationship (cited in Ellis,p. 674) between extraversion/ introversion and academic proficiency.
The Big Five Model, an important theory of personality in psychology, has five dimension of personality (openness to experience, conscientiousness, extraversionintroversion, agreeableness, and neuroticism-emotional stability) (Big Five Personality Test, June 05, 2015). This model has been modified and used by Verhoeven and Vermeer (2002) where it was found that children showing interest in belonging and identifying with their target language speaking peers achieve success in learning (p. 373).
Recent studies have seen more success than that of the previous ones in correlating language with personality traits. However, there are limitations like situational dependence of personality, variables like attitude, motivation, situational anxiety influencing the effect of personality, and methodological deficiency.

Anxiety
The learning situation affects the learning process of both naturalistic and classroom learners. Language, according to Pavlenko (2006b), is an "inherently emotional affair" (cited in Ellis, 2012, p.691). Researchers like Horwitz, and Young (1991), Arnold (1999), and Young (1999) (cited in Ellis, 2012, p. 691) believe anxiety to be SLA's most noticeable affective aspect. The three types of anxiety that are present are -trait anxiety, state anxiety, and situation specific anxiety. Trait anxiety is the tendency of being anxious all the time, whereas state anxiety is what a learner feels in a particular moment as a reaction to a certain situation (Spielberger, 1983, cited in Ellis, 2012. Finally, situation specific anxiety is the apprehension a learner feels in situations like examination, public speaking, etc. (Ellis, 2012, p. 691).
Among the many techniques of measuring the correlation between anxiety and achievement diary data, questionnaire responses correlating to achievement, experiments, report of learners' response to language learning condition are mentionable. Spielmann and Radnofsky's (2001) ethnographic studies containing "rich description of learners' reactions to their learning situations" address "three issues: 1) the source of language anxiety, 2) the nature of the relationship between language anxiety and language learning, and 3) how anxiety affects learning" (cited in Ellis, 2012, p. 692).
Another notable study, the diary study by Bailey (1983, cited in Ellis, 2012 has the analysis of 11 learners where she found that when learners find themselves more proficient than their peers, their anxiety decreases. She mentioned tests, teacher-student relationships, etc. to be some sources of anxiety. Ellis and Rathbone (1987) found from their study that teachers' questions can be another source of anxiety. However, it is really hard to identify the sources of anxiety because Horwitz (2001), from the review of her studies, revealed that in most of the cases the tasks that were considered "comfortable" by some were considered to be "stressful" by others(p.118).

CROSSINGS : VOL. 6, 2015
Among the many instruments of measuring anxiety level Horwitz, Horwitz and Cope's (1986) Foreign Language Classroom Anxiety Scale is notable. Their 33itemed questionnaire tries to relate to the three sources of anxiety (communication apprehension, test, and fear of negative evaluation) for speaking and listening in L2 acquisition (Ellis, 2012, p. 693). On the other hand, Cheng, Horwitz, and Schallert (1999) developed a questionnaire to identify the relationship of reading and writing anxiety with general language anxiety (Ellis, 2012, p. 693).
Language learning and anxiety are related to each other and three positions have been identified regarding the relationship between anxiety and language learning. The first position, anxiety facilitates learning, was supported by Eysenck (1979) who said that "low level anxiety" motivates learners to give more effort (Ellis, 2012, p. 694). MacIntyre (2002), Chastain (1975), Kleinmann (1978) assumed a similar position in their studies. The second position, anxiety, has a negative impact on language learning, and was supported by Chastain (1975)and Horwitz (1986) who found a negative correlation between anxiety and grades or marks. Ely (1986a) found that learners having high anxiety levels took less risk.That is, their motivation was negatively affected (Ellis, 2012, p. 694). The third position, language anxiety, the result of difficulty with learning rather than its cause, was supported by a series of studies conducted by Sparks, Ganschow, and Javorsky (2000) which claims that anxiety regarding L2 learning is a result of language difficulties faced by the learners (Ellis, 2012, p. 695).
An important model on anxiety and the language learning process was proposed by MacIntyre and Gradner (1991a), known as the developmental model, which tries to relate learners' developmental stage and situation specific learning experiences with learner anxiety. This model justifies Parkinson and Howell-Richardson's (1990) diary studies which revealed that anxiety develops because of learners' "bad learning experience" (Ellis, 2012, p. 695). Elkhafaifi's (2005 study showed that beginner learners had more listening anxiety than intermediate or advanced learners as "anxiety reduces as they develop" (Ellis, 2012, p.695).
MacIntyre and Gardner (1991b) developed their model based on their study in which they used video cameras to observe anxiety levels in the three stages (input stage, processing stage, and output stage). They found the anxiety level to be highest just after introducing the video camera. However, gradually, learners overcame the anxiety and compensated it by increasing performance (Ellis, 2012, p. 696).

Willingness to communicate
Willingness to communicate (WTC), in other words, "the intention to initiate communication given a choice" (MacIntyre, Baker, Clement, and Conrad, 2001) is considered to be a variable that is "determined by other variables" (cited in Ellis, 2012, p. 697). The factors influencing WTC are situation specific (Ellis, 2012, p. 697).
One of the prominent studies on this variable was done by Yashima (2002) who

Mohammad Mahmudul Haque | Anxiety and Willingness to Communicate in Language Learning: A Case Study
illustrated in her study the necessity of knowing what learning a language means in a context before imposing a definition/model developed elsewhere. In other words, the definition and attitudes of language learning should be bottom up as opposed to top-down. Yashima's (2000) study reveals how international posture (general attitude of the international community) figures both as a direct and indirect variable depending on the context (p.62). However, Kang's (2005) study on four Korean adult males learning English where they were paired with native speakers to communicate freely showed no direct relationship between international relationship and WTC (Ellis, 2012, p. 698).
As far as learning the language to communicate is concerned, Ellis (2012) related WTC to CLT (Communicative Language Teaching) as he said that learners who are willing to communicate are benefitted from CLT whereas learners who are not willing to communicate learn better from more traditional approaches. Comparing the difference of WTC inside and outside of classes, MacIntyre et al. in their study found WTC to be a "stable, trait like factor" (cited in Ellis, 2012, p. 698), the same both inside and outside the classroom for Anglophone learners of L2 French in Canada.
Adding an interesting dimension, Dörnyei and Kormos' (2000) study revealed a relation between WTC and attitude towards the task. They found that learners with a positive attitude towards the task had more willingness to communicate whereas the correlation was close to zero when the learners had a negative attitude towards the task (Ellis, 2012, p. 698).

The current study
To develop a general understanding of the roles these IDs played in Masha's acquisition of English, the current two-phase study adopts a concatenated approach, or a research-then-theory approach through a structured questionnaire interview with a general research question, "To what extent do learning styles, personality, anxiety, and willingness to communicate account for Masha's L2 achievement?" The questionnaire has 33 questions with separate sections on each of the IDs mentioned above. Analyzing the interviewee's responses to these questions, the researcher narrowed down the focus of the research question to the two most important IDs influencing Masha's learning of English, namely 'anxiety,' and 'willingness to communicate.' As the researcher analyzes the raw data collected through the questionnaires, he found that Masha was mostly an autonomous learner who learned best by working on her own. She found classroom learning boring and frustrating whereas she learned unconsciously through TV or movies and found it fun. She was very conscious about making errors in class, whereas she did not care much about the errors she made outside classes during her spontaneous speech. She also learned through application, so, naturally, she did not find the grammar-focused classroom instructions very engaging, frequently getting distracted. Moreover, being quite a CROSSINGS : VOL. 6, 2015 talkative person outside class, Masha was quite reserved in the classrooms. Finally, although unwilling to communicate in the classroom, Masha was quite enthusiastic about out-of-class communication.
Summarizing her responses from the general questionnaire, the researcher could detect a very different behavioral pattern in Masha's attitude inside and outside of classrooms. Based on this finding, he decided to further explore Masha's 'anxiety' and 'willingness to communicate' in and out of the classroom.The research question was revised to the following:

Ÿ To what extent do 'anxiety' and 'willingness to communicate' account for
Masha's language development in and out of class?
The researcher adapted each of the scales into two parts: one part investigating Masha's behavior in the classroom and the other part outside of class. Each of the parts had 10 statements to specifically find out Masha's behavioral differences in and out of class (please see the fully developed scales as appendix B, C, D, and E). In designing both the scales, the same items (for example, her level of confidence in and out of class) were used for the researcher to be able to compare the findings with each other. The researcher adapted a well-organized pattern for the interviewee to feel comfortable in answering the questions. The scale survey results both in 'anxiety' and 'willingness to communicate' strongly correlate with the findings of the general questionnaire interview in that Masha's anxiety level is very high when she learns in class and extremely low (in fact the lowest) during her outside class interactions. Similar results were found with her 'willingness to communicate': very low in class yet, very high outside of classes.

Anxiety
Masha's response to classroom anxiety is similar to one of the students, namely Monique, in Ellis and Rathbone's (1987) study in which the learner "felt stupid and helpless in class" (cited in Ellis, 2012, p. 692). Also, Eyseneck's findings that low level anxiety can lead to more effect is similar to Masha's since her low anxiety outside the classroom was proved to be facilitative in acquiring English. However, unlike Horwitz, Horwitz, and Cope's (1986) findings, Masha's high anxiety level may not entirely reflect her "apprehension at having to communicate spontaneously" (as cited in Ellis, 2012, p. 692). It is applicable only when she is in the classroom.

Willingness to Communicate
As presented in McIntyre's (2001) research, WTC is influenced by variables like "communication anxiety" (cited in Ellis, p. 697) which is strongly reflected in Masha's case as communication anxiety in the classroom prevents her from participating actively in the classroom activities. It can also be concluded that given her autonomous and experiential learning style, she did not enjoy the teachercentered deductive presentations of grammar rules. Also, Masha's personality had a decisive effect on her language learning. Being an extrovert outside class and quiet in the class, Masha did not like to spend time in classrooms, and therefore, she did not quite enjoy her "academic success" as noted by Griffith (1991) with regards to introverted learners (cited in Ellis, 2012, p. 674).

Conclusion
Finally, as the researcher compared his findings to that of his own learning background and IDs, he finds a number of similarities as well as differences. The major similarities are: Ÿ Both of them are the products of teacher-centered pedagogy Ÿ Both of them studied similar textbooks Ÿ Both of them had a high level of anxiety in the classroom Ÿ Both of them were unwilling to communicate in class While observing their differences, the researcher found the following: Ÿ The interviewee had access to satellite TV channels while the researcher did not, and thereby could not avail the opportunity of input flooding Ÿ The interviewee's anxiety level was low outside the class, but the researcher had a high level of anxiety both inside and outside of his class till his tertiary level of education Ÿ The interviewee was very willing to communicate outside the classroom whereas the researcher was not until the later part of his tertiary level of education.
Based on the similarities and differences between the researcher and the interviewee, it may be surprising to note that, despite the fact that the researcher did not have much opportunity for input for a long time, and that his anxiety level was pretty high both inside and outside of class, the researcher still managed to learn English and reached an advanced stage of learning. So, it can be concluded that it may be difficult to measure the effect of IDs on language learning since the apparent negative impact of an ID may not be negative and vice-versa.

Limitations of the study
This study has a few limitations. Firstly, the researcher of the study, being the primary instrument for data collection and analysis, may not have fully overcome the human subjective bias in selecting and organizing issues that he found pertinent. Therefore, though the researcher was conscious about not influencing the study at all, there may have been unconscious attempts to manipulate the subject's answers, thereby affecting the reliability, validity, and generalizability of the study. Secondly, since it is a case study, it may never claim its findings to be truly representative of similar sets of subjects as Hamel (1993) said, "…case stud [ies] [have] basically been faulted for its lack of representativeness" (p. 23).However, the researcher has listened to the tape recorded interview repeatedly to detect and thereby remove any single example of subjective bias to make the findings as objective as possible.