Exploring self-perceived communication competence in foreign language learning

Speaking self-perceived communication competence (SPCC) is a construct with many potential implications for foreign language learning, but one that has been little studied. SPCC itself is a major predictive factor in willingness to communicate, a construct which has been widely conceptualised and researched. This study (N = 103) used a repeated measures ANOVA to investigate SPCC and its correlation with actual L2 speaking proficiency over the course of a year; there was no significant correlation. Qualitative data was then treated with grounded theory to establish why SPCC was inaccurate and to provide pointers as to how SPCC accuracy might be improved. The findings are discussed with reference to the literature in an attempt to establish a deeper understanding of SPCC, particularly in the Japanese context, its formulation and its implications for foreign language learning.

on SPCC has been done for the L1, this study will concentrate on SPCC in foreign language learning (FLL), a crucial component in willingness to communicate (WTC) (McCroskey & Baer, 1985;Yu, Li, & Gao, 2011), which in turn is instrumental in the success or failure of FLL (MacIntyre, Clément, Dörnyei, & Noels, 1998).Self-perceived communication competence, particularly its accuracy, could be an important area for FLL as accurate and inaccurate self-beliefs can help or hinder approaches to learning (Mercer, 2011); however, being cognisant of strengths and weaknesses allows students to "adjust their own cognition and thinking to be more adaptive to diverse tasks and, thus, facilitate learning" (Pintrich, 2002, p. 222).This article will argue that if a student of foreign languages is able to gauge their SPCC accurately, L2 speaking will improve through a greater willingness to speak rendering more L2 experience and therefore higher proficiency.Through reviewing the existing literature and discussing new data it will attempt to find pedagogical ways to facilitate this.

Literature Review
Self-Perceived Communication Competence in Willingness to Communicate WTC, like SPCC, is a concept that originates in the L1 communication field (McCroskey & Baer, 1985) and was conceptualised for FLL by MacIntyre et al. (1998).WTC "is the main cause of second language use" (Yu, Li, & Gou, 2011, p. 253), as language learners with a higher degree of WTC will be more active in the L2.Developing WTC in learners is therefore a desirable goal for language teaching (MacIntyre et al., 1998;Mercer, 2011) as greater L2 experience is likely to lead to greater proficiency.WTC in FLL has many contributing constructs, self-evaluative, motivational, contextual, personality-based and situational, and is commonly presented as a heuristic pyramid (MacIntyre et al., 1998, p. 547), representing a person's state of mind as they decide whether to utilize their L2 or not.Yashima's (2002) study is one of the best known studies on WTC in the Japanese context, where this study also took place.She carried out a comparison of WTC in Canada and Japan and found that in most cases the original Canadian WTC model was also applicable to the Japanese context.Among other similarities, international posture and international communication interest contributed to WTC in both contexts and L2 self-confidence was more predictive of WTC than actual proficiency.However, where Japan did differ was on the matter of motivation and WTC; motivation was only directly correlated to WTC when coupled with self-confidence.Hence, it is suggested that self-confidence may have particular weight in the Japanese context.
It is the self-confidence construct in WTC that is key for this study as Yu, Li, and Gou (2011, p. 256), in their study on the personality-based variables and the correlations underlying WTC, identified speaking self-confidence as being the same construct as SPCC; this study will do the same.The construct has two contributory factors: self-evaluation of L2 proficiency and foreign language anxiety (MacIntyre et al., 1998).

Self-evaluation
Self-evaluative "beliefs are quite vital in deciding human activity especially [given that] humans tend to regulate the level and the distribution of effort spent vis-à-vis the effects expected from their actions" (Anyadubalu, 2010, p. 194); "people must feel sufficiently competent at the instrumental activities to achieve their desired outcomes" (Deci, 1995, p. 64).Anyadubalu (2010) found that higher self-evaluative feeling equalled lower anxiety and better performance and Hashimoto (2002, p. 57) found that an "increased perceived competence will lead to increased motivation which in turn affects frequency of L2 use in the classroom." The role of culture and self-system in self-evaluation is an important one (Mercer, 2011) as it is "instrumental in . . ., motivation and in the regulation of interpersonal processes such as person perception, social comparison, and the seeking and shaping of social interaction" (Markus & Kitayama, 1991, p. 230).There are two self-views; the Independent, which is characteristic of North American and some other European cultures, and the Interdependent, characteristic of Japanese, but also many Asian, African, South American and southern European cultures (Markus & Kitayama, 1991).Independent selves view the self as a distinct entity and seek to "discover and express [their] unique attributes" (Markus & Kitayama, 1991, p. 226).Interdependent selves "insist on the fundamental connectedness of human beings to each other [and recognize that] one's behaviour is determined, contingent on, and, to a large extent organised by what the actor perceives to be the thoughts, feelings, and actions of others in the relationship" (Markus & Kitayama, 1991, p. 227).
While linked by self-view similarities, however, these geographically widespread cultures are clearly each unique; forming, maintaining and perpetuating their interdependent self-view in differing ways.In Japan selfevaluation is characterised "not by seeking positive self-regard but rather by maintaining a chronic self-critical view" (Heine, Lehman, Markus, & Kitayama, 1999, p. 767) and "possessing, let alone enhancing or maintaining, a positive evaluation of the self disconnected from the social context is not a primary concern for Japanese" (p.770).Self-evaluative characteristics include "self-criticism, self-discipline, effort, perseverance, the importance of others, shame and apologies, balance and emotional restraint" (p.769); this leads to a concern and awareness of one's weaknesses as opposed to one's strengths.In education, this manifests itself as a hesitancy to assume superior proficiency than classmates (Heine, Takata, & Lehman, 2000) and a student awareness and concentration upon what is not yet known rather than knowledge already acquired (Aspinall, 2006).It is in fact "considered immature and bad manners for the learner to 'show off' something they have learned, or be ostentatious in any way" (Aspinall, 2006, p. 263).These cultural tendencies to self-critical and humble behaviour clearly have great implications for self-evaluation and classroom behaviour as they contribute to SPCC in Japan.Horwitz, Horwitz, & Cope, (1986, p. 125) defined anxiety as "the subjective feeling of tension, apprehension, nervousness, and worry associated with an arousal of the autonomic nervous system."It may in part be caused by low self-evaluation (Anyadubalu, 2010;Pellegrino, 2005;)."The special communication apprehension permeating [FLL] derives from the personal knowledge that one will almost certainly have difficulty understanding others and making oneself understood" (Horwitz et al., 1986, p. 127).Horwitz et al. (1986) named this "foreign language anxiety" and described it as a "distinct complex of selfperceptions, beliefs, feelings and behaviours related to classroom language learning" (p.128).It is particularly focussed on speaking as the most active and public of the language skills.

Anxiety
Anxiety can refer to both context specific anxiety, and also to trait anxiety, a personality-based construct; either kind can be debilitating in FLL (Dörnyei, 2005).Anxious students seem to speak less and due to lack of experience become more anxious; they then self-evaluate themselves as less competent (Dörnyei, 2005;Kitano, 2001;Yu, Li, & Gou, 2011).Anyadubalu (2010) found that high levels of anxiety adversely affected acquisition and performance while Andrade and Williams (2009) found that "higher levels of anxiety tended to indicate lower levels of proficiency" (p.5); anxiety was related to fear of being negatively evaluated while conversing with others.
Context specific anxiety is a temporary state influenced by immediate environment (Ushioda, 2003) and "may be increased by many factors such as unpleasant prior experiences, intergroup tension, increased fear of assimilation or, an increased number of people listening" (MacIntyre et al., 1998, p. 549).This commonly occurs in stressful situations like tests, or when put on the spot in conversations (Horwitz et al., 1986).However, conversely there are times when context specific anxiety can be facilitative due to performance pressure and adrenalin (Horwitz et al., 1986).Andrade and Williams (2009) actually found that many students expect anxiety provoking situations in the FLL classroom and are mentally prepared beforehand.For this reason, "the majority of students do not feel an intense, persistent, hindering anxiety" (p.11) that only affects a small minority and in differing degrees.
Culture can also play a role here, especially in an FLL situation if a person is out of their own cultural milieu, for example learning their L2 abroad or from a foreign teacher; such multicultural settings possibly generate a "complex construct that combines language anxiety, self-perceptions of L2 proficiency, and attitudinal/motivational components" (Dörnyei, 2005, p. 200).Jamshidnejad (2010) reported that in some such situations, and also when interlocutors' proficiency and social status were perceived as higher, some language learners simply give up speaking through anxiety.

Ways to Promote Accurate SPCC in the Classroom
Accurate SPCC seems to be facilitated through lessened anxiety and better self-evaluation, therefore educational methods to lessen the former and promote the latter seem particularly important.The literature points to two main contributory areas; firstly the educational environment, classroom and teaching method, as in "language acquisition, the person cannot be meaningfully separated from the social environment within which he/she operates" (Ushioda, 2010a, p. 16).Secondly, educator attitude and approach, which has perhaps the most significant bearing on a learning situation (Horwitz, 2001).
Educational environment.To construct an educational environment conducive to SPCC's contributory factors, Palacios (1998) recommended that curricula should encompass clear goals so that students are aware of the purpose and potential outcomes, both short and long term, of their study.Learner logs and reflection journals can provide a powerful tool for students to compare their progress with curriculum goals and therefore feel their proficiency gains (Kitano, 2001).
Many authorities recommend using student centred methodologies in supportive learning environments (Deci, 1995;Dörnyei, 2005).Anxiety can be lessened through the initial use of pair and group work rather than whole class activities (Anyadubalu, 2010), and this could be particularly beneficial for lower level learners (Andrade & Williams, 2009).Scaffolding learners' involvement in discussions from pairs to small groups and finally to whole class discussions increases student confidence and enables better self-evaluation (De Saint Léger & Storch, 2009).
Learner autonomy also provides a powerful tool to increase feelings of empowerment, well-being and self-evaluation (Deci, 1995;Kitano, 2001;Mills, Pajares, & Herron, 2007)."Synthesis occurs when there is enough support in the social context so that the natural, proactive tendencies are able to flourish.However in the absence of adequate support, not only will intrinsic motivation be undermined, but so too will the development of a more integrated or coherent sense of self" (Deci, 1995, p. 83).To this end, using supportive, scaffolded classroom tasks with effective language learning strategies and peer tutoring embedded in lesson materials to motivate learners to think for themselves (Kitano, 2001;Mills et al., 2007;Ushioda, 2010a) should lead to an increase in intrinsic motivation as autonomy increases (Fukuda, Sakata, & Takeuchi, 2011).
Ushioda (2010b) recommends a Vygotskian approach, stimulating students at an individually appropriate level.Feelings of competence occur not when someone has done something "trivially easy, [but] when one has worked toward accomplishment" (Deci, 1995, p. 66), suggesting that differentiation and teaching to the top not the middle, particularly in mixed proficiency classes, would contribute to maintaining this stimulating level for as many learners as possible.Encouraging students to cooperate rather than compete and enabling stronger students to use their L2 skills to support weaker students should also benefit SPCC accuracy through more practical usage and improved interpersonal and social skills.
Many of the methodologies mentioned here were conceived of and developed in cultures characterized by the independent self-view and may experience challenges in other contexts, such as Japan one (Goto Butler, 2011), due to educational culture such as teacher-centred traditions (Aspinall, 2006;Goto Butler, 2011;Rohlen & LeTendre, 1995).However, educators who acknowledge local cultural norms and work with them sensitively and adaptively rather than assume that all self-views will react in the same way are likely to be able over time to enact the methods they aspire to (Aspinall, 2006;Goto Butler, 2011).
Educator approach.There has been little scientific work done on educator factors in motivation (Dörnyei, 2005), but the literature affords some pointers stemming mainly from the basis that if learning does not take place or is hindered, self-evaluation will fall, anxiety will rise and therefore SPCC accuracy will suffer.Dweck's (2006) concepts of fixed and growth mindset seem to have a strong bearing on educator approach.A growth mindset represents the belief that abilities and outcomes result from the individual's own effort, that a person's true potential is unknown.A fixed mindset on the other hand is characteristic of people who believe their personal qualities are unchangeable, they would rather not challenge themselves or leave themselves open to the possibil-ity of failure.Therefore, promoting growth mindset would seem to be one way of improving self-evaluation and lessening anxiety through more meaningful learning and feelings of achievement; this centres on praise techniques as unwarranted praise encourages the attitude that if you are already brilliant, why try harder?(Dweck, 2006).Most importantly, praising intelligence and talent rather than effort "implies that we're proud of [the student/child] for their intelligence or talent rather than for the work they put in" (Dweck, 2006, p. 177).Mercer and Ryan (2010, p. 442) in their study of mindset in EFL wrote that praise should be limited to the growth-orientated process, focussing "feedback on learners' efforts, the process of learning, and beliefs about developing one's ability through hard work."Interestingly, the attitude that anyone can do anything if they try hard enough is a strong characteristic of the Japanese educational context (Aspinall, 2006) along with a tendency to not lavish praise on learners (Rohlen & LeTendre, 1995); perhaps the growth mindset factor could be one of the reasons the Japanese education system performs so well in international comparisons?
Supportive and understanding teachers may be the biggest factor in reducing anxiety (Horwitz, 2001) as "problematic" students probably have some kind of anxiety at the root of their problems (Horwitz et al., 1986).Knowing students well as individuals, being flexible (Horwitz, 2001) and exhibiting interest in them for their own sake can contribute to lessened anxiety.Error correction techniques should therefore focus on reducing defensive reactions in students as well as improving L2 proficiency but not by correcting every little mistake (Horwitz et al., 1986).Horwitz (2001) also recommended against teaching the foreign language as a massive memory exercise; realistic, context specific and appropriately paced curriculums seem to work best.
The role of cultural and ethnic differences in teacher-student relations and anxiety caused by it is mentioned by Horwitz (2001).In Japan the number of non-Japanese language educators is large due to government policy and the commercial profitability of the EFL industry.These teachers are overwhelmingly native speakers of English, normally from independent-self cultures, in contrast to their students from an interdependent-self culture.There can be deep misunderstandings between people of differing self-views (Heine et al., 1999) and it seems unlikely that these would not manifest themselves in FLL.
Interdependent self-view could have a major bearing on educator approach in the classroom, for example in group and discussion work where students may be less anxious and perform more effectively in groups established over time rather than constructing new groups for each activity.It may also be that to ask a student's opinion on a subject suddenly without giving time for the student to study the social context may produce silence.This may not mean that the student has no opinion, or that they cannot speak the L2 (Jamshidnejad, 2010), but simply that for the interdependent self, opinion and expression will depend on subtle variations in the social context so the student may be unsure what to say.It would seem sensible to ask complex questions to the whole class instead of individuals, giving students time to formulate their opinion in contextual relation first.
Emotions and expressions could also be an area where self-view misunderstandings may flourish.For "independent selves, emotional expressions may literally 'express' or reveal the inner feelings such as anger, sadness, and fear . . .for interdependent selves however, an emotional expression may be more often regarded as a public instrumental action that may or may not be related directly to the inner feeling" (Markus & Kitayama, 1991, p. 236).This suggests that educators need to become adept at reading the situation to judge whether a student has misunderstood or is simply waiting to read the social context before acting.This will also be relevant when offering choice.In Japan it is the responsibility of the teacher to establish what is correct in a given situation rather than offer a choice (Markus & Kitayama, 1991); if offered a choice therefore a student may well act baffled, thereby giving the non-Japanese teacher the impression that the student cannot react linguistically rather than culturally.Similar cases are reported in Jamshidnejad's (2010) Iranian Study.It is suggested that putting students under pressure in these way is likely to cause anxiety and negative self-evaluation.
Finally, due to interdependent self-view, it is a marked characteristic of the Japanese educational system that discipline and class harmony are kept not through what might be termed overt behaviour control techniques, but through the forming of groups and relationships which bind students and educator together (Rohlen & LeTendre, 1995).Therefore, the group membership is another area of potential misunderstanding; if a non-Japanese teacher is not aware of the key role of groups and group membership, they may not be able to formulate their own position in the class in relation to it.This could cause discomfort for all concerned, and non-Japanese teachers will need to be sensitive to this fact and work with it.
Self-perceived communication competence is a construct highly significant for WTC and therefore the success or failure of FLL; it contains two factors, self-evaluation and anxiety.Both can fluctuate depending on cultural and environmental variables, but both have a strong effect on each other.There may be many ways language educators can promote accurate SPCC through reducing anxiety and improving self-evaluation; primarily however it seems to depend on a culturally sensitive, supportive, clearly structured, studentcentred approach which promotes autonomy and growth mindset and retains learner awareness of progress.

Research Questions
The aim of this study is to establish further how SPCC can be made more accurate; to this end the research questions to be addressed are: 1. Did this research population have an accurate SPCC? 2. What reasons can be determined for the answer to Research question 1? 3. What can the data tell us about further ways to foster accurate SPCC in FLL?

Setting and Participants
The participants (N = 103) were first year university students (18 or 19 years old), all unknown to the researcher and enrolled in the English department at a private foreign language university near Tokyo.The university employs a high number of non-Japanese educators and markets the institution very successfully as a place where "communicative" English taught by native speakers "will" facilitate effective language learning; the inference is that contact with native speakers is the key to this.Many students choose the university precisely for this reason and are often recommended to do so by their high school teachers and parents.
The "English communication" classes that the participants were members of put a heavy emphasis on speaking and were mandated to be conducted in English by both teachers and students.Participants average TOEFL scores were: listening 47.3, writing 43.7 and reading 46.4,classing them as A2-B1 (elementary, independent users), on the Common European Framework of Reference for Languages (CEFR).

Instrument
This study used a questionnaire (see Appendix A) to measure participant self-confidence in speaking as well as the other three language skills and grammar.It was created from data collected in two focus groups (n = 9) and was used previously in Lockley and Farrell (2011).Its Cronbach's reliability was .94 for the speaking component on that occasion.For this research, the questionnaire was administered at the beginning of semester one in March 2010 and again to the same participants at the end of semester two in February 2011.It collected numerical data to establish self-confidence values and also non-compulsory written qualitative data which sought to obtain reasons for the numerical data given.The scores for self-confidence were obtained by asking three virtually identical but differently worded questions; a single scale with three items for each skill.Participants indicated their self-confidence on a 5-point Likert scale which allowed a score of 15 points for each skill when the three question scores were combined.This article was restricted to selfconfidence in speaking; SPCC (Yu, Li, & Gou, 2011).Other data were not used.
To establish actual proficiency, the Kanda English Proficiency Test (KEPT) speaking examination was used.The students took the exam shortly before the first semester and shortly after the second semester, coinciding closely with the questionnaire administration.KEPT requires three or four examinees to hold an impromptu conversation for seven minutes after reading a short topic prompt.The two independent assessors grade fluency, lexis/grammar, pronunciation, and conversation skill (see Appendix B) out of 4 (for a possible total of 16).Bonk and Ockey (2003) found that the Rasch model enables examinees to be reliably separated by ability and Van Moere (2006) found that the test was a reliable measure of a candidate's ability in L2 speaking.
A repeated measures analysis of variance (ANOVA) investigated the relationship between KEPT and SPCC, both in March 2010 and February 2011 and then the correlation between the two KEPT and the two SPCC values.Although the number of participants and the size of the scale were relatively small, the repeated measures strengthened the study by rendering more data per subject and more power.For descriptive statistics see Table 1.Given the exploratory nature of the study and the importance of ensuring that findings were grounded in the actual experience of learners (Pellegrino, 2005), grounded theory following Dörnyei's (2007) method was used for the qualitative data.Dörnyei advocated a 3-stage process, firstly open coding, secondly axial coding and finally selective coding.Dey (2007), described this process as allowing "comparison and contrast, links and connections" (p.173) to emerge, enabling construction of a coherent narrative and informed discussion.

Quantitative Data
The descriptive statistics (Table 1) show that in March 2010 the average SPCC score achieved by participants was only 6.05 from a possible 15 (37.81%); this had increased slightly to 7.28 (45.5%), in February 2011.Participant KEPT scores were 6.89 (45.93%) of the total possible, during the first data collection and 7.93 (52.87%), during the second.This shows that the SPCC was in fact lower percentage-wise than actual proficiency although both SPCC and proficiency did rise over the year Table 2 shows the ANOVA results for SPCC and KEPT and the correlation between the two.The rise in the SPCC speaking and KEPT test scores were both individually statistically significant, p < .05but there was no correlation between them, p > .05.This means that statistically students did improve both their SPCC and their actual proficiency, but that the two had no significant effect on each other.Higher actual proficiency did not lead to higher SPCC, or vice versa, and SPCC was neither accurate at the beginning or the end of the year.The answer to Research question 1 is that this research population did not have an accurate SPCC.

Qualitative Data
The qualitative data brings a deeper and more faceted picture to the study.Three categories emerged from the coding of the first data set, each with several subcategories: (a) negative self-evaluation of speaking, (b) positive self-evaluation of speaking and, (c) anxiety.These categories emerged organically but correspond to the formulation of SPCC in the literature.
Negative self-evaluation was the most numerous (n = 72) and there were six subcategories, firstly (n = 19) attribution to lack of experience; this included comments like "I've never talk to native English speaker.So I don't experience to use English."The second subcategory showed that many (n = 18) had difficulty expressing themselves; frustratingly their language level did not match what they wanted to say.This was represented by comments such as "I have a lot of things I want to talk, but I can't make a sentence."The third subcategory included unattributed statements of poor proficiency (n = 15) such as "I am not good at speaking."The fourth attributed poor speaking to issues with specific language skills (n = 8); for example, "It takes time to translate into English.My pronunciation is not good."The fifth (n = 8) represented concern about lack of speed when taking part in conversation and included comments such as "I can't find right English words in an instant."The final subcategory (n = 4) represented overly high expectations of what language learners at their level of L2 proficiency should be able to do; for example "I can't speak fluency." The second largest category (n = 15) was positive self-evaluation of speaking.There were four subcategories of which the largest (n = 7) attributed positive self-evaluation to experience; a representative comment was "I have traveled to some foreign countries and I speaked foreigners there."The second subcategory (n = 5) was personal attribution, for example, "I like to talk and try to use words or phrases that I have learnt."The third subcategory (n = 2) was unattributed confidence such as "because I am confident that I say my opinion in English."The final subcategory included only one comment attributing confidence to success on a test, "My ability of speaking is just above passing the exam, Eiken second grade [CEFR B1/2]" which is a good level to have reached for learners of this stage.
The third main category was anxiety (n = 10) with three subcategories.The first was grammar anxiety, for example, "I cannot speak English smoothly when I care too much about grammar."The second was context specific anxiety hindering speaking performance, for instance, "when I speak to someone in English, I am always very nervous."The third comprised two identical comments of trait anxiety: "I am shy." In February 2011 four main categories emerged.They were again (a) negative self-evaluation, (b) positive self-evaluation and, (c) anxiety.However, this time a new category (d) both positive and negative self-evaluative comments, also appeared.It is worth noting that the comments in general were noticeably longer and more facetted than they had been the previous year.
The largest main category was again negative self-evaluation (n = 65), and once more comprised six subcategories of which the largest (n = 24) attributed poor speaking to specific language skill deficiencies.A representative comment was "recently, I want to speak English more properly, follow grammar.But this made can't speak English because I have to think before I speak." The second subcategory (n = 13) was unattributed poor proficiency, for example "I cannot come up with phrases smoothly when talking with someone."A third subcategory (n = 11) was again frustration over not being able to explain ideas that students were cognitively able to conceive, for example "I like to communicate with many people, but often I cannot tell what I want to say."The fourth subcategory (n = 7) represented overly high expectations for students of their proficiency; all comments were virtually identical: "I cannot speak English fluently."The fifth subcategory (n = 5) was to do with native speakers with comments like "it is difficult to explain my feeling in English.Recently I hesitate to talk with [native speakers]."The final subcategory (n = 5) was again to do with speed of conversation, for example, "it takes much time for me to speak English because I often stop to think what to say next." The second main category was positive self-evaluation (n = 15) comprising four subcategories, firstly unattributed proficiency (n = 7); an example comment was "I can speak exactly what I want to say."The second subcategory was experience as attribution (n = 4), "because I like to speak English and I lived in America."The third subcategory was confidence through test results (n = 2), for example, "I get good score of listening test" and the final subcategory (n = 2) was again to do with native speakers, this time a sense of confidence through being able to understand, for example, "I can understand what [native speaker] teacher said." The third main category was anxiety (n = 10) with four subcategories.This time anxiety appeared to be both facilitative and debilitative, particularly in relation to the first subcategory, grammar anxiety (n = 3).Issues with grammar also appeared in the negative self-evaluation category, but were only included in the anxiety category where anxiety, or lack of it, seemed to be the main focus.For two participants, lack of anxiety about grammar was facilitative: "I don't think about grammar much when I speak, but I think speak naturally is more important than grammar," but for one other it was still debilitative: "I'm worried whether I make a grammar mistake or not when I'm speaking."A second subcategory was anxiety about native speakers (n = 3), and one comment was "I am good enough at speaking English with my friends but I cannot speak English well when talking with [native speakers] because I get nervous."The third subcategory concerned lack of confidence (n = 2): "I am not confident about my ability to speak English."The final subcategory was possible context anxiety (n = 2), represented by "when I feel nervous, I cannot speak even very simple English." This time there was one other main category, so there were no distinct subcategories.This main category comprised both positive and negative selfevaluations (n = 6); two examples of this category are "I can speak one to one, but it is hard to speak to whole class" and, "my English is not enough but I can talk with native speakers without problems."

Selective Coding
This section will attempt to interpret the data to answer Research question 2, which was why was SPCC inaccurate?The most noticeable factor on both data sets was the overwhelming instance of negative self-evaluation.Perhaps however this was to be expected given the above noted propensity of Japanese people to be self-critical (Heine et al., 1999), conclude that they are performing worse than they actually are (Heine et al., 2000) and be modest about achievements (Aspinall, 2006).Horwitz et al. (1986) also suggested that lower self-evaluation may be a common characteristic of all language learners.A perhaps significant trend was for the subcategories from February 2011 to show a more attributed tendency and better awareness of deficiencies in specific language skills; there was also the appearance of the new category of both positive and negative self-evaluations.These could perhaps point to greater skill-level internal comparisons which may develop with language proficiency in some learners (Mercer, 2011;Mercer & Ryan, 2010); others however "may simply tend towards global self-descriptors" (Mercer, 2011, p. 103) perhaps therefore accounting for the continuing presence of some unattributed comments.It is possible that this greater awareness of the construction of the L2 actually caused increased anxiety through the realization of how much hard work and time it takes to learn a foreign language.
A minor subcategory indicated that grammar anxiety may have become a facilitative rather than debilitative factor for some participants by February 2011, with students actively attempting to suppress anxiety to speak smoothly.Furthermore, there were also a small number of comments that performance on a test had contributed to positive self-evaluations.Horwitz et al. (1986) mentioned that knowledge of errors made on tests can contribute to anxiety which in turn can lead to more errors being made on tests in the future and it appears that the opposite may also be possible.The number of comments to do with learner frustration and inability to express themselves in English had fallen somewhat too, and comments about difficulties with the speed of conversation had also lessened.This could show that a small number of participants had become less anxious through greater proficiency during the year but anxiety about limited expression may in fact be characteristic of FLL as the awareness that "range of communicative choice and authenticity is restricted" (Horwitz et al., 1986, p. 128) in itself causes anxiety and lower self-evaluation.
The words native speaker occurred repeatedly in both data sets and in multiple categories; it is possible that these could be representative of the complicated construct forged in multi-cultural settings (Dörnyei, 2005) but it could also represent instances of the so-called "native speakerism" (Holliday, 2006).Native speakerism refers to the idealizing, even idolizing, of The Native Speaker and the variety of language they speak, almost always, and certainly in this context, a native speaker of English."The 'native speaker' ideal plays a widespread and complex iconic role . . .an underlying theme is the 'othering' of students and colleagues from outside the English-speaking West according to essentialist regional or religious cultural stereotypes" (Holliday, 2006, p. 385).In Japan native speakers and English proficiency are connected with not only linguistic idealism, but also social capital, inclusion in an imagined global community, and romance (Kubota, 2011).Even current government policy assumes that involving native speaking teachers and assistants will automatically increase English proficiency despite the previous government finding this was a waste of money (Mie, 2013).The amount of negative self-evaluation comments concerning lack of experience with native speakers on the first data set and the institutional native speaker emphasis attest to the likely presence of some kind of native speakerism among these participants.
By the second data set lack of experience, the largest attribution in the first data, had disappeared; instead a small number of comments about hesitating to talk to native speakers from feelings of linguistic inadequacy and talking to native speakers causing anxiety had appeared.On the other hand, the ability to understand and communicate with native speakers lead to positive selfevaluations for two students; it seems perhaps that native speaker related beliefs had worked for a small minority who felt they were reaching their ideal, but had had a debilitative effect for many.Perhaps the experience of having native speaker teachers in itself actually caused anxiety and low self-evaluation in some students; comparison against a more proficient and honestly speaking linguistically privileged (due to having the desired language as a mother tongue) Other may have caused anxiety, and unrealistic judgments (Horwitz et al., 1986), about their own potential proficiency at this stage of language learning.It may also be that even if comments did not explicitly mention native speakers, the drop in attributions of positive self-evaluation to experience, the overly high expectations, and some negative self-evaluations (against a native speaker ideal) may represent instances of latent native speakerism.

Discussion
The first research question concerned how SPCC and actual L2 speaking proficiency (KEPT) correlated in this population; although all participants' SPCC and proficiency had improved over the year, the SPCC was inaccurate.The answer to Research question 2, why was the SPCC inaccurate, seems to lie in a multitude of factors that took place over the course of the academic year to ensure continued anxiety and lower self-evaluation tendencies.These factors may have included continued frustration at limited communication ability; the realization that learning a foreign language, even intensively in a dedicated institution is challenging; and context specific anxiety connected to native speakerism, which may also have contributed to a variety of other selfevaluation lowering instances.Positive self-evaluation and lessened anxiety were perhaps contributed to by an affirmation of proficiency through test results, a liberation from grammar anxiety, and native speakerism beliefs and expectations being seemingly affirmed and fulfilled.
The third research question concerns deriving pointers from the data and literature as to what can be done to foster a more accurate SPCC.An important point is that much of the negative self-evaluation may be culturally and perhaps also subject specific, which could therefore mean that the Japanese self-view may admit more negative self-evaluation than some other contexts to obtain an accurate SPCC.Promoting a greater educator awareness of both context specific self-evaluatory tendencies and SPCC in general would seem a good idea so that lesson and curriculum planning can take them into account.The rest of the pointers will only be useful when educators become more cognizant of these issues.
One of the major trends was a dissonance between communication ability and desired linguistic expression.One possible remedy for this may lie in an increased offering of content and language integrated learning (CLIL), to create "a climate which fosters continuous language growth" (Mehisto, Marsh, & Frigols, 2008, p. 32) and facilitate the learning of language of interest and relevance to students' lives.CLIL can also improve critical thinking skills, higher level vocabulary and real life speaking experiences (Mehisto at al., 2008).
Another issue seemed to be overly high expectations for this proficiency level.Kitano (2001) urges educators to "watch for learners who immediately set their goals as high as the level of native speakers, because this unrealistic expectation inevitably makes them perceive their ability as insufficient and causes them anxiety" (p.559).Kitano (2001) suggested that such students should be counselled in realistic "standards or short-term goals in language learning and incorporate standards of evaluation that encourage this" (p.559).These assessments could be formative, acting as bridge to further learning by showing students where they went wrong and how to improve at the next attempt.This has been shown to aid large improvement in learner achievement (Black & Wiliam, 1998).Combined with scaffolded learning and embedded reflection (Mills et al., 2007), it is suggested that students would be better able to understand their level of language and its potential uses.
Many students also seemed to have a lack of understanding about the system of language itself, perhaps evident in the amount of unattributed comments on both data sets and lack of skill-level introspection for the majority of participants.Although this may be normal in lower level learners (Mercer, 2011), there is still no reason why educators should not try teaching more about languages and their formulation or focus learning strategies on fostering more attributed thinking.Foreign language learning curriculums could include an "introduction to the L2" course, describing the makeup of a language and providing a rationale for why learning the language is important.Brooks-Lewis (2010) for example taught ments of the history of English to her students, helping to provide "a foundation for the constructing of learning" (p.148).
Native speakerism was an unanticipated finding of this study although perhaps not surprising in an institution which puts such a weight on its native speaker educators; many of the students were attracted to the university by this fact and their expectations will have been shaped by it.Perhaps then when these students realized that in fact simply having a native speaker teacher was not a panacea for FLL, disillusionment with previously held native speakerism beliefs set in.This could have contributed to feelings of inadequacy and "otherness", as well as alienation from target cultures (Holliday, 2006), which in turn could have led to lower self-evaluation (Mercer, 2011).Native speakerism is not confined to this institution nor to Japan (Holliday, 2006), it is a widespread social problem and is difficult to combat.However, if educators were more aware of the phenomenon and willing to manage it constructively in their own contexts on an individual basis, it is possible that it may become less of an issue for students in those educators' classes.It would require a good deal of introspection by teachers as to how they might be promoting it, if at all, and a willingness to then enact specific and individualized strategies to confront it.This could well be a difficult prospect, as, of course, one of the groups who benefit consciously or unconsciously from the phenomenon most, in social standing, employment prospects, sometimes even sexually (Appleby, 2013), are the educators themselves and there must be an acknowledgement of complicity by some educators in the situation.
Given the feelings of otherness that native speakerism renders, promoting a greater sense of connection and "identity with the target [language] community" (Mercer, 2011, p. 27) may aid self-evaluation.This could perhaps be done through CLIL lessons in the history, society and geography of target language countries, which can be powerful tools to promote connectivity (Mehisto et al., 2008).In EFL contexts, such as the one where this study took place, it could also be emphasised that in this day and age English is used as much to talk to non-native speakers as an international lingua franca rather than purely as a tool to speak to native speakers.
The final issue, which may also have implications for the othering of students and educators, is that of culture and differing self-view.In this context the native speaker teachers were overwhelmingly from cultures characteristic of the independent self-view and the students from the interdependent.The findings of this study suggest that it seems important for educators to be as sensitive as possible to cultural differences (see also Horwitz, 2001) and not to judge classroom occurrences by their own cultural expectations.This may be a worthy but challenging proposition as attempting to understand differences in self-evaluation practices "can evoke puzzlement, disbelief and pejorative assessments of the other world" (Heine et al., 1999, p. 769).
So in what way could educators from different national and ethnic groups (this does not only refer to those from English speaking countries, it may be also equally applicable to Japanese people teaching in the UK, for example) achieve greater familiarity and cultural expertise?It would be unrealistic in the short term to expect widespread educator retraining, although perhaps language teacher training programs might put more emphasis on this in the future.Instead perhaps institutions might include cultural familiarization training for new staff and encourage educators to read more widely during service.This would benefit all stakeholders through increased student satisfaction and potential SPCC increases leading to increased experience, proficiency and therefore student satisfaction.Educators themselves could also enact simple in-class actions such as consulting students as to their past language learning experiences (Sampson, 2010).

Limitations
This was a small-scale study in a Japanese university, is not globally generalizable and its findings need be seen through culture tinted lenses; what is right in one context may not be right for another.Furthermore, it should be remembered that cultural distinctions are "general tendencies that emerge when the members of a culture are considered as a whole" (Markus & Kitayama, 1991, p. 225), and not necessarily characteristics of specific individuals.
It must also be acknowledged that although a considerable time (11 months) passed between the two questionnaire administrations, there is the possibility that the two may have affected each other.A further limitation is that participants needed to be identified to match up the first data collection with the second, and it is possible that this lack of anonymity could have affected the data.

Conclusion
This article presents a review of the factors involved in L2 speaking SPCC and refers to the literature and new data for ways that accurate SPCC might be promoted in the classroom by educators.It is suggested, with reference to the literature, that this may lead to rises in linguistic awareness, self-evaluation and ultimately to more effective language learning.
The study established that SPCC was inaccurate in this population and suggested that generally low self-evaluation and anxiety, formed through unrealistic expectations, frustration at limited proficiency, native speakerism and perhaps a Japanese tendency to self-criticism, contributed to this.It suggested that educators should employ scaffolded curriculums with embedded learning strategies and reflection, put more emphasis on what a language is and where it comes from, and consider teaching higher level CLIL type lessons to improve vocabulary and familiarity with cognitively challenging subjects and target culture.Furthermore it suggested that in contexts where native speakerism may be an issue, educators should try to manage student expectations and assumptions.Finally, where student and educator are of different background, educators should be encouraged to adapt or at least become more familiar with student cultural norms, in particular when they pertain to students' self-views and educational culture.

Questionnaire
For the following items, circle the number (from 1 = Strong Disagree to 5 = Strongly Agree) that best describes you.

Table 2
SPCC and speaking proficiency