Foreign Language Teaching Anxiety of Non-Native Pre-Service and In-Service EFL Teachers

Despite a robust body of literature on anxiety in language learning, as of yet, teacher’s aspect regarding anxiety seems to have received scant research attention. Apparently, absence of a valid and reliable measuring instrument impeded the empirical research on foreign language teaching anxiety. However, preliminary work well-documents that even teachers get stressed and feel teaching anxiety in their English as a Foreign Language (EFL) classrooms. Thus, the purpose of this study was to investigate the level of foreign language teaching anxiety that non-native pre/in-service EFL teachers experience. The data were collected from 30 in-service and 60 pre-service EFL teachers in two northern cities of Turkey; Trabzon and Yalova. The Foreign Language Teaching Anxiety Scale (FLTAS) was used as the main data collection instrument. No significant difference was found regarding the participants’ gender. When the participants’ scores were compared by the department they were enrolled, it was found that English language teaching department graduate teachers had significantly lower teaching anxiety levels. Significant negative associations between foreign language teaching anxiety and duration of experience as well as graduation department were found. It is believed that the results of the study will be of great contribution to further research into teaching anxiety and have important implications for policymakers of teacher training programs.


Introduction
Anxiety, as a psychological or emotional construct, has driven an ever-increasing research interest in a wide range of disciplines, and extends from clinical psychology, through classroom to language classroom and, relatively new, to foreign language classroom. Among other individual difference variables such as motivation (Yan & Horwitz, 2008), apptitude (Dörnyei, 2005) and attitude (Gardner, 1985), which are thought to have a crucial impact on achievement, anxiety has been widely accepted in second and/or foreign language learning within the past few decades. However, extant work on foreign language anxiety seems to have focused on learners rather than the teachers. Put differently, research on Foreign Language Teaching Anxiety (henceforth FLTA) is sorely missing. Several reasons can be cited for this paucity of literature. One reason may be the roles attributed to teachers within newly emerging so-called humanitarian approaches to language teaching (Larsen-Freeman, 2001) which encouraged focusing mainly on learners, yet another reason might be that lack of a valid and reliable measuring instrument seems to have impeded experimental studies on FLTA.

Background to the Study
The relationship between anxiety and learning has become an increasingly important field of study within the past few decades. Although some suggest contrary, it has been well documented that there is a negative correlation between anxiety and learning, at least to a certain degree (Piniel, 2006;Ohata, 2005). However, anxiety research has focused mainly on alleviating learners' anxiety and increasing their performance, while very little attention has been paid on teacher aspect. Particularly anxiety experienced by foreign language teachers seems to have received scant attention. Apart from a limited number of studies conducted mostly, if not all, with pre-service teachers in various disciplines (Bowers, Eicher, & Sacks, 1983;Ngidi & Sibaya, 2003;Wadlington, Slaton, & Partridge, 1998) apart from English language teaching such as mathematics teachers (Akinsola, 2014), agriculture faculty (Bernstein, 1983), even psychology teachers (Gardner & Leak, 1994), there is a palpable dearth of literature on anxiety in English as a foreign language teaching.
The negative correlation between teaching anxiety and teaching performance have been amply demonstrated in a considerable body of work. It is almost axiomatic that as the teaching anxiety increases, the teaching performance of teachers decreases (Ameen, Guffy, & Jackson, 2002;Fish & Fraser, 2003;Gardner & Leak, 1994;Williams, 1991). In the same vein, Horwitz (1996) posits that "Language teachers with higher levels of language anxiety may also communicate negative messages about language learning to their students." (p. 366). From a wider perspective, it is also claimed that foreign language teachers themselves are considered as constant and/or lifelong learners of the target language and the anxiety they experience "has the potential to affect the way teachers teach, the amount of language students receive, and the role models to whom role-model learners are exposed" (Horwitz, 1996, p. 371). Moroever, recent studies have also documented significant differences between pre-service and in-service teachers regarding their teaching practices as well as performance (Akinsola, 2014;Horwitz, 1996;Wadlington, Slaton, & Partridge, 1998). What is more, the recent body of literature provides substantial evidence that there is a general tendency that as the teachers get experienced their anxiety levels decrease (Fish & Fraser, 2003;Gardner & Leak, 1994) while some others consider it as an ongoing problem (Bernstein, 1983). However, the teacher's anxiety per se as such a vital variable in language teaching seems to have received little research attention, if not totally neglected.

Anxiety
Within this vast area of disciplines numerous definitions as well as taxonomies of anxiety have been presented by scholars. Scovel (1978) offers a neurological definition of anxiety as an emotional state which "is generated through the arousal of the limbic system" (p. 135). In their treatise, Horwitz, Horwitz & Cope's (1986) define anxiety as "subjective feeling of tension, apprehension, nervousness, and worry associated with an arousal of the autonomic nervous system" (p. 125). All in all, common sense suggests that anxiety creates somehow uneasiness or insecurity on individuals. Two main taxonomies depending on the nature of the anxiety have been commonly accepted. On one hand there is the dichotomy of trait and state anxiety (Spielberger, 1983). Dörnyei (2005) defines the dichotomy as "Trait anxiety refers to a stable predisposition to become anxious in a cross-section of situations; state anxiety is the transient, moment-to-moment experience of anxiety as an emotional reaction to the current situation." (p. 198). Moreover, MacIntyre and Gardner (1991) highlight another type of anxiety "situation-specific anxiety" which suggests mainly that individuals may get anxious in specific stiuations. On the other hand, there is facilitating and debilitating anxiety (Alpert & Haber, 1960) which suggests that anxiety can facilitate learning to a certain degree while after a threshold it is believed to hinder learning.
Although introduced with pioneering seminal review of Scovel (1978) as a separate construct, it was not until Horwitz, Horwitz and Cope's (1986) most renowned treatise, had anxiety unique to foreign language learning been established as a unitary construct. Furthermore, despite an ever-expanding accumulation of research which had also offered insightful discussions on the relationship between learning and anxiety, research on foreign language learning anxiety flourished only after Foreign Language Classroom Anxiety Scale (FLCAS) (Horwitz, Horwitz & Cope, 1986). Since then, a proliferation of studies with mounds of data on the relationship between language learning anxiety and achievement have been published (Dörnyei, 2005).
It took a decade for scholars to address, if not notice, the affect that foreign language teachers themselves experienced in their classrooms. Horwitz (1996) was one of the first to provide a description of anxiety experienced by foreign language teachers and remarked, "Even though language teachers are supposed to be high-level speakers of their target language, language learning is never complete, and most nonnative language teachers are likely to have uncomfortable moments speaking their target language." (p. 365). In line with this, in a fairly recent study Ameen, Guffey and Jackson (2002) highlight that "Teaching anxiety appears to be a common occurrence among the vast majority of accounting professors and often manifests itself in the form of physical and psychological symptoms or reactions" (p. 17). Interestingly enough, along with psychological symptoms of distress, apprehension and being upset some of the physical symptoms reported in the study are heart-rate acceleration, gastrointestinal distress, or being flushed. Bernstein (1983) further remarks that in severe conditions anxiety might even end up with physical illnesses such as "ulcers, colitis, cardiac arrhythmias, headache, chronic pain, hypertension" as well as problems of "disrupted family… colleagial relations, and sexual dysfunction" (p. 5). All in all, FLTA seems to have the potential to constitute a significant problem for language teachers.

Native and non-native teachers
It is estimated that over two billion people speak English as a second or foreign language (Crystal, 2008). Research reveals that an overwhelming majority of second language teachers is non-native speakers of English (Ranta, 2010, Tsui, 2003, as has been the lingua franca of the world (Seidlhofer, 2004). Correspondingly, many scholars have harbored deep-seated doubts regarding the ownership of English (Widdowson, 1994). Along with the arguments on the ownership of English (Widdowson, 1994) serious criticism was leveled against the standardized and idealized native-speaker norms (Alptekin, 2002;Leung, Harris, & Rampton, 1997;McKay, 2011) which have granted a highly privileged status to the "White, monolingual English speaker" (Leung, Harris, Rampton, 1997) teachers. Canonical view suggests that language teacher as the main, if not the sole, model of the language in EFL and ESL contexts should be native-speaker who are assumed to be a good model of the target language. In line with this, Kuo (2006) favors a native-speaker model for phonology as well as grammar to "ease or smooth the flow of conversation, to reduce the listener's burden of processing information, and to satisfy learners' needs that stretch beyond merely international intelligibility" (p. 220). Besides, nonnative speakers are considered to be using a somewhat deviant accent or a "deficit model" (Bernat, 2008) of the target language which might be allotting them a subordinate position. However, contemporary sociolinguistic perspective (Jenkins, 2002;Smith, 1992;Seidlhofer, 2004) suggests "intelligibility" as the major premise of language teaching methodology although some others consider intelligibility as a vague term whose borders are not well-defined (Pickering, 2006). On the other hand, non-native speaker teachers are believed to be holding some better qualities such as "a more thorough knowledge of grammar, empathy for the EFL (English as a Foreign Language) learner having been one themselves, being a bilingual resource in the classroom, and understanding the local curricula and contextual demands" (Bernat, 2008, p. 1) as well as better knowledge of the language and better training and experience in teaching skills (Jenkins, 2006). From a different perspective, Norton (1997) remarks the difficulty of designating a native or non-native speaker considering lingua franca footing of English in various parts of the world such as India, Pakistan, Canada. Davies (2003) further brings the native speaker per se as a term to the table for reconsideration. On the whole, there is no consensus reached yet on the current controversy about the native or non-native speaker status of English language teachers and seems to remain so at least in the near foreseeable future.

Foreign Language Teaching Anxiety
Common sense has it that all teachers including English as foreign language teachers experience a certain degree of anxiety particularly at the early periods of their teaching experience. However, only a handful of publications can be found regarding the native versus non-native speaker English language teachers' teaching anxiety. Greis (1985) depicts a general framework for anxiety experienced by teachers highlighting the high level of anxiety of inexperienced teachers in the early stages and remarks that "Anxiety may be felt by any beginning teacher, whether native or non-native. However, when put next to native speaker, the non-NETs [non-native ESOL Trainees] often experience a strong sense of fear that they will not attain the same level of proficiency and that the ESL students may reject them preferring a native speaker as a teacher" (p. 318). By the same token, there seems to be a consensus reached that non-native speaker English language teachers experience anxiety about their perceived insufficient language proficiency whose most visible signal is their foreigner L1 accent (Braine, 2010;Inbar-Lourie, 2005;Reves & Medgyes, 1994). In the same vein, discussing findings of her study, Bernhat (2008) argues that "More than half the respondents felt that they do not belong in front of the classroom, and this may be either a result of lack of teaching experience or feelings of inadequacy due to being NNS (though subsequent interview data showed that the latter cause dominated)" (p. 4). In a nutshell, this remarkably sparse literature on non-native speaker English language teachers clearly demonstrates a prevailing general anxiety due to perceived insufficiency in language proficiency; however, little is empirically known about their FLTA.

Measuring FLTA
Research on foreign language teaching anxiety is still in its infancy. Extant writings cite frequently two studies, thus, scales that specifically measure foreign language teaching anxiety. One study that is highly referred is Horwitz's (1996), which is from Western world. The second study is from Turkey; Ipek's (2006) PhD dissertation. The need to spell out such a differentiation originates from the almost axiomatic impact of culture in human cognition which is duly noted in a robust and extensive body of published works by eminent scholars (Hofstede, Hofstede, & Minkov, 2010;King & McInerney, 2014;Nisbett, Peng, Choi, & Norenzayan, 2001). In line with this, the literature is also replete with contentions of scholars highlighting the differences in unfitting concepts between the Western and Eastern worlds, or in Hofstede's (1984) terms, individualist and collectivist cultures. More specifically put, ithat communicative language teaching, once highly-praised, was a Western artefact and would hardly be a panacea for EFL classrooms of Eastern culture. In the same vein, what may be anxiety arousing for teachers in western culture might not be so in a non-western context or vice versa. However, there is another caveat germane to this discussion worth noting here. Strange to recount, the studies that report to have used Horwitz's scale cite her 1996 work. Nevertheless, Horwitz presents no scale yet acknowledges the type of anxiety that is peculiar to foreign language teaching in the aforementioned study. Moreover, to the best of present authors' knowledge, no foreign language teaching anxiety scale has been published as of yet by the renowned scholar Elaine K. Horwitz. All things considered, İpek's (2006) scale emerges as a more viable option with reliability and validity of the scale were already provided including factor analysis for construct validity.
As a field of scholarly investigation, foreign language teaching anxiety had not kept pace with learning anxiety. Before the new millenium, foreign language teaching anxiety experienced by teachers had received only cursory research attention. One reason to be cited is that lack of an appropriate measuring instrument seems to have hampered the research on foreing language teaching anxiety for a considerable amount of time. Among the few work, Aslrasouli and Vahid (2014) conducted a study with 114 EFL university teachers in Iran context and used a scale to measure teacher anxiety developed by Ferguson, Frost and Hall (2012). The researchers found that 57.65% of the teachers were somehow experiencing teaching anxiety caused by various affective variables including interpersonal relations, employment structure as well as language proficieny and knowledge, as they report. The study also probed into the association between year of experience and teacher anxiety and found negative significant correlation (r = -0.21, p < 0.05), though the scale they used was not specific to FLTA. Correlation in this study can be considered weak as in social sciences correlation value over .40 is considered to have more practicality (Fraenkel & Wallen, 2006). The researchers also delved into the impact of gender variable on teaching anxiety by running a Point-biserial correlation and found no significant correlation (r = -0.11, p > 0.05).
Fairly recently, Öztürk (2016) undertook a research with 103 state university language instructors in Turkey. The study had a mixed-methods research design and for the quantitative data collection part the researcher used İpek's (2006) instrument FLTA scale. The findings of the study indicate, as Öztürk coins, a "moderate" level (M= 1.85) of FLTA experienced by the in-service university language instructors. The study also delved into the impact of variables such as gender and educational background which did not yield significant results while year of experience was found to have a significant impact on their anxiety level.
Concurrently, Aydın (2016) excavated the sources of foreign language teaching anxiety with 60 preservice teachers studying in DELT in a qualitative study. The researcher reached twelve topics that arouse anxiety in teachers ranging from lack of teaching experience, fear of making mistakes, lack of learners' motivation, participation interest and involvement in activities to personality type and difficulties in time management. Another finding emerged from the study was that pre-service teachers felt anxiety almost all the time; before, during, after, they were involved in teaching activities.
On the whole, looking across this sparse body of work, it is evident that scholarly world has turned a blind eye to the anxiety experienced by foreing language teachers. Although studies have lately recognised foreign language teaching anxiety, there is still an apparent dearth of systematic investigations into this area.

Purpose of the study
The main purpose of the study was to explore Foreign Language Teaching Anxiety levels of pre-service and in-service English language teachers in Trabzon and Yalova. Another purpose of the study was to investigate whether EFL teachers' FLTA level exhibit any difference with regard to some variables such as their gender, employment status (pre-service/in-service), the department they graduated from. The study also delved into whether there is a correlation between EFL teachers' FLTA level and their age as well as year of experience.
This study was designed to test out the research hypotheses presented below: 1. There is no significant difference in EFL teachers' foreign language teaching anxiety level with regard to their, a) gender b) pre-service and in-service status c) graduation department 2. There is no correlation between EFL teachers' foreign language teaching anxiety level and their a) age b) year of experience 2.Methodology

Participants
A convenience sampling method was exploited in the study as is well identified in the relevant literature that it is difficult to reach a true random sampling in social sciences (Dörnyei & Taguchi, 2009;Fraenkel & Wallen, 2006). The research population comprised of both in-service and pre-service EFL (English as a Foreign Language) teachers. A total of 90 in-service and pre-service EFL teachers in Yalova and Trabzon cities participated in the study. Forty-six of the participants were enrolled at or the graduates of the department of English language and literature (henceforth DELL), while 44 were enrolled at or the graduates of the department of English language teaching (henceforth DELT) in education faculties of various universities in Turkey. Fraenkel and Wallen (2006) remark that "The minimum acceptable sample size for a correlational study is considered by most researchers to be no less than 30" (p. 335). In line with this, Cohen, Manion and Morrison (2007) highlight that "causalcomparative and experimental methodologies require a sample size of no fewer than fifteen cases" (p. 102). Moreover, it is worth to note that a uniform curriculum has been followed in English Language Teaching departments of education faculties in all over Turkey since 1997 (YÖK, 1997), while English Language and Literature departments do not have such tight and strict curricula. What is more, preservice teachers from both departments take two semesters of practicum in their senior year. Thirty of the participants were in-service teachers working at various schools in Yalova and Trabzon while 60 of them were pre-service EFL teachers studying in faculty of education and faculty of letters at Karadeniz Technical University. The age range of in-service teachers was 22-50 with a mean of 29.9 (SD = 8.21) while the year of experience range was 1-25 with a mean of 6.5 years. The age range of preservice teachers was 21-30 with a mean of 23.41 (SD = 1.35) years.

Instrument
Foreign Language Teaching Anxiety Scale (FLTA) developed by Ipek (2006) was used to measure FLTA of participants as the main data collection instrument in the study. There were 26 items in the scale. Participants indicated the degree of anxiety they felt on a 5-point Likert type items questionnaire with 1= strongly disagree, 2=disagree, 3= neither agree nor disagree, 4= agree, 5= strongly agree items to choose from. Reliability and validity of the scale were statistically proved (İpek, 2006). Thus, the scale has been used in various studies and found highly reliable; in İpek's (2006) study high reliability was reached (α=.93). Similarly, Merç (2010) reached a high internal consistency coefficient level (α=.92), while Öztürk (2016) mentioned no reliability statistics. A Cronbach's Alpha Reliability test yielded a highly reliable level in this study (α=.92). Along with Foreign Language Teaching Anxiety Scale, participants were asked to express their gender, year of experience and the program, ELT or DELL, they graduated from in a demographic information form.

Procedure
The data were collected from two northern cities of Turkey; Trabzon and Yalova in the spring semester of 2017-18 academic year. The participants were handed out questionnaires by the researcher in person. No personal information was asked from the participants. Each participant was given a bar of chocolate as an incentive. The data were analyzed in Statistical Package for the Social Sciences (IBM SPSS ® , version 22.0). Descriptive statistics were generated for all variables and internal consistency reliability analysis was calculated using Cronbach's Alpha. The significance level was stated as a p value of <.05. Following the tests of normality, several parametric statistical tests: independent samples ttest and Spearman correlations were computed for inferential statistics as well as non-parametric Kruskal Wallis H test for smaller groups which had less than thirty participants (Larson-Hall, 2015).

Findings
Normality test of Shapiro-Wilk, which is considered as quite appropriate for small sample sizes (Larson-Hall, 2015), was computed and results revealed that the difference between the obtained data and normal distribution was not statistically significant (p =.055). To confirm normality, Skewness (= .21) and Kurtosis (= -.01) levels were inspected and found that normality was not violated. A general rule of thumb for Skewness and Kurtosis is that less than ±2 indicates normality is not violated (Brown, 1997;Weinberg & Abramowitz, 2002). Descriptive statistics were calculated and overall mean anxiety scores of participants were found to be moderately low (M = 2.42, SD = .6283). Furthermore, a onesample t-test was run to determine if a statistically significant difference existed between foreign language scores from the sample used in this study and the general population t(89), -8,74, p = ,000.
An independent samples t-test was computed to test the research hypothesis 1-a, if there was statistically significant difference in the foreign language teaching anxiety levels of the participants with regard to their gender. The null hypothesis was accepted at p = .172 which indicated that there was no statistically significant difference regarding participants' gender (Table 1). To test the research hypothesis 1-b, if there was statistically significant difference between in-service and pre-service EFL teachers' FLTA level, an independent samples t-test was run. The results revealed no significant difference (p = .872) with in-service teachers having slightly higher teaching anxiety level (Table 2). In pursuit of testing the research hypothesis 1-c regarding the department participants were enrolled or graduated from, an independent samples t-test was performed. Significant difference was not found (p =.589) between the graduates of DELL and DELT in general (Table 3). However, the table clearly demonstrates that teachers who were enrolled at or graduated from DELT had higher level of foreign language teaching anxiety.  (Field, 2009;Larson-Hall, 2015), indicated that the level of FLTA of DELT graduate in-service teachers was significantly lower than DELT enrolled pre-service teachers (U = 133.00, z = -2.47, p < .05) and ELT enrolled pre-service teachers (U = 70.50, z = -3.91, p < .05). However, the anxiety level of DELT graduate in-service teachers was not significantly different from DELL graduate in-service teachers. The research hypothesis 2-a that there is no correlation between EFL teachers' FLTA level and their age was tested by utilizing Spearman correlation. The result of the Spearman correlation indicated a negative statistically significant association between FLTA and participants' age (r = -.23, N = 90, p = .03), demonstrating that the older the participants the lower anxiety they experience. As a rule of thumb, correlations to be considered meaningful are expected to be no less than .30 in social science research (Cohen, Manion & Morrison, 2007;Dörnyei & Taguchi, 2009) and Fraenkel and Wallen (2006) consider correlation between .41 and .60 as "large enough to be of practical as well as theoretical use" (p. 249).
Spearman's rank-order correlation was computed in order to assess the research hypothesis 2-b that there is no correlation between EFL teachers' FLTA level and their year of experience. The null hypothesis was rejected as the result identified a statistically significant negative correlation between the teachers' FLTA level and their year of experience (r = -.35, N = 90, p =.001). Put differently, the more experienced the teachers get the less teaching anxiety they experience.

Discussion
Preliminary descriptive statistics, which, were calculated to examine central tendencies in the study, revealed a relatively low foreign language teaching anxiety level for the participants (M = 2.42, SD = .6283). However, when compared with the results of Öztürk's (2016) study, who used the same scale and found mean score of "1.85" with a standard deviation of .56, the participants in this study had, if not suffer from, a considerably higher level of FLTA (Foreign Language Teaching Anxiety).
To explore any gender-related difference in participants' foreign language teaching anxiety levels, an independent samples t-test was implemented. The t-test results revealed no significant difference for gender. However, as in line with the extant literature (Rubin, Slovin, & Krochak, 1988) female participants found to have higher anxiety levels. There is a general tendency that females experience, particularly in self-reported studies of various fields, more anxiety than males. As the research on the impact of gender has yielded controversies, this result is in line with some literature where no significant difference, regarding, gender was found (Akinsola, 2014;Aslrasouli & Vahid, 2014;Öztürk, 2016), while it is in contradiction with some other works where significant differences were found (Ameen, Guffey, & Jackson, 2002;Fish & Fraser, 2001;Ngidi & Sibaya, 2003). However, it should be noted that females' high anxiety level in self-reported studies might result from females' general tendency to express and share their feelings while males generally have a tendency to hide their feelings, particularly their weaknesses.
Another independent samples t-test was carried out explore whether there was significant difference between pre-service and in-service EFL teachers' FLTA scores. Although in-service teachers had slightly higher levels of FLTA, the difference did not reach statistically significant level. This result is in partial contrast with existing literature which demonstrates that as the teachers get experienced their anxiety level decreases (Akinsola, 2014;Ameen, Guffy, & Jackson, 2002;Aslrasouli & Vahid, 2014;Fish & Fraser, 2003;Gardner & Leak, 1994;Kesen & Aydın, 2014;Öztürk, 2016). However, the result may be due to the fact that there are some novice in-service teachers within the research sample of this study. Moreover, this finding can be interpreted as a cause of anxiety experienced due to supervision process (Kayaoğlu & Kobul, 2013) as being supervised per se is widely accepted as an anxiety-provoking situation (Akinsola, 2014;Gebhard, 1990).
It was hypothesized that there might be a difference in participants' FLTA levels with regard to the department they were attending or graduated from. No statistically significant difference emerged with respect to the department; either English language teaching or English language and literature. However, further test of Kruskal Wallis H test and Post Hoc Mann-Whitney U tests revealed that participants who were graduates or enrolled in the English language teaching department had higher FLTA levels than of department of English language and literature. To the researchers' best knowledge, no published study systematically investigated and documented the difference in teaching anxiety levels of pre-service and in-service teachers. This result might have been obtained due to the folkwisdom of "ignorance is bliss" which refers in this context the courses taken within the language teaching department curriculum. From the first year on, the participants follow a curriculum which offers teaching-focused courses. There is a uniform curriculum followed in English Language Teaching departments of education faculties accross Turkey since 1997 (YÖK, 1997). These teaching courses, either theoretical or practical, offered participants to take throughout their bachelor's program might be creating some kind of self-monitoring effect (Synder, 1974) or false consensus effect (Ross, Greene & House, 1977) that would affect them negatively, i.e. lowering their perceived self-efficacy beliefs (Bandura, 1993;Pajares, 1996) or causing some kind of anxiety. It is well-documented that when people know less about something they have a tendency to a have an inflated self-assessment (Kruger & Dunning, 1999) and thus, experience less cognitive inhibition (Wood, Matthews, & Dalgleish, 2001). Thus, the higher anxiety that the DELT participants experience can be attributed to their cognitive load which might be serving as a burden (Kalyuga, Ayres, Chandler, Sweller, 2003) because they are trying to handle too much information at a time or some kind of psychological inhibition or sheer anxiety as in tunnel vision (Savage, Potter, & Tatler, 2013;Smith, 2004) for the participants. In line with this, Dadandı, Kalyoncu and Yazıcı (2016), in a study conducted with 677 pre-service teachers of various majors, found that students from teaching departments had higher concern levels than the students from departments whose students take these teaching classes only in their fourth and last year under the name of "pedagogic formation" program in Turkey.
The results of this study have identified a negative and significant correlation between teachers' age and FLTA based upon a Spearman's rank-order correlation. This result resonates with relevant literature which well documents that there is an association between anxiety level and age (Ameen, Guffy, & Jackson, 2002;Fish & Fraser, 2001;Gardner & Leak, 1994;Ngidi & Sibaya, 2003).
Another finding emerged from this study was a moderately strong negative association between participants' FLTA level and year of experience which suggests that the more experience the teachers get the less teaching anxiety they experience. This result corroborates the findings of existing literature (Aslrasouli & Vahid, 2014;Ameen, Guffy, & Jackson, 2002;Fish & Fraser, 2001;Gardner & Leak, 1994) while it contradicts with Kesen and Aydın's (2014) study where they found that less experienced instructors were reported to have lower level of FLTA than experienced college instructors. The difference might be lying on the difference in research sample; teachers working at primary or secondary level education versus tertiary level college instructors. Teachers working at different education level schools might have different aims, objectives, levels, needs, interests, concerns and background etc.

Conclusion
The current study was undertaken mainly to investigate the FLTA level of pre-service and in-service non-native English language teachers. Another purpose of the study was to probe into any difference in FLTA level of participants regarding their gender, employment status and the department they were enrolled at or graduated: department of English language teaching or department of English language and literature.
The results of the study clearly indicate that, regardless of pre/in-service status, English language teachers self-report to experience FLTA, at least to a certain degree. This result is in accordance with the extant literature (Ameen, Guffy, & Jackson, 2002;Bernstein, 1983;Borg, Riding, & Falzon, 1991;Bowers, Eicher, & Sacks, 1983;Fish & Fraser, 2001;Gardner & Leak, 1994;Ngidi & Sibaya, 2003;Wadlington, Slaton, & Partridge, 1998). However, the study revealed no significant difference in participants' FLTA level with regard to gender. Although this finding differs from some published work where significant differences were reported (Ameen, Guffey, & Jackson, 2002;Fish & Fraser, 2001;Ngidi & Sibaya, 2003), it substantiates some other research where no significant difference was found (Akinsola, 2014;Aslrasouli & Vahid, 2014). This demonstrates that the impact of gender on teaching anxiety will continue to stay at the forefront of controversy in the foreseeable future. Conjointly, further research with different and wider samples would contribute substantially to enhance our understanding of the role of gender in FLTA.
One of the most noteworthy findings of this study is that no statistically significant difference was found between pre-service and in-service teachers' FLTA levels though in-service teachers were found to have less FLTA level compared to pre-service teachers. However, a Kruskal-Wallis H test yielded statistically significant difference between the groups when they were grouped according to their employment (pre/in-service) status and enrolled/graduation department (χ 2 = 19.34, df = 3, p = .000) separately. Post hoc pairwise comparisons revealed that in-service DELT graduate teachers were found to be experiencing less FLTA than DELL graduate teachers. Moreover, pre-service teachers enrolled at both DELT and DELL departments were found to have more FLTA level compared to in-service DELL and DELT graduate teachers.
Another remarkable finding of the study was that the result of Spearman rank-order correlation indicated a significant moderate negative correlation between teachers' FLTA level and their age (r = -.23, N = 90, p = .03). Put it differently, the older the teachers get the less anxiety they experience. This finding supports further evidence to the earlier research findings (Ameen, Guffy, & Jackson, 2002;Fish & Fraser, 2001;Gardner & Leak, 1994;Ngidi & Sibaya, 2003).
The current study also identified, grounded on Spearman's rank-order correlation, a significant moderate negative relationship between teachers' FLTA level and year of experience (r = -.35, N = 90, p =.001). This finding corroborates the findings of a previous studies which have long established the negative correlation between teaching anxiety and year of experience (Akinsola, 2014;Aslrasouli, & Vahid, 2014;Fish & Fraser, 2001;Klassen & Chiu, 2010). The common sense has it that as the teachers get experience they get more familiar with anxiety-inducing situations such as disruptive student behavior, classroom management, heavy work load, following the curriculum, planning lessons, being supervised and can cope with them more effectively (Akinsola, 2014;Fish & Fraser, 2001;Gebhard, 1990;Kesen & Aydın, 2014). Thus, the results provide further support to the Latin proverb "Usus magister est optimus" which means practice makes perfect.

Implications
The results obtained from this study have the following pedagogical as well as practical implications: Foreign Language Teaching Anxiety per se should be acknowledged as an anxiety arousing experience for EFL teachers. Awareness should be created in teachers and teacher trainers and some further precautions can even be taken. In the broadest perspective, the curricula of pre-and in-service training programs might be suggested to include topics that would raise awareness of FLTA nation-wide. Put differently, EFL teachers as well as teacher trainers might be suggested to be trained in the sources, causes, effects and results of FLTA. Accordingly, methods, techniques or procedures such as coping strategies for alleviating anxiety should be recommended and included in curricula for better and more effective teaching process. For instance, reflective teaching (Richards & Lockhart, 1994) and clinical supervision (Gebhard, 1990;Kayaoğlu & Kobul, 2013) can be recommended as feasible alternatives for coping strategies as they are already acclaimed to improve teachers' affective states as well as teaching process. Notwithstanding, empirically grounded research can be further conducted to investigate the outcomes of reflective teaching and clinical supervision in alleviating EFL teachers' teaching anxiety level.
Some other potentially anxiety-provoking factors or situations might have gone unnoticed within the FLTAS. It can, thus, be suggested that more insights can be gained through in-depth individual and/or focus group interviews with the teachers pertaining to their additional anxiety sources/reasons. More broadly, future research conducted in mixed method design (Fraenkel & Wallen, 2006) studies would be of great contribution to excavate more complex and intricate data as well as more comprehensive identification of FLTA.
Self-efficacy belief can also be a factor that might have influence on teaching anxiety. Thence, a study probing into the correlation between self-efficacy and foreign language anxiety might help to gain a more accurate picture of the mechanisms underlying teacher psychology.
There is abundant room for more investigations that will delve into the impact of native versus nonnative teachers dichotomy on FLTA. Put more explicitly, whether FLTA levels of native and non-native teachers reveal significant differences remain unanswered at present. It can be further suggested that the difference might be further investigated including pre-service in-service dichotomy.
Another suggestion for further research pertains to the DELT and DELL pre-service teachers' anxiety level. A longitudinal follow-up study which will investigate whether there will be any change in FLTA levels of the same participants in their career, is thought to be of great contribution to the field.
Last but not least, the present study features some limitations that merit mention. First, limitation of this study is that the data were collected from two cities of Turkey. In line with this, another limitation that needs to be acknowledged is the data were collected from only pre-service teachers studying in Trabzon while in-service teachers were from both Yalova and Trabzon. However, it would be rather untenable to conclude that the data were thus invalid. Our results are encouraging however, it would be better to be validated by research with larger samples from various regions of Turkey and with mixed method research design, which is believed to provide more vivid picture of the situation.