Analyzing Foreign Language Test Anxiety among High School Students in an EFL Context ( Note 1 )

With the increasing need to learn languages as a result of globalization there is a great demand on the part of the learners to communicate in a second/foreign language, which is also supported downwards by the governments and upwards by the parents. Among the many aspects of foreign language learning, affective factors are researched a lot as they are dependent on contexts, individual differences, cultural background, teaching methodology etc., which cause a variation in the results. The current research focuses on test anxiety as one of the major affective factors. Thus it aims to identify the level of test anxiety and its relationship with gender, grade level, and academic achievement. Moreover, the causes of test anxiety were investigated according to students’ own perceptions. A test anxiety scale and semi-structured interviews were conducted to gather the qualitative and quantitative data. The overall results showed that the participants had a moderate level of test anxiety. Females were found to be more anxious than males only in some aspects; low achievement scores provoked test anxiety with regard to a few items, and 9 graders were found to be more anxious than the 10 graders. According to participants’ own perceptions, test validity, time limit, teacher attitudes, test techniques, proctors, length of the test, testing environment and clarity of test instructions were the causes of test anxiety.


Introduction
Internationalization and globalization are two accompanying terms for the world in recent years.Being the lingua franca, the efforts to teach English as a Foreign Language (EFL) gained speed and displayed itself with a change in the curricula of most countries toward teaching even younger learners in primary school (Gürsoy & Akın, 2013).Language learning is a process including linguistic and nonlinguistic elements that consists of cognitive, metacognitive and affective factors (Amiri & Ghonsooly, 2015;Gürsoy & Akin, 2013;Zafar & Meenakshi, 2012).Among these components, anxiety is one of the affective factors that has an effect in language learning.Test anxiety, being a part of language learning anxiety, is one of the important determinants of student success (Aydın & Zengin, 2008;Brown, 2000).Although test anxiety is an important component of language learning, the number of research conducted on the issue is limited to make effective assumptions.Thus this study incorporates to this gap in the literature by concentrating on test anxiety in an EFL context by underlying the need to develop more student appropriate assessment.From this angle, designed as a mixed-method research, the current study aims at identifying the level of test anxiety among the learners of a state vocational high school in a large city in Turkey and investigating the relationships between text anxiety and achievement scores, grades and gender.Finally, it identifies the reasons of test anxiety based on the participants own perceptions.

Background
Foreign language learning is a multilayered process affected by individual differences such as learners' language aptitude, beliefs, motivation levels, expectations, affective agents etc. (Aydın & Zengin, 2008).Foreign Language Learning Anxiety (FLLA) is one of these affective factors that has a dynamic impact on this process.Test anxiety, as an important component of FLLA can develop due to several reasons such as the contextual factors, personal factors, and learners' previous experiences.Turkey is an exam-oriented country in which the Turkish learners experience test anxiety at different levels due to this context-specific situation.This is mostly because of the high-stake examinations such as "University Entrance Examination" and/or "Transition from Primary to Secondary Education Exam" (Ergene, 2011) that students enter at the end of the secondary school and/or high school.The majority of Turkish students take these exams and work for these exams during their education, which paves the way for the transfer of the affect to different courses at different grades.To pass these high-stake exams is a life changing experience for the learners, hence they are mostly associated with the feeling of anxiety.
Anxiety, being a complex term with many components, has been defined as "an unpleasant emotional state or condition that is characterized by the subjective feeling of tension, apprehension, and worry, and by activation or arousal of the automatic nervous system" (Spielberger, 1972, p. 482).According to Gass and Selinker, anxiety may be "a matter of personality, an emotional reaction to a situation, or a combination" (2008, p. 400).Hence, it can be divided as trait, state, and situation-specific anxiety.
The trait anxiety is caused when an individual tends to perceive the stressful conditions as hazardous or threatening (Cao & Liu, 2015).State anxiety, on the other hand, can be defined as the feelings of nervousness and worry at the very moment of the triggering situation (Cao & Liu, 2015); and the last one, situation-specific anxiety, is the one felt for a particular situation (Maclntyre & Gardner, 1991).
From another perspective, it is also possible to group anxiety in two different forms; facilitative and debilitative anxiety as Scovel stated in 1978.In fact, the names represent their meanings.As Amiri and Ghonsooly (2015) explained in their study; the facilitative anxiety is helpful for a language learner, on the other hand, debilitative anxiety hinders success.
Given these definitions, foreign language learning anxiety can be classified as a situation-specific anxiety (Aydın, Yavuz, & Yeşilyurt, 2006) and is defined as "a distinct complex of self-perceptions, beliefs, feelings, and behaviors related to classroom language learning arising from the uniqueness of the language learning process" (Horwitz, Horwitz, & Cope, 1986, p. 128).Horwitz et al. (1986) presented three components of language learning anxiety: communication apprehension focuses on interpersonal interactions; fear of negative evaluation is the feeling of fear for evaluation by others and avoidance of the evaluative conditions; test anxiety is the anticipation of a failure in academic evaluation.
As an indispensable part of every learning environment, assessment can cause anxiety to many learners.Using assessment tools like quizzes, examinations, reviews or observations can be a source of test anxiety.Although the term has been defined by many researchers (see, Aydın & Zengin, 2008;Horwitz et al., 1986;Sarason, 1984;Young, 1991).Zeidner and Mathews describe test anxiety as behavioral, phonological and psychological reactions to the evaluative situations such as exams or tests in relation to the negative effects or anticipation of these items (2005), while Dusek (1980) defines test anxiety as the displeasing apprehension or psychological condition that has physiological and behavioral characteristics which are experienced at the time of formal testing or in other evaluative conditions.
There are many factors that might trigger test anxiety.Some of these are described in the study of Young (1999).He stated that these factors for the learners are test validity, test techniques, length of the test, time limit, clarity of test instructions, and the testing environment.He stated that the students considered the test invalid if they felt highly stressed in the tests.Similarly, Ohata (2005) pointed out the pressure on students caused by the time limit.In an earlier study, Young (1991) reported on the importance of students' familiarity with the test format and the questions as another factor for test anxiety.
The relationship between the test anxiety and gender has been studied by many researchers.However, the results in these studies are conflicting.In some studies, females are found to be more anxious than males (see, Chang 1997;Daly, Kreiser, & Roghaar, 1994;Felson & Trudeau, 1991;Hembree, 1988;Lowe & Lee, 2008;Lowe, 2015;McDonald, 2001).In other studies, such as Aydın and Takkaç (2007) the relationship between gender and test anxiety showed a minimal difference and it changed mutually according to the situations; for some of the items males scored higher than females, while, for other items females' scored higher.
Another supposed factor for the researchers is the academic achievement.According to Collins and O'Brein (2011), academic achievement is "the attainment of knowledge, competencies, and higher-level status, as reflected in grades, degrees, and other forms of certification or public acknowledgment" (p.3).Most of the research conducted to identify the relationship between test anxiety and academic achievement, showed a negative correlation between them (Aydın, Yavuz, & Yeşilyurt, 2006;Ergene, 2011;Hill & Wigfield, 1984;Marcos-Llinás & Garau, 2009;Şener, 2015;Tugan, 2015).Accordingly, if a learner was highly anxious, it lowered the academic achievement.
In order to understand the variables of language learning in a broader sense, the related literature has also investigated the relationship between language anxiety and grade level.There is only a few research that focused on language anxiety in terms of students' grade level (Aydın, 2013;Er, 2015;Gürsoy & Akın, 2013).Grade level of the students in the Turkish context is important due to the differences in program, content, preparation for national examinations, etc.As these sub-factors could affect the students at different grade levels it is possible that students' test anxiety is affected by them.

Methodology
The present study was conducted at a Vocational High School in which the students' primary concern is not English, but other vocational lessons.This study was implemented with the 9 th and 10 th grades particularly in an attempt to see the possible effects of the Transition from Primary to Secondary Education Examination on the 9 th grades.A mixed-method research design was used in the study.Qualitative and quantitative data were collected to answer the following questions: 1) What is the level of test anxiety among 9 th and 10 th grade students who learn EFL at a state vocational high school in Bursa?
2) Does gender, course grades, or grade level (being 9 th or 10 th grade) have an effect on EFL students' test anxiety?
3) What are the causes of test anxiety according to the participants' own perceptions?

Participants
The participants in the study consisted of 138 vocational high school students selected via non-random convenient sampling.77 (55.7%) were female and 61 (44.2%) were male students.The group also included 54 (39.1%) 9 th grade students and 84 (60.9%) 10 th grade students.

Data Collection
Both qualitative and quantitative data were collected in the study.Quantitative data were collected via a questionnaire and analyzed statistically by using SPSS 20.The qualitative data were gathered with the help of semi-structured interviews.Systematic random sampling was applied while choosing the participants for the interview.This type of sampling was especially applied in order to prevent subjectivity and increase the validity of the data.Every 4 th participant in the classes that the study took place was included in the interview.The interviews were conducted with 15 males and 19 females (N=34).14 of the participants were 9 th grades while 20 of them were 10 th grades.

Instruments
The instrument used in the study was a test anxiety scale that was developed by Sarason in 1984 and is a valid scale that has also been used by many researchers (Aydın et al., 2006;Aydın & Takkaç, 2007;Aydın & Zengin, 2008;Şener, 2015).The instrument has two parts.The demographic information of the participants was collected in the first part and included 3 questions related to gender, course grades and grade level of the students.The second part consisted of 22 questions and aimed to find out the test anxiety level of the participants.In this part the students were asked to indicate how frequently they feel anxious in the given statements (1: never, 2: rarely, 3: sometimes, 4: often, 5: always).The content and face validity of the instrument was established in different studies that used the same instrument (Aydın et al., 2006;Aydın & Takkaç, 2007).As the next step, the instrument was piloted on 15 10 th grade students from the same school.Lastly, it was used with the participants and found highly reliable with the 0.86 alpha value.
The interview, provided the qualitative data for the study and conducted to support the findings from the quantitative part.34 participants were asked three open ended questions in the interviews.The interviews lasted 10-15 minutes and the researchers kept field notes on a template in order not to miss any valuable data.

Data Analysis
The data analysis process was done in two different steps.As the first step, SPSS software program was used to analyze the quantitative part of the study.The reliability coefficient, the mean scores, frequencies were found.The results of the Levene's test showed that the data were normally distributed.As a result of the homogeneity of the variances, parametric tests were implemented.Dual comparisons were made by using T-tests to find the effects of some variables such as gender, grade and achievement scores on students' anxiety level.As a second step, the interviews were content analyzed by identifying recurring themes in the data and putting those into groups and sub-groups.

Results
In order to answer the first research question that aims to identify the anxiety level of the students', frequency analysis was conducted and means were calculated.According to the results of the SPSS analysis, the mean of the responses given to the items was M=2.62, which indicates a moderate level of anxiety.Table 1 presents the results of frequency analysis of the questionnaire.Although the anxiety level of the students was generally moderate in the total number of participants from 9 th and 10 th grades, six of the items had higher means than the general average of the scale.Accordingly, more than half of the students believed that they could have done better after taking the test (53.6%).The participants claimed that if they felt they were not ready for the tests, their anxiety level increased relatively (41.3%); during the test, they had a difficulty about remembering the knowledge that they had already gained before the exam (33.3%); they felt crammed before a test (35.5%); and tests were bothering (31.1%).Another important factor for the participants was the time limit.The participants stated that they felt under pressure because of the time limits in a test (27.6%).The other scores in the scale showed that most of the participants did not have a high level of anxiety.
The interview results supported the answers given in the questionnaire.The first question aimed to find out if they ever had test anxiety.All of the participants stated "Yes" for the first question in the interview.None of the participants stated "No" which shows that they had experienced test anxiety at least once during their learning experience.
The second interview question aimed to find out the frequency of their test anxiety.16 of the participants stated that they "rarely" felt anxious about English tests.Nine answered the 2 nd question as "sometimes", six participants described the frequency of their test anxiety as "often" and only 2 participants stated that they were "always" anxious about English tests.These results were also reflected in the results of the test anxiety scale (M=2.62).
Finally, the third interview question addressed the participants' own evaluations for the reasons of test anxiety.
The most common answer compiled from the interviews was about the instructions.The participants (16 out of 34) stated that they had difficulty in understanding the instructions.Interviewee 7 (I7) stated that "most of the time I do not read the instructions written above the questions because I think they are more difficult than the questions in the exam".13 stated that they felt anxious when they do not study enough, which overlapped with the item 20 in the test anxiety scale.12 participants stated insistently that, before and during English exams, even if they were prepared for the exam, they felt confused, nervous and stressed.These responses fit the items 1, 3, 11, 13 and 21 of the test anxiety scale as well.10 participants also expressed that after English exams, they still had worries because they couldn't remember what they had done in the examination.These statements are also related to the items 11 and 14 in the scale.The interview revealed another interesting fact about the exam proctors.Since the proctors did not know English, they could not answer the questions of the participants during the exam.Two participants also stated that in written English tests, some of the proctors made fun of the participants' level of English.This was also found out to be another cause for their test anxiety.Time limit was also another factor that caused anxiety among the participants (8 out of 34), which is relevant to the item 6 in the scale.I22 explained that not for the other exams but English, "I can feel the tick tocks of the clock in my heart".6 participants told if they saw a new word or question type in the exam they had never seen before, these made them worried and caused apprehension.I17 explained it like "if I see a new word in the exam, I forget all the other words I know and cannot focus on the exam".This was also revealed in the items 1, 4, 10, and 13 of the scale.Finally, the length of the exam was also included in four of the participants' answers.These participants told that they would feel safe if the content was less in English exams.

Test Anxiety and Gender
The second research question aimed to answer if there was a significant relationship between test anxiety level of the participants and their genders.The values in Table 2 indicate that 6 items of the scale showed significant differences between test anxiety and gender.According to the results of the t-test, female students seemed to have more test anxiety than male students in all these 6 items.

Test Anxiety and Course Grades
With the aim of answering the second research question, the relationship between test anxiety level of participants and their course grades was analyzed (Table 3).The overall scores of the participants from the English lesson were used in the analysis.For students to be considered successful from any lesson including English, they need to get 50 over 100 in Turkish education system.Thus the course grades were grouped as "0-49" (unsuccessful) and "50-100" (successful).An independent samples t-test was conducted to compare course grades and test anxiety.According to the findings, 3 items in the test anxiety scale showed statistically significant results with the course grades of the participants.These items (10, 15 and 19) were about how often their performance was affected by their emotions, how often they were bothered because of the tests and how often the test affected their performance.
The groups with lower scores (scores between 0-49) said that during a test their emotions affected their exam performance in a negative way, while the group with the higher scores (scores between 50-100) did not seem to have much difficulty about controlling their emotions and anxiety during an exam (item 10).Item 15 showed that the group with lower scores felt disturbed because of the exams while the participants with the higher scores did not get bothered as much as the lower group did.Finally, item 19 suggested that the participants with the lower scores believed that tests had a negative impact on their performance, whereas the participants with the higher scores did not believe that test affected their performance negatively so much.
The second question of the interview was related to the frequency of the test anxiety as well.Students in the interview were grouped according to their achievement scores.There were 15 participants between the 0-49 scores in the interview accordingly, 2 interviewees (Is) "always", 2 Is "often", 7 Is "sometimes" and 4 Is "rarely" experience anxiety.However, 12 participants that had grades between 50-100 scores told they "rarely", 2 Is "sometimes", and 5 Is "often" feel anxious.The results indicate that the lower achievers experience anxiety more often than the higher achievers.

Test Anxiety and Students' Grade Level
According to the values in Table 4, six items of the scale showed significant relationships with the participants' grade levels.The results of the t-test indicated that the participants in 9 th grade seemed to experience anxiety more than the 10 th grades.When compared to the 10 th grades, 9 th graders felt more confused before the exam, besides, they also worried more during the exam.Additionally, students at the 9 th grade were also more likely to be affected by their feelings negatively and said that they easily forget what they knew while taking an exam.In contrast to this, students at 10 th grade seemed to control their nervousness and were more secure about what they knew while taking an exam.Lastly, when they were ready for the exam, 9 th grade students felt test anxiety more than the 10 th grades did.Likely, 10 th grade students felt less worried and anxious even when they were not prepared enough for the exam than the 9 th grade students did.In the interviews, there were 20 10 th grade students and 14 9 th grades students.Among the 14 9 th grade participants, 5 of them said they were "rarely" anxious, 2 of them said they "sometimes" felt worried, 6 of them said, they "often" felt the tension and 1 of them said she was "always" anxious.Among 20 10 th grade participants, 11 of them said they got "rarely" worried, 7 of them said "sometimes" they got anxious, 1 of them said "often" and another 1 out of 20 said "always".This information obtained from the interviews was also in agreement with the test anxiety scale.

Discussion
Having an exam-oriented education system in Turkey, teachers and students rely mostly on test results when determining the success of the learners rather than using alternative forms of assessment.At different grade levels students take centralized exams either to get into better high schools or university.These high-stake examinations forces teachers and students to work for the exam, putting the exam results in the center of education.Foreign language learning as one of the compulsory courses of the curriculum have its share in this environment with teachers focusing on accuracy rather than the communicative aspect of the language.This situation may cause learners to transfer their anxiety to the foreign language classroom, which itself already have its own factors that might cause apprehension.Although the present study shows that the students have a moderate level of test anxiety, academic achievement, gender differences and grade levels as variables are important factors that affect test anxiety level which can be a result of the aforementioned educational context.
According to Amiri and Ghonsooly (2015), English classroom anxiety is much more significant than any other anxiety type since learning a language is a multi-dimensional process and could be stressful; tests, being in an evaluative situation, could increase this stress.One way to reduce the test anxiety is to develop skills in assessment and measurement.The validity of the test plays an important role in learners' anxiety level.In addition to the testing skills, the learning environment such as the rapport between the students and the teacher, classroom participation, and creating a risk-taking environment (Salehi & Marefat, 2014) are crucial factors in reducing the affective filter.
The overall results of the current study revealed that the students in a state vocational high school in a large city in Turkey had moderate level of test anxiety.However, when dual comparisons were made statistically significant differences were found in certain aspects of test anxiety.The results showed that the students feel confident most of the time if they had good grades in the exam.They believed that having a test was a necessity.In other respects, students believed that not being well prepared for the exam increased their anxiety, they thought they were confused before the examinations and unfortunately, in their perspective, they felt they could do much better after the test.
When the relationship between test anxiety level and gender was analyzed the study had similar results with that of Aydın (2013), so that females were found to be more anxious only in some aspects in the present study as well.
It was found that in contrast with male learners, female learners mostly compared themselves with the other learners, whether they were well prepared or not, they were still anxious about the exams and got more confused during the examinations, they were also more sensitive to the time limits.This result was consistent with the other studies in the literature (see, Chang, 1997;Daly, Kreiser, & Roghaar, 1994;Felson & Trudeau, 1991;Hembree, 1988;Lowe & Lee, 2008;Lowe, 2015;McDonald, 2001).
Additionally, test anxiety was found to be significantly related to academic achievement with regard to a few factors/items.The findings indicated that the higher achievers were experiencing test anxiety less than the lower achievers.The more the scores were higher, the less the students worried about the tests.The result of the study was in agreement with the literature in terms of academic achievement (see, Aydın, Yavuz, & Yeşilyurt, 2006;Ergene, 2011;Hill & Wigfield, 1984;Marcos-Llinás & Garau, 2009;Şener, 2015;Tugan, 2015).The findings proved that academic achievement was an important factor for test anxiety in some aspects of the construct.The learners feel more secure if they know that they can be successful.According to Tugan (2015), test anxiety is an obstacle for excellent academic achievement and prevents learners from acquiring consistent learning capabilities.This fact reveals that lower achievers may need more guidance before the tests and the testing conditions as well as the test itself should be well-designed.In terms of guidance, the instructions used in the exams need to be used, exemplified, and/or demonstrated in the classroom as well to familiarize the students with them, as in the interview they complained about the fact that they cannot comprehend them.In terms of the testing conditions the proctors can be given clear instructions about how to conduct the exam.If possible, language teachers could be assigned as proctors to help students with their questions.Moreover, they can be informed about how a proctor's attitudes towards students affect their exam performances.Another important factor is the content of the exam; validity of the exam needs to be considered carefully.Lastly, in order to help learners feel more confident before and during the exam, the study skills should be well-founded by the students with the help of the teachers.
The analysis of the relationship between test anxiety and grade indicated some significant differences.9 th graders felt more anxious than 10 th grade students did.During the examinations, 9 th grade students had more negative apprehension and sometimes felt confused before the examinations.This finding is opposite the other studies in the literature.Conducting their studies with children, Er (2015) and Gürsoy and Akın (2013) concluded that the older the students get toward adolescence, the more language anxiety they have.However, with older learners as in the present study the reason why 9 th graders are more anxious than 10 th grades may have some other explanations such as the context.The centralized exam called "Transition from Primary to Secondary Education Exam" in the Turkish education context is a high-stake test as a result of which the learners are assigned to high schools according to their scores after the 8 th grade.Most of the time this exam is perceived as a race to get into better schools.Higher achievers get into Science Schools or Anatolian High Schools, whereas lower achievers are enrolled in Vocational High Schools.Therefore, it is possible that the 9 th graders might have transferred their recent anxiety-raising experience to the language learning classroom.The reflection of learners' feelings caused as a result of this exam might be a reason for this context and participants at 9 th grade to feel more anxious than their older counterparts, which needs to be examined in the future studies in the same context.
The interviews aimed to identify the participants' own perceptions regarding the reasons causing test anxiety.Their attributions were pointing mostly to the factors other than their selves such as the time limit, content of the course, proctors who did not know English, unclear instructions and validity of the tests.The students felt pressured and lost concentration in the test environment because of the proctors' behaviors or a new question type that they did not come across with before.What's more the instructions made learners feel stressed and the time limit and the length of the examinations prevented students focusing on the content of the test.These facts are also in line with the literature (Ohata, 2005;Young, 1991).One way to reduce the instruction-related test anxiety might be to use similar instructions in the classroom.In addition, the use of English when communicating in the classroom and giving instructions in L2 would increase exposure and input, moreover, it can increase the familiarity of the students with the chunks used in questions.

Conclusion
The overall results of the study indicate that the participants of the study have moderate test anxiety.However, dual comparisons showed significant differences between participants according to some variables such as grade level, gender and achievement.Thus 9 th graders were found to be more anxious than 10 th graders; females than males and lower achievers than higher achievers in certain aspects of the construct.The education system as well as the examination system might have an impact on students' test anxiety.What's more, factors related to the teacher such as the teaching methodology, the use of L2 in the classroom, relevance of tests to classroom practices, etc., might play a role on the results.As the participants stated in the study, the most important determiner of test anxiety is the instructions given in the test, which might be related to teacher's methodology and use of L2 in the classes.The other affective factors stated in terms of test anxiety are content validity, test taking environment, proctors, time limit and the length of the test.Therefore, further studies in similar contexts can be conducted to identify the effects of teacher-related factors on test anxiety.
The results have implications for language teachers and policy makers.Increased awareness on our teaching and reflective practice might reduce the negative outcomes of the teaching-learning process.Therefore, it is suggested that teachers are engaged in Continuous Professional Development to bridge the gap between theoretical knowledge gained in teacher education institutions with practical knowledge developed in schools.Moreover, the policy makers can overview the current examination system to reduce the negative effects of the system on students' affect.
A few limitations of the study must be noted.Since the data were collected from a single type school, it is difficult to generalize the findings to other teaching environments.Thus further studies need to be conducted in similar contexts to compare contextual differences deriving from different school types.Additionally, it should be noted that, other factors such as social, cultural, political and psychological agents must be taken into account by the forthcoming studies.

Table 1 .
The test anxiety level of the participants

Table 2 .
The relationship between test anxiety and gender

Table 3 .
The relationship between test anxiety and course grades

Table 4 .
The relationship between test anxiety and students' grade levels