AGE , GENDER AND GRADE EFFECT ON FOREIGN LANGUAGE ANXIETY AMONG CHILDREN

The number of studies on the relationship between certain demographic variables, age, gender and grade, and the levels of foreign language anxiety (FLA) in the English as a foreign language (EFL) context regarding adult and young learners is fairly limited, whereas the findings obtained from prior studies did not reach a consensus on the influences of FLA, and studies mainly focused on adult language learners rather than young learners. This study aims to explore the effect of age, gender and grade differences on FLA and its types, communication apprehension, fear of negative evaluation and test anxiety, among Turkish children who learn English as a foreign language. The participants were 494 children enrolled at primary and secondary schools. The data collection tools were a background questionnaire interrogating the variables mentioned and the Children’s Foreign Language Anxiety Scale (CFLAS). After computing the reliability coefficient of the scale and total variance, independent samples t-test and ANOVA were used to see the 134 TEFLIN Journal, Volume 28, Number 2, July 2017 relationships between the levels of FLA and its types and subject variables, age, gender, and grade. The results showed that age, gender, and grade constitute significant differences regarding FLA and its types, when several items in the CFLAS were considered. In the light of findings, several practical recommendations are listed.

While foreign language anxiety is a considerable issue that has attracted much attention in an EFL research context, no consensus has been obtained in terms of its causes and effects.Research also mainly focused on anxiety among adult EFL learners, while children have drawn little attention in the mentioned research context.However, it is clear that foreign language learning during childhood facilitates target language proficiency (Cameron, 2003;Lightbown, Spada, Ranta, & Rand, 2006).Thus, it is evident that research is strongly necessary for understanding the causes and debilitating and facilitating effects of FLA before adolescence (Lenneberg, 1967).In this scope, research lacks how FLA affects the foreign language learning process during childhood.Speaking more specifically, there are no data on the relationship between certain demographic variables and the levels of FLA among children.
In the Turkish educational context, several considerable reforms were carried out in 2012.In the EFL teaching and learning contexts, first, English class became an obligatory class from second grade to 12th grade.Secondly and more importantly, EFL curricula of primary and secondary education were designed in accordance with the Common European Framework of Reference for Languages that highlights communication and interaction instead of grammarbased language learning.However, while those dramatic changes were carried out, interest in research not only on a general educational context but also on EFL learning context remained quite scarce.How those changes affect the EFL learning process needs to be clarified in a research context.In other words, how FLA affects the learning process and whether there is a relationship between FLA and certain variables such as age, gender and grade should be investigated.Below, after introducing several terms regarding foreign language, a research synthesis was presented.
For Gardner and MacIntyre (1993), FLA is an apprehension encountered during a situation that requires the use of a foreign language when the user is not proficient in the target language.In other words, FLA is situation-specific because of specific situations and events (Aydin, 2008;Ellis, 1994).For Horwitz, Horwitz, and Cope (1986), FLA is a complex of perceptions, behaviors, beliefs and feelings that relate to classroom learning and has three varieties.First, communication apprehension is experienced when the speaker suffers from the lack of mature communication skills.Second, fear of negative evaluation is experienced when the language learner feels incapable of having a proper social impression.Lastly, test anxiety is an apprehension that is experienced regarding failing in tests (Horwitz & Young, 1991).
In the broadest perspective, research shows that how FLA affects the learning process is still an unresolved issue and that FLA is affected by certain factors such as perceptions and attitudes towards foreign language learning, learners' beliefs, interpersonal relations, learning and teaching contexts, classroom practices, examinations, learners' perceptions of the target culture, teachers, motivation, teaching materials and proficiency levels (e.g.Oxford, 1992;Price, 1991;Sparks & Ganschow, 1991;Young, 1990;Young, 1991).Research also indicates that FLA prevents oral and written production (Aydin, 2008;Koch & Terrell, 1991;Steinberg & Horwitz, 1986;Young, 1991), while it is concluded that a moderate level of FLA facilitates oral production (Gregersen, 2003).More interestingly, there exist several studies concluding that FLA has no influence on speaking skills (Gardner, Lalonde, Moorcroft, & Evers, 1987;Matsumoto, 1989;Young, 1986).To conclude, as the results obtained from prior research have not reached a consensus on the effects of FLA, more research seems necessary.More importantly, it should be underlined that all of the studies mentioned above focused on adult language learners, while there is no data in relation to FLA among children and the variables of age, gender, and grade.
In a Turkish EFL context, the fairly limited number of studies appeared on FLA.Those studies mainly focused on the relationship between anxiety levels and gender (Aydin & Takkaç, 2007;Öztürk & Gürbüz, 2013), the sources and effects of FLA (Aydin, Yavuz, & Yesilyurt, 2006), FLA effect on achievement (Çakıcı, 2016;Dalkılıç, 2001), and technology effect on FLA (Aydin, 2011).Moreover, several studies explored anxiety types (Aydın, 2008;Koralp, 2005) and writing anxiety (Atay & Kurt, 2006;Kırmızı & Kırmızı, 2015).The main concern with the limited number of studies cited above is that all of the studies dealt with adult learners, while children who learn English as a foreign language were not considered as sample groups.
In terms of children as foreign language learners, the number of studies is also quite limited.As an example, Chan and Wu (2000;2004) explored FLA among children and concluded that they mainly suffered from fear of negative evaluation.They also found that parental expectations and personality were significant variables that caused FLA, whereas they underlined that teachers were not aware of FLA among children.In the Turkish EFL context, Aydin (2012;2013) investigated test anxiety among children and found that they experienced test anxiety at a moderate level.Aydin (2012;2013) also noted that item instructions and learners' attitudes towards their teachers, validity of tests, test length, test techniques and time limitations during tests were the factors that influenced the level of anxiety.However, the above-mentioned studies have several limitations.First, in their studies, Chan andWu (2000, 2004) used the FLCAS developed by Horwitz (1986) for adult learners as a data collection tool.Second, Aydin (2012;2013) focused on test anxiety rather than FLA in a broader perspective.In other words, similar to the studies conducted by Chan andWu (2000, 2004), Aydin also used Test Anxiety Scale developed by Sarason (1978) for adult foreign language learners.To conclude, it can be stated that no moderation and adaptation of the scale for children were considered in the studies cited above.
Concerning the effect of age on FLA, research suggests that age is a considerable predictor that relates to foreign language anxiety among foreign language learners.However, as Dewaele (2007) stated age is also a neglected variable regarding FLA, while it is evident that adult and young learners cannot be treated equally in regards to their responses to language learning and FLA.In his study, Dewaele (2002) found that mature learners found it harder to accommodate to the rules of a foreign language; therefore, their anxiety levels tended to be higher than younger groups.Similarly, Onwuegbuzie, Bailey, and Daley (1999) noted that mature language learners tended to indicate more anxiety levels than younger learners considering FLA.To add, MacIntyre and Gardner (1994) stated that the older the language learner got, the greater the anxiety level was because adult learners took more time while processing information and put more importance on being accurate (Salthouse & Somberg, 1982).In regards to children, the number of studies concerning age effect on FLA is quite limited, since age variable is generally analyzed in the context of adult learners.However, Dewaele (2007) highlighted that young learners may develop more anxiety than adult learners on the ground that language is a new situation for them, and children can be influenced by affective states easily when introduced to new situations.However, in a descriptive study, Aydin (2013) stated that young learners were more relaxed in terms of test anxiety.In other words, when learners were compared in terms of age, older students felt more anxious about English tests.In addition, Chan and Wu (2000) found that FLA was mostly triggered by the pressure put on children's shoulders in a way that they could not succeed in their English classes.
Research on gender, as one of the factors influencing the language learning process and the level of anxiety among foreign language learners, shows conflicting results.For instance, several studies focused on gender effect on the level of FLA (Gandhimathi, 2016;Dewaele, MacIntyre, Boudreau, & Dewaele, 2016) and concluded that gender and the level of FLA were interrelated.In this sense, Gerencheal and Horwitz (2016) conducted a study among university students and found that females had a higher level of anxiety when compared to male learners.Similarly, Ezzi (2012) investigated the relationship between FLA and concluded that female students felt more anxious.On the other hand, Mesri (2012) revealed that there was a considerable difference between female and male students' anxiety levels.However, there exist contradictory findings indicating male learners were more worried (Na, 2007;Fariadian, Azizifar, & Gowhary, 2014), while it was found that there was no correlation between gender and foreign language anxiety (Loo & Kitjaroonchai, 2015;Piechurska-Kuciel, 2012).Critically enough, gender differences regarding FLA among children have drawn little attention.Alshahrani and Alshahrani (2015) investigated the relationship between gender and FLA levels among elementary school students.They concluded that gender differences did not play a significant role on FLA.On the contrary, Abu-Rabia ( 2004) investigated the effect of gender and teachers' behaviors on FLA among seventh graders and revealed that gender was a significant predictor of FLA among children.
Research shows that grade is a considerable variable that has an influence on FLA, while it should be underlined that few studies appeared on the issue.For instance, Liu (2006) conducted a study on Chinese EFL learners and found that as the proficiency level of undergraduate learners got higher, the participants tended to be less anxious.Marcos-Llinás and Garau ( 2009) investigated the effect of the proficiency level in FLA among college students learning Spanish as a foreign language.They noted that FLA differed across proficiency levels.That is, while advanced learners showed a higher level of FLA, begin-ning and intermediate learners experienced a low level of anxiety.Finally, MacIntyre, Noels, and Clément (1997), in their study, asked adult Anglophone students learning French as a foreign language to go through a self-assessment task and used perceived competence as a variable in the measurement of FLA.The results of the study indicated that language competence and anxiety were interrelated in a way that anxious students had a tendency to underestimate their competence while less anxious students tended to overestimate their competence.In terms of children, research does not provide a multitude of studies on the issue.As an example, Abu-Rabia ( 2004) investigated the effect of gender and teachers' behaviors on FLA with specific respect to seventh graders.It was found that gender and teachers' behaviors were significant predictors of FLA among children, while the study did not provide a cross-grade comparison, as previously mentioned.In another study, Aydin (2013) reached results showing the relationship between grade and test anxiety among children who study in primary school in Turkey.He noted that higher grade students seemed more worried about foreign language tests and pop-up quizzes than the lower grade students.
Based on the above discussion, several reasons guided the current study.First and in the broadest perspective, the number of studies on the relationship between certain demographic variables and the levels of FLA both in the global and Turkish EFL contexts regarding adult and young learners are fairly limited.In a narrower sense, it seems necessary to explore whether there is a relationship between FLA and certain variables such as age, gender, and grade.Second, whereas the findings obtained from prior studies did not reach a consensus on the influences of FLA, studies mainly focused on adult language learners rather than young learners.Thus, children need to be paid special attention in terms of FLA and the relationship between the levels and FLA and demographic variables, namely, age, gender, and grade.Third, the number of studies on FLA among children is fairly limited, whereas scales used to measure children's behaviors with regard to FLA have certain limitations, as noted above.To conclude, it is evident that research is strongly necessary to explore whether there exists a relationship between the levels of FLA among children and subject variables, age gender and grade.The study interrogates one research question: Does the level of foreign language anxiety among children who learn English as a foreign language differ in accordance with age, gender, and grade?

METHOD
The participants in the study were 494 Turkish children who learn English as a foreign language at primary and secondary schools in Balikesir city center in Turkey.The sample group consisted of 238 (48.2%) girls and 256 (51.8%) boys.Their mean score for age was 9.82 within the age range between seven and 12.The participants were 91 (18.4%) 2 nd , 98 (19.8%) 3 rd , 87 (17.6%) 4 th , 80 (16.2%) 5 th , 86 (17.4%) 6 th , and 52 (10.5) 7 th grade students.The rationale behind choosing the mentioned sample group was that all of the participants were under their puberty and that they had obligatory EFL classes.
Two research instruments, a background questionnaire, and CFLAS, were used to collect data.First, the background questionnaire was used to obtain demographic information about participants' ages, birth dates, and grades.Second, the CFLAS which consisted of 18 items was aimed to interrogate the students' foreign language anxiety levels.Facial expressions from one to five (1=very unhappy, 2=unhappy, 3=neither happy nor unhappy, 4=happy, 5=very happy) were included in the scale.The CFLAS were adapted from Horwitz' (1986) FLCAS.The adaptation steps were the translation process of the FCLAS (Aydin et al., 2016a), simplification, moderation, a pilot study of validation (Aydin et al., 2016b) and the validation of the scale (Aydin et al., 2016c) that consisted of 20 items.However, it should be added that two items that did not function were removed from the scale, after carrying out a principal component analysis and the Varimax method.
The study followed this procedure.First, a document was obtained from the Ethics Council of Social and Educational Sciences of the Graduate School of Social Sciences of Çanakkale Onsekiz Mart University.Second, a written permission from the authorities of the National Education Ministry was provided.In the third step, students, parents, teachers and school administrators were informed about the purpose and significance of the study and research ethics.Last, after the students were invited to participate, the background questionnaire and CFLAS were administered to the participants during the academic year of 2015 -2016.
Three steps were followed regarding the statistical procedure of the study.First, before computing independent t-test and ANOVA, the reliability coefficient of the CFLAS was found.The reliability coefficient was found to be 0.78 in Cronbach's Alpha indicating a high level of reliability.Regarding subscales, the reliability coefficients were found to be 0.79 for communication ap-prehension, 0.74 for fear of negative evaluation and 0.74 for test anxiety.The total variance of the CFLAS also indicated that the scale obtained validity (% of the variance= 48.58).Finally, independent samples t-test and ANOVA were performed to see the relationship between the items in the CFLAS and demographic variables, age, gender, and grade.

Findings
The findings obtained from the study are divided into three subsections: First, data on the relationship between age and the levels of FLA, communication apprehension, fear of negative evaluation and test anxiety were presented.Then, findings on the relationship between gender and FLA levels were given.Last, data on the relationship between grade and the levels of FLA and its types were presented.

Age
Ten items in the scale were found to be significant regarding the age variable as highlighted in Table 1.To begin with, considering examinations in English classes (p=.00), children at the ages of seven and 12 felt less worried.As for speaking English in the class (p=.00), the participants felt quite worried, while the participants at the age of 12 were less worried.It was also found that when the participants were addressed by their teachers in English classes (p=.00), the level of anxiety was high as well, with the highest score at the age of ten.What is more, younger students felt more anxious when they were given a chance to speak in English classes (p=.00).Additionally, the participants at the ages of eight, nine and 10 suffered more from anxiety (p=.00).Similarly, the participants at seven years felt more anxious when they had to speak in the target language without any preparation in advance (p=.01).The participants at seven years felt more worried about the idea of failing in English exams (p=.02), perceived more anxiety when they did not understand what the teacher was correcting (p=.01) and felt more anxious when they were asked a question for which they did not prepare in advance (p=.00).Finally, findings showed that the youngest age groups suffered more from anxiety when other students laughed while they were speaking English (p=01).
As shown in Table 1, in the scope of FLA types, various findings were found.First, results indicated that age as a variable had significant value in terms of communication apprehension.For instance, younger learners stated that they did not feel anxious when they were speaking English in the class (p=.00), whereas they felt even less anxious when they were given a chance to speak in their English classes (p=.00).Age also differed regarding speaking without any preparation in English classes (p=.02) and being asked a question which they had not prepared in advance (p=00).Older participants also stated that they felt more anxious than the younger ones when they did not understand what the teacher was correcting (p=.01).Second, in terms of fear of negative evaluation, students at older ages experienced a higher level of anxiety compared to the younger ones when other students laughed at them while speaking English (p=.01).Third, regarding test anxiety, findings showed that nine-yearold students did not feel anxious when they had English examination when compared to the older ones (p=.00) while they suffered from failing in tests (p=.02).

Gender
As shown in Table 2, considering the items in the scale, it was found that six items were significantly different in terms of gender.The first item focusing on how the participant felt when they failed English classes (p=.00) revealed that female students felt tenser in the face of failure in English classes.In addition, the second item regarding their feelings when they did not understand what the teacher was correcting indicated a statistically significant difference between genders (p=.00).That is, female students felt less confident of teachers' corrections.In addition, the item on how they felt if other students laughed when they spoke English indicated a significant difference (p=.00).In other words, female students felt more anxious when compared to male students.The item on how students felt when they were well-prepared for the English class was also significantly different considering genders (p=.00).However, female students were more secure about themselves when they were well-prepared for English class than male students.Furthermore, the fifth item that inquired students on how they felt when they volunteered answers in English class indicated a significant difference between genders (p=.01).In other words, findings showed that female students felt more confident when they volunteered in their English classes.Last, the sixth item inquiring about students' feelings when they felt well-prepared for English exams indicated a statistically significant difference considering gender (p=.00).That is, female students felt more selfconfident when they were well-prepared for English exams.
Table 2 also shows that, in the scope of FLA types, various findings were reached.First, in terms of communication apprehension, female students felt more worried due to teacher corrections than males (p=.00), while they felt more confident of volunteering in classes than male learners (p=.01).Regard-ing fear of negative evaluation, female students seemed more sensitive than males when other students laughed at them (p=.00), while they felt more confident in front of their classmates than males when they prepared for their classes (p=.00).In terms of test anxiety, findings showed that females felt more selfconscious about themselves if they failed in English classes than males (p=.00), while male participants felt more confident of getting prepared for examinations (p=.00).

Grade
As indicated in Table 3, seven items in the CFLAS are significantly correlated to grade.First, in terms of having English examinations (p=.00) and being asked a question when the participants were not prepared (p=.00), lower grade students felt more worried.Second, findings showed that lower grade students felt more comfortable than upper graders when they were speaking in their classes (p=.00).On the contrary, regarding learning rules to speak in the target language upper graders felt more worried when compared to lower grade students (p=.00).What is more, higher grade students seemed more unhappy than lower grade students, when other students laughed at them during speaking (p=.00).Last, lower graders felt more anxious when their teachers called (p=.00) and when they were given a chance to speak in their classes (p=.00).
Results also showed that grade is a considerable variable in terms of communication apprehension, fear of negative evaluation and test anxiety among children.First, in the scope of communication apprehension, lower graders felt more nervous when they were called by the teacher (p=.00), when they were forced to speak English in classes (p=.00) and while they were learning rules (p=.00), while they experienced less anxiety in speaking activities (p=.00).Second, lower grade also experienced fear of negative evaluation as they suffered more than higher grade students when other students laughed at them (p=.00).Third, findings indicated that higher grade learners felt more anxious in terms of test anxiety.That is, higher graders experienced more anxiety during their examination (p=.00) and being asked a question when the participants were not prepared (p=.00).

Discussion
In a broader sense, it is obvious that the results obtained from the study may contribute to the related literature in a global EFL context as the number of the studies on age, gender and grade effects on FLA is too limited to draw conclusions.What is more, the study is also significant, as research lacks the above-mentioned issues in the Turkish EFL context.The results of this research reveal that age, gender and grade are considerable factors regarding FLA, communication apprehension, fear of negative evaluation and test anxiety among children who learn a foreign language.In addition, it should be underlined that previous research shows that a moderate level of FLA facilitates oral production among adult students (Gregersen, 2003) and that FLA has no influence on speaking skills (Gardner et al., 1987;Matsumoto, 1989;Young, 1986), while this study reveals that children are adversely affected from anxiety during speaking.In this sense, it is evident that research on anxiety among children requires special attention, as the findings reached in this study do not provide a consensus.When the conclusions of the study are compared to the ones in prior research, it can be stated that the findings support that children suffer from fear of negative evaluation as Chan andWu (2000, 2004) note.However, it should be added that children also suffer from communication apprehension and test anxiety as found by Aydin (2012;2013).In terms of age, contrary to the findings that younger learners are less anxious (Dewaele, 2002;Onwuegbuzie et al., 1999;MacIntyre & Gardner, 1994;Salthouse & Somberg, 1982), this study concludes that they experience anxiety.That is, as noted that children can be influenced by affective states easily when they are introduced to new situations (Dewaele, 2007), the study reveals that younger learners are affected more by FLA.Regarding gender, on the contrary to the findings obtained from previous research that shows no correlation between gender and FLA (Alshahrani & Alshahrani, 2015;Loo & Kitjaroonchai 2015;Piechurska-Kuciel, 2012), this study reveals that the level of FLA among children shows gender differences.In addition, while it is claimed that male learners seem more anxious (Na, 2007;Fariadian et al., 2014), the study suggests that female students are more worried about the fear of failure, teachers' corrections and negative evaluations and unpreparedness.

CONCLUSIONS AND SUGGESTIONS
Four conclusions were obtained from the study.First of all, in the broadest perspective, the level of foreign language anxiety among children significantly differs in terms of age, gender, and grade, when several items in the CFLAS were considered.Regarding age, younger students feel more worried during speaking activities, examinations, and teachers' corrections, when compared to older learners.In terms of gender, female students seem more anxious due to the fear of failure, teachers' corrections and negative evaluations and unpreparedness, while they feel more secure during speaking when compared to male students.In terms of grade, lower grade students suffer more from anxiety due to examinations, unpreparedness, teachers' corrections, while they feel less worried during speaking activities, learning grammar and negative evaluations when compared to the students in higher grades.The second conclusion reached in the study is that the level of communication apprehension considerably differs regarding age, gender, and grade.Speaking more specifically, younger learners feel more secure regarding oral communication and interaction with teachers and classmates, while older learners suffer more from anxiety regarding communication with their teachers when they encounter teachers' corrections.In terms of the relationship between gender and communication apprehension, female students seem more worried during communication with their teachers during corrections, while male students feel less confident when volunteering in speaking.In terms of grade, lower graders feel more nervous during grammar activities and when they are forced to speak and called by teachers, whereas higher grade students suffer more during speaking activities.Third, the level of fear of negative evaluations among children seems considerably different in terms of age, gender and grade.That is, younger learners suffer more from being laughed by other students during conversations when compared to older students.When gender is considered, female students seem more sensitive to being laughed by other students and unprepared speaking in front of their friends.In a similar way, lower grade students feel more anxious than higher grade learners when their classmates laugh at them when they speak.The last conclusion of the study is that the level of test anxiety considerably differs when age, gender, and grade, are considered.For instance, interestingly enough, both the youngest and the oldest ones suffer from failing in tests and examinations in English.Female students are more confident of tests in English, while male learners feel less worried during test preparation.In terms of grade, higher grade students seem more anxious during examination and unexpected questions, when compared to lower grade students.
Some practical recommendations regarding the findings of the study can be listed.First of all, it can be recommended that teachers should raise their awareness that age, gender, and grade are considerable factors relating to the level of FLA among children.In narrower scope, they should pay special attention to younger learners while conducting speaking activities, preparing and administering examinations and making corrections.Second, they should be aware that female students need special attention regarding test anxiety.That is, they should motivate and guide female students for preparation to classes, use positive correction strategies and design tests that have facilitating effects on learning.Teachers should also avoid explicit grammar instruction for lower grade students and focus on communicational and contextual activities for lower grade students.More importantly, lower grade students should never be forced to speak and exposed to unexpected and threatening questions during classes.Third and last, as younger learners and females suffer from negative evaluation, teachers should provide a code of conduct in a democratic atmosphere.They should also use strategies to guide older students who get prepared for central examinations about how to prepare them for the examinations.In conclusion, it can be stated that EFL teachers should raise their awareness of a pedagogical perspective with respect to age, gender and grade effects on FLA among children.
One of the limitations of this study is that the participants were limited to 494 Turkish EFL learners in the age range of seven and 12.The scope of this research was limited to the descriptive data collected from a background questionnaire and CFLAS adapted by Aydin et al. (2016a;20016b;2016c).The data included children's age, gender, and grade and their relationships with the level of FLA regarding communication apprehension, fear of negative evaluation and test anxiety in the Turkish EFL context.Further research concentrating on other factors that may have influences on their levels of FLA such as children's perceptions, attitudes, beliefs, interpersonal relations, target culture, teachers, motivation levels, teaching materials and proficiency levels is warranted.Moreover, to deepen the understanding of FLA among children, in addition to descriptive studies, qualitative and experimental studies are necessary for various educational, cultural and contextual settings.