Anxiety over EFL speaking and writing: A view from language classrooms

The assumption that foreign language learners experience a high level of anxiety mainly when faced with speaking activities implies that research should focus on those learners prone to anxiety over that skill. Despite not being widely investigated, foreign language writing anxiety also seems to be a concern for a large number of students. Drawing on questionnaire findings, the study reported in this article examined the nature of, and the connection between the English language classroom speaking and writing anxiety of 128 Greek EFL learners in private language school settings. Speaking anxiety was operationalised by Horwitz, Horwitz, and Cope s (1986) Foreign Language Classroom Anxiety Scale, and wri ng anxiety was measured by Gungle and Taylor s (1989) ESL version of the Daly and Miller s (1975) Writing Apprehension Test. Interconstruct and intraconstruct associations between the two instruments were examined through principal components analysis with varimax rotation and correlations check. A significant and high correlation was found between classroom anxiety and speaking anxiety, thus indicating that the English language classroom context is a source of speaking anxiety. Writing anxiety was found to load primarily on items relating to attitudes towards writing in English followed by self-derogation for the process and fear of negative evaluation by the teachers and/or by fellow students. On the basis of the findings, suggestions are made concerning the reassessment of the influence that writing anxiety exerts on classroom performance and the adoption of teaching techniques that promote topic-centred process writing.

The last three decades have seen a growing number of studies concerned with the interplay between individual differences and foreign or second language learning and teaching (Dörnyei, 2005;Ellis, 2008;Gardner & Lambert, 1972;Skehan, 1989).Those publications aimed at showing that learners differ upon certain dimensions and, therefore, achieve different levels of performance and success in second/foreign (L2) learning.Individual differences research typically includes anxiety whose impact on language learning has been widely investigated by both practitioners and researchers (Horwitz, 2001(Horwitz, , 2010;;Horwitz, Horwitz, & Cope, 1986;MacIntyre, 1999;MacIntyre & Gardner, 1991;Young, 1990).Horwitz, Tallon, and Luo (2010) argued that "approximately one-third of students studying a foreign language experience at least a moderate level of foreign language anxiety" (p.99).
To date, however, relatively few studies have explored the nature and effect of writing anxiety (WA) on English as a Foreign Language (EFL) learning.By contrast, anxiety over speaking has received the most empirical attention in the literature to date (Gregersen & Horwitz, 2002;Horwitz et al., 1986;Phillips, 1991;Woodrow, 2006).Additionally, the number of studies examining the relationship between classroom anxiety and WA in instructed EFL learning contexts is scarce.The current study was designed in order to address this gap.The paper begins with a brief theoretical overview of foreign language classroom anxiety (FLCA) with specific attention to anxiety over speaking and writing.The remainder of the present paper deals with an empirical investigation of the issue described above.

Speaking and Writing Anxiety from a Classroom-oriented Perspective
FLCA is defined as "a distinct complex of self-perceptions, beliefs, feelings, and behaviours related to classroom language learning arising from the uniqueness of the language learning process" (Horwitz et al., 1986, p. 128).It occurs when students attempt to successfully use a second or foreign language which they have not yet adequately or fully mastered.Hence, FLCA influences language achievement (Gardner, 1985).MacIntyre and Gardner (1991) have posited that "anxious individuals think about their own reaction to a task in addition to the demands of the task itself" (p.297).Negative self-related cognition intrudes on their task performance in class and, consequently, anxiety rises.This is what ultimately differentiates language anxiety from other forms of anxiety, suggesting therefore that second language contexts should be studied in isolation.MacIntyre and Gardner (1991) further argued that "if anxious students could focus on positive experiences in the second language, rather than on negative ones, the debilitating effects of language anxiety could be reduced" (p.297).Thus, foreign language anxiety is a distinct variable in the language learning process and has detrimental effects on learning.
Language anxiety researchers have suggested considering language anxiety from a situation-specific perspective (Ellis, 2008;Horwitz, 2001;Horwitz & Young, 1991;MacIntyre, 1999).In other words, the fact that language anxiety manifests itself only in specific situations, in this case foreign language learning, should be taken into account.Students who have it may not have difficulty in other classes and do not have poor study habits.On the contrary, they are very competent and resilient in most other contexts.The key assumption about FLCA refers to students' immature second or foreign language attempts to communicate.According to Horwitz et al. (1986), "as an individual's communication attempts will be evaluated according to uncertain or even unknown linguistic or socio-cultural standards, second language communication entails risk-taking and is necessarily problematic" (p.128).More recently, Dörnyei (2001) argued that language classrooms are "inherently face threatening environments" where learners are expected to perform through the use of a "severely restricted language code" (p.91).Therefore, difficult or insufficient language input along with concerns about others' evaluations in the EFL setting result in the creation of FLCA.
Most anxiety research, however, has placed disproportionate emphasis on anxiety associated with second or foreign language speaking (Horwitz, 2001;Horwitz et al., 1986;Kitano, 2001;Phillips, 1992;Woodrow, 2006), which has consequently led to the development of questionnaires dominated by speakingrelated items.This speaking-centred approach to FLCA raises certain theoretical and empirical questions regarding the (in)appropriateness of these instruments to identify students' anxiety about performing writing, listening, or reading.This also rationalizes the fact that slightly different variables with other names recognized in applied linguistics research, such as L2 willingness to communicate, are speaking-oriented, thus precluding the possibility of discussing writing as a form of communication as well.Current paradigms focusing on oral communicative competence in language teaching goals, though, could provide a powerful counter-argument to this insistence on L2 speaking.
To return to writing anxiety, and given that "language educators tend to think about developing their students' foreign language competence in four areas" (Horwitz et al., 2010, p. 106), a shift of the research focus onto the neglected domain of writing is necessary.Early research on communication apprehension (McCroskey, 1970) was the stepping stone to studying WA.The latter construct initially emerged in the literature under the term writing apprehension (Daly & Miller, 1975) to denote "the dysfunctional anxiety that many individuals suffer when confronted with writing tasks" (Cheng, 2002, p. 647).By analogy with FLCA, writing apprehension appeared as a distinct type of anxiety arising from the uniqueness of the written communication process.Madigan, Linton, and Johnson (1996) argued that "distress associated with writing and a profound distaste for the process" constitute the two main effects of anxiety about writing on prospective writers (p.295).
With reference to WA about English, Leki (1999) claimed that writing, albeit the most private and self-controlling of the four skills, causes EFL learners to experience a kind of "writer's block" (p.65).Additionally, through factor analytic procedures, Cheng, Horwitz, and Schallert (1999) found a significant moderate correlation (r = .65)between second language classroom anxiety and second language WA proving that these two anxiety constructs are related but are not identical.Cheng et al. (1999) concluded that "some anxious second language student writers may suffer chiefly from low writing-related selfesteem, some from negative affect toward the writing activity and some from fear of evaluation" (p.436).The WA experience among L2 learners should, therefore, be studied by seeking sociolinguistic dimensions when tackling EFL writing tasks, such as students' relying heavily on others' criticism of their work.It would be fair enough then to assume that, apart from speaking, EFL writing also exhibits a considerable degree of learner self-exposure.
Finally, Cheng et al. (1999) attributed a language-skill-specific character to WA because they found that it highly correlated with writing achievement.As a consequence, language anxiety researchers have proposed differentiating language-skill-specific anxiety from general classroom anxiety that seems to be more associated with speaking (Cheng, 2004;Cheng et al., 1999;Horwitz, 2001).In particular, Cheng et al. (1999) reported that: some language learners may feel particularly anxious about speaking in the second language, and some about writing. . . . the discrepancy between a learner's first and second language competence in different skill areas, a language learner's varied experiences in acquiring each of the four language skills, and his or her history of success and failure in performing each skill might lead to differentiated attitudes, emotions, and expectations about each of the language skills.Language-skillspecific anxiety might well be one of the negative emotions and attitudes formed during the process of second language learning.(pp.Having provided an overview of writing and speaking anxiety, I will next deal with empirical evidence on the aforementioned topic.

The Present Study: Purpose and Research Questions
Conducted as a partial replication of Cheng et al.'s (1999) study, the research goal of the current attempt was to examine the constructs of language anxiety and WA, as well as their conceptual links, situating both in the same English language learning context, that is, EFL learning in Greece.An individual characteristic of the Greek foreign language education system is "a thriving private sector of foreign language institutes providing intensive foreign language tuition . . .Courses offered at private language institutes are not compulsory, are mostly exam-oriented and give the opportunity to students to sit for exams which will allow them, if successful, to obtain a language certificate" (Mattheoudakis & Alexiou, 2009, pp. 230-1).Students' anxiety levels, therefore, appear to intensify as a response to success-orientedness and long-term exam-orientedness imposed on them by the system.The researcher's general objective led to the following research questions (RQs): 1. Is Greek EFL learners' speaking anxiety part of classroom anxiety? 2. Which factors influence Greek EFL learners' WA? 3. To what extent is Greek EFL learners' WA related to classroom anxiety?

Participants
The participants of the study were 128 EFL students enrolled in general English classes in two private language schools in Greece.All informants were adults; the average age was 24.12.There were 44 male and 84 female students.All of them gave their consent to participating.The minimum length of exposure to English was two years with a maximum length of eight years.A few false beginners (N = 11) also took part in the project.The participating students' proficiency level ranged from B1 to C2 (in CEFR1 standards, from lower intermediate to upper advanced).A description of the band levels can be found in Appendix A.

Instruments
A background questionnaire to capture demographic data and two Likert-type questionnaires to measure anxiety specific to the English language classroom, as well as to WA were administered to the students.Classroom anxiety was operationalised by a modified version of the Foreign Language Classroom Anxiety Scale (FLCAS; Horwitz et al., 1986) and WA was measured by an adapted version of the English as a Second Language Writing Apprehension Test (ESLWAT; Gungle & Taylor, 1989).
The adapted FLCAS contained 29 items.Test anxiety items included in the original questionnaire were eliminated due to lack of consistency with the research purpose; the focal point of the study was to measure speaking and writing anxiety resulting from EFL lessons as such and not from testing situations.Moreover, items related to interaction with native speakers were also excluded, because they refer to hypothetical situations that do not occur in the participating English language classrooms.In this study, the term language used in the original FLCAS was replaced by English language.For instance, the original FLCAS item "I feel confident when I speak in my foreign language class" was modified to "I feel confident when I speak in my English language class." The ESLWAT adapted questionnaire that was administered to the students consisted of 17 statements.The original instrument included 26 statements answered on a 6-point Likert scale.The researcher decided to eliminate those items that were not congruent with the English language classroom context, such as "I would enjoy sending my English writing to magazines to be evaluated and published."Additionally, only one out of two or more tautological items that highly correlated during the piloting stage of the project, such as "I don't like my English compositions to be evaluated" and "I am afraid of writing essays in English when I know they will be evaluated," formed part of the ESLWAT final version.The double-barrelled item "I have no fear of my English writing's being evaluated by my teacher and/or by my peers" was further split into two more specific items: "I have no fear of my English writing's being evaluated by my teacher" and "I have no fear of my English writing's being evaluated by my peers."

Scoring Method
The two questionnaires were answered on a 5-point Likert scale2 , ranging from 1 (strongly disagree) to 5 (strongly agree).The total classroom anxiety score and the total WA score were derived by summing the students' responses to respective items.Positively worded items, such as "It wouldn't bother me at all to take more English language classes" and "I look forward to writing down my ideas in English" were reversed and recoded.The descriptive statistics for both instruments are summarized in Table 1.

Procedure
Multivariate statistics, and in particular factor analysis, were applied to the classroom and WA questionnaires.Principal components analysis with varimax rotation was selected for the intraconstruct examination of the FLCAS and the ESLWAT in order to identify those components that best define each of the two anxiety measures.Further, correlation coefficients check was computed to examine the interconstruct relationships among the entire pool of items from the FLCAS and the ESLWAT.The data were entered into and analysed through SPPS version 16.0 for statistical analysis.

Findings RQ1: Is Greek EFL learners' speaking anxiety part of classroom anxiety?
To address this research question, the FLCAS was subjected to exploratory principal components analysis with varimax rotation.Table 2 displays the item loadings on FLCAS1.The remaining two factors with their item loadings can be found in Appendix B.

Items
Factor loadings I can feel my heart pounding when I'm going to be called on in my English language class..668I always feel that the other students speak English better than I do.
.664 I am afraid that the other students will laugh at me when I speak in English.
.651 I feel more tense and nervous in my English language class than in my other classes.
.644 I keep thinking that the other students are better at English than I am.
.639 I tremble when I know that I'm going to be called on in my English language class.
.602 I never feel quite sure of myself when I am speaking in my English language class.
.599 I feel confident when I speak in my English language class.
.529 It embarrasses me to volunteer answers in my English language class.
.524 I worry about the consequences of failing my English language class.
.482 I often feel like not going to my English language class.
.407 I don't worry about making mistakes in my English language class. .340 The number of components to be extracted was guided by the scree plot and a three-component solution, accounting for 42.51% of the total vari-ance, was selected.The first component (FLCAS1) consisted of twelve items accounting for 18.5% of the total variance.Most of these items seem to share a feeling of speaking anxiety caused by low perceived self-efficacy and fear of negative evaluation by the peers, thereby signifying that anxiety about speaking in English forms an integral part of FLCA.This factor was labeled Speaking anxiety and fear of negative evaluation.

RQ2: Which factors influence Greek EFL learners' WA?
This research question was addressed by considering the ESLWAT intrastructure factor loadings examined through principal components analysis with varimax rotation.

Items per factor
Factor loadings Factor One.Attitudes towards writing in English I like to write down my ideas in English.
.716 Writing in English is a lot of fun.
.714 I enjoy writing in English.
.712 I look forward to writing down my ideas in English.
.711 I like seeing my thoughts on paper in English.
.607 Handing in a composition written in English makes me feel good. .582 Factor Two.Self-derogation when writing in English When I hand in an English composition, I know I'm going to do poorly..785My mind seems to go blank when I start to work on a composition in English.
.761 I have a terrible time organizing my ideas in an English composition course.
.647 Taking an English composition course is a very frightening experience.
.574 It is easy for me to write good compositions in English. .573 Factor Three.Fear of negative evaluation I have no fear of my English writing's being evaluated by my teacher.
.824 I have no fear of my English writing's being evaluated by my peers. .653 As evidenced in Table 3, a three-component solution, which accounted for 51.97% of the total variance, was also selected for the ESLWAT.The first factor (ESLWAT1) included six items addressing learners' likes and dislikes with respect to writing in English.This factor was given the label Attitudes towards writing in English to encompass both positive and negative student attitudes to writing.Five ESLWAT items were selected to define the second component (ESLWAT2).Here, the items refer to negative self-perceptions and concern about potential failure in writing classes.Therefore, this factor presents a selfderogation dimension of writing in English.The last ESLWAT component (ESLWAT3), consisting of two items, is characterized by a strong evaluation apprehension element either by the teacher or by the peers.This component was named fear of negative evaluation.

RQ3: To what extent is Greek EFL learners' WA related to classroom anxiety?
To address this research question, Pearson correlations were computed between the overall FLCAS and ESLWAT, as well as their subcomponents (see Table 4).A significant and moderate correlation (r = .54)was found between the FLCAS and the ESLWAT.In addition, the strongest correlation was between the overall FLCAS and ESLWAT2 (self-derogation when writing in English; r = .55)and between the overall ESLWAT and FLCAS1 (speaking anxiety and fear of negative evaluation; r = .47).The overall FLCAS was highly correlated (r = .91)with FLCAS1 (speaking anxiety and fear of negative evaluation), less highly correlated (r = .71)with FLCAS2 (teacher-related anxiety), and moderately correlated (r = .45)with FLCAS3 (comfortableness with taking English classes).As far as the WA correlations are concerned, there were significant and high correlations between the overall ESLWAT, ESLWAT1 (attitudes towards writing in English) and ESLWAT2 (self-derogation when writing in English; r = .82and r = .81,respectively).Lastly, a significant and moderate correlation (r = .60)was found between the overall ESLWAT and ESLWAT3 (fear of negative evaluation).Table 4 presents the correlation matrix.

Discussion
Corroborating previous research (Cheng et al., 1999), this study showed that FLCA and WA about English are two related but distinguishable variables.The factor analysis and the correlation coefficients check proved that FLCA consists of a strong speaking anxiety element, as opposed to WA, which depends on learners' attitudes towards the writing class.
With reference to the first research question, the classroom anxiety instrument loaded primarily on items related to speaking anxiety, thereby suggesting that speaking anxiety emanates from the broad English language classroom context.The strong association of the FLCAS with speaking anxiety is consistent with Cheng et al.'s (1999) findings, as well as with Aida's (1994) conclusion that "the FLCAS appears to measure anxiety primarily related to speaking situations" (p.163).Statements, such as "I am afraid that the other students will laugh at me when I speak in English, "or" I never feel quite sure of myself when I am speaking in my English language class," could pave the ground for a discussion of linguistic, as well as sociolinguistic factors associated with L2 speaking.On the one hand, EFL speaking anxiety is caused by learners' efforts to produce as accurate an utterance as possible in the classroom paying attention to both pronunciation and language choice and use.On the other hand, self-consciousness and anxiety about EFL speaking could be defined as a state of ongoing social comparisons among learners and as a fear of negative evaluation and loss of face in the event of a mistake.The sociolinguistic dimension of speaking anxiety could thus reduce learners' levels of intended effort when speaking in class and ultimately result in poor achievement.If we then hypothesise that speaking anxiety stems from fear of peer criticism, research is warranted to investigate the teacher's role not only as a language educator, but also as a moderator of certain classroom events that could lead to personal feelings of inadequacy as a learner.MacIntyre, Clément, Dörnyei, and Noels (1998) have indeed underlined the teachers' responsibilities in connection with learners' willingness to participate in communication events in class by saying that: the ultimate goal of the learning process should be to engender in language students the willingness to seek out communication opportunities and the willingness actually to communicate in them.A programme that fails to produce students who are willing to use the language is simply a failed programme.(p.547) WA in turn was shown to relate to students' attitudes to writing classes, self-derogation when writing in English, and fear of negative evaluation.The emergence of attitudes as the first component of the ESLWAT indicates that teachers should implement writing teaching strategies that will grow enjoyment of EFL writing.First of all, it is important to ease students' fears about producing written work in the L2 that is below what they could have produced in their native language.Adult learners are often hesitant to produce language that is below part their mother tongue (L1) skills.Process writing could help students incorporate skills at a natural pace and encourage them to improve their understanding of writing as a task and of the materials covered.Rather than focusing students on working on perfection in the first draft, various writing sessions could be set and writing tasks could be broken into smaller manageable units.Further, Rankin-Brown (2006) suggested that teachers "assign papers that address topics students are already familiar with" (p.5).Through theme-centred modules, students develop critical thinking skills and writing strategies to implement with topics they already know and are willing to discuss.
Viewing WA as a whole, the three components that the factor analysis of the ESLWAT yielded seem to partially support Cheng et al.'s (1999) claims about language-skill-specificity with regard to second language WA.Despite the fact that no correlations have been calculated between WA and writing achievement, it could be presumed that the learners' degree to which they have developed cognitive and linguistic abilities with regard to L1 writing, as well as the knowledge or potential lack of knowledge of strategies to tackle L2 writing could account for their WA, thus suggesting that WA be treated as a language-skill-specific type of anxiety.
Additionally, the findings call for a reevaluation of the role that WA plays in learners' writing performance and consider both speaking and writing as involving an equal amount of self-exposure when practiced in class.Developing practical writing skills, such as techniques for generating and expressing ideas, and having acquired the input needed to proceed with writing tasks would definitely make students feel psychologically secure in the EFL classroom.Nevertheless, WA is also dependent on a number of non-linguistic factors, the most prominent being students' low self-confidence when writing in English due to one's own evaluations and trepidation about being less competent than their peers and not having the right answer to a question set by the teacher.Given that students may see the foreign language classroom as a place where any correction equals failure, teachers should make the classroom as non-threatening as possible.Measures such as selective error correction, by taking into account the main foci of the writing activity and by accompanying it with comments that do not immediately accentuate the errors, could be taken by teachers to help students overcome their WA.However, teaching methodologies themselves could not contribute directly to allaying FLCA and WA.The important role of other individual student traits in lessening language anxiety should be further addressed.Future studies targeting interactions between anxiety, self-esteem, language learning strategies, and personality could supplement the above findings.

Conclusion
Taking into account that L2 writing anxiety is an under-researched topic in second language acquisition literature, this study aimed to shed light on this ostensibly innocuous construct.Horwitz et al. (1986) have made a strong case for the role played by speaking in the creation of language anxiety.To date, however, writing anxiety has been less frequently addressed among language anxiety researchers.The current study constitutes a first attempt to examine issues pertaining to writing and classroom anxiety in the Greek English language learning context.The suggestions for future studies made above can help us to increase even more our understanding of the dynamic and multifaceted construct of EFL anxiety.

Table 1
Descriptive statistics for the FLCAS and the ESLWAT

Table 4
Correlations among overall FLCAS, overall ESLWAT, and their subcomponents