AN EXPLORATION INTO FACTORS ASSOCIATED WITH READING ANXIETY AMONG TAIWANESE EFL LEARNERS

The present study was an elaboration on factors influencing foreign language (FL) reading anxiety in EFL learners (n = 202) in Taiwan. Data were collected from three measures, a background information questionnaire, the Foreign Language Reading Anxiety Scale (FLRAS) and a semistructured interview. The results showed that FL reading anxiety was associated with time spent on reading English, but not with gender and academic major. Moreover, FL reading anxiety was found to be positively correlated with text features and personal factors, but negatively correlated with reading strategy use, indicating that the more reading strategies a learner employs, the lower FL reading anxiety level he/she will experience. Unfamiliar vocabulary, unknown topics, long and complicated text structure as well as the fear of making mistakes were identified as the major factors of FL reading anxiety. This study contributes to extending our understanding of factors that might evoke EFL reading anxiety.

For students learning English as a Foreign Language (EFL), skill in reading is evidently one of the most important skills to be acquired.This is true especially for university students (Mohammadpur & Ghafournia, 2015), since most aca-demic texts are written in English and most foreign language (FL) learners are required to read a great many English texts.However, for many EFL learners, reading English texts is a challenging task (Bektas-Çetinkaya, 2011).One of the challenges is reading anxiety that some FL learners might suffer from.It has been recognized that FL reading anxiety affects FL reading performance negatively (Sellers, 2000;Shi & Liu, 2006;Zhao, Guo, & Dynia, 2013).According to Horwitz, Horwitz, and Cope (1986), FL anxiety was 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" (p.128).In Zhou (2017), FL reading anxiety is distinguished from FL anxiety, which is the specific anxiety that "learners experience while they are reading a passage in the target language" (p.156).FL reading anxiety can be described as a feeling of one's own weakness and inability to cope with reading in a foreign language.In other words, it is perceived as a type of anxiety specifically associated with FL reading.
Some researchers have pointed out the influence of FL anxiety on the success of learning a new language (Kuru-Gonen, 2005;Mohammadpur & Ghafournia, 2015;Rafik-Galea, 2010;Rajab, Zakaria, Rahman, Hosni, & Hassani, 2012) and FL reading performance (Bernhardt, 2005(Bernhardt, , 2011)).It is widely accepted that anxiety may result in different learning processes and outcomes (Alsowat, 2016;Bektas-Cetinkaya, 2011;Brantmeier, 2000Brantmeier, , 2005;;Kilinc & Yenen, 2016;Lin & Ho, 2009).In some cases, anxiety acts as a giant gate, preventing EFL learners from adopting effective learning practices (Kuru-Gonen, 2005).It is highly likely that poor readers experience higher levels of reading anxiety than good readers, since they might be more prone to distractions and other negative effects associated with anxiety.Hence, identifying the causes of reading anxiety might provide insight for instructors, enabling them to decrease EFL learners' reading anxiety and assist the learners in their struggle with reading.
Previous studies have indicated that the influencing factors on foreign language (FL) reading anxiety include gender, experience with the target language country and course level (Joo & Damron, 2015;Zhao, et al., 2013), worries related to comprehension, unfamiliar topics, unknown pronunciation, and feeling uncomfortable in reading aloud (Zhou, 2017), personal perception of reading proficiency (Lien, 2016), text feature and reading course (Al-Shboul, Ahmad, & Nordin, 2013;Kuru-Gonen, 2009;Nazarinasab, Nemati, & Mortahan, 2014), and grade and gender (Elaldı, 2016).Variables examined in this study included age, gender, reading text, personal factors, and reading strategies used, which are variables frequently observed among EFL learners in Taiwan.
The relationship between age and reading anxiety was examined by Jafarigohar and Behrooznia (2012), because it was assumed that the performance would be slightly different in reading among different age groups of learners.In Jafarigohar and Behrooznia's study, 112 participants were involved, whose age ranged from twenty to forty with the majority being twenty to twenty-eight years old.However, the result differed from that expected, since no relationship was found between age and reading anxiety.
Concerning gender and reading anxiety, several studies (Capan & Karaca, 2013;Jafarigohar & Behrooznia, 2012;Lien, 2011;Lu & Liu, 2015) revealed mixed results.Jafarigohar & Behrooznia's study (2012) showed that female participants felt more anxious and worried than male participants did.Zheng (2008), however, claimed that male students suffered from higher levels of anxiety than female students.Lien's (2011) and Capan and Karaca's (2013) study showed no significant difference in reading anxiety among female and male subjects.Yet, female participants seemed to be more anxious overall than male participants (Lien, 2011).These controversial findings may be attributed to Williams ' (1996) conclusion that "male students are more likely to resist being identified with anxiety, since it is viewed as a feeling related more to females than males" (as cited in Capan & Karaca, 2013, p. 1362).
Factors related to reading text have been extensively examined, including unknown vocabulary, an unfamiliar culture and unfamiliar topics and text features.In reading, vocabulary is the fundamental element because all information is delivered through words.Learners encountering unknown vocabulary will more likely struggle with reading.In the studies conducted by Huang (2012) and Kuru-Gonen (2005), the majority of the students agreed that unknown vocabulary could lead to reading anxiety.The results indicated that the more unknown vocabulary a student encountered, the higher reading anxiety a student might feel.Furthermore, the findings of interviews in Al-Shboul et al.'s (2013) study showed that unknown vocabulary was indeed one of the major causes of reading anxiety, since the majority of interviewees responded that not knowing the vocabulary was a real problem in understanding the texts.
Moreover, it is generally accepted that there is a close relationship between cultural background knowledge and FL reading comprehension, as cross-cultural competence is an essential element for successful second language learning.Apart from unfamiliar scripts and writing systems, unfamiliar cultural material was identified by Saito, Horwitz, and Garza (1999) as a potential source for eliciting FL reading anxiety.Huang (2012) postulated that cultural shock is a common factor which can lead to a higher anxiety level in EFL learners.Cultural shock may make them feel more anxious when they are reading texts because of the fear of losing their own identity and their own language in cross-cultural scenarios.Consequently, unfamiliar cultural material seems to have the potential to elicit reading anxiety to some degree for FL learners.
Further factors related to the text are unfamiliar topics and text features.Some EFL learners feel anxious when they come across a topic with which they are not familiar.In the study of Al-Shboul, et al., (2013), the interviewees responded that they would like to read short and easy stories.The EFL learners participating in Kuru-Gonen's (2005) study claimed that the text features, such as length, can also cause their reading anxiety.The majority of the participants did not like to read long texts, which would increase their reading anxiety.
As far as personal factors are concerned, fear of making errors, negative background experience and lack of confidence are potential causes of FL reading anxiety.In several studies, fear of making mistakes has been identified as one of the major sources of EFL reading anxiety for Turkish learners (Al-Shboul et al., 2013;Cubukcu, 2007), Chinese learners (Liu, 2007), and Iranian learners (Celebi, 2009).It was confirmed that being afraid of making mistakes would make the learners feel anxious, because they wished to avoid being laughed at.
Negative background experience refers to the experience or bad situation that EFL learners face during the process of learning a new language.A bad situation could, for example, be in the form of being ridiculed by others or being scolded by teachers.Kuru-Gonen (2005) claimed that due to bad experiences or situations, EFL learners may create a sense of negative feeling which may prevent them from enjoying reading.Bad experiences may also include those EFL learners who do not have a good habit of reading or those who experience negative incidents related to reading, which results in their having some prejudices against reading in the target language.The more negative background experience an FL learner has, the higher reading anxiety level he might experience.Another frequently mentioned affective factor in FL reading research is self-esteem.Research has demonstrated a close relationship between self-esteem and FL reading performance (Hisken, 2011;Juyandegan, 2016), indicating learners with higher self-esteem perform better than learners with lower self-esteem in reading.Similar results were obtained by Zhao, Liang, and Callan (2008) who found a significant negative correlation between FL reading anxiety and reading performance.It is postulated that FL learners who have more confidence are better in having an ability to transform negative emotions (e.g., confusion and anxiety) into positive emotions (Huang, 2012).However, whether or not low self-esteem would trigger FL reading anxiety still remains uncertain.
Reading strategy use is another factor worth investigating.In general, successful use of reading strategies can help EFL learners improve their reading comprehension (Al-Nujaidi, 2003;Brantmeier, 2000;Huang, 2012;Lee, 2007;Wu, 2005), since appropriate use of reading strategies may ease reading difficulties and thereby reduce the anxiety evoked by a reading task.Interestingly, in a study conducted by Ghonsooly (2012), the results showed no significant relationship between EFL learners' FLRAS (Foreign Language Reading Anxiety Scale) scores and SORS (Survey of Reading Strategies) scores.However, Lien (2011) found in her study a negative correlation between FL reading strategy use and reading anxiety.Kuru-Gonen (2005) argued that if an EFL learner relies too much on a dictionary or translation, it is most likely that there is no benefit or enjoyment from reading.Also, they may spend too much time translating and using a dictionary, which results in forgetting what they have read previously.
Based on the above literature review, some issues concerning the sources of FL reading anxiety remain controversial.In view of this, the purpose of the present study is to examine the factors influencing reading anxiety among Taiwanese EFL learners.Specifically, this study investigates the relationship between background variables, reading text, personal factors and strategies used with the students' reading anxiety level.Two research questions for the purpose of the present study are as follows: 1. To what extent is FL reading anxiety related to background variables, specifically gender, age, academic major and time spent weekly in reading English materials among Taiwanese EFL learners?2. To what extent is FL reading anxiety related to reading text, personal factors, and strategy use among Taiwanese EFL learners?

METHOD
The participants of the present study were 202 EFL students at a university in southern Taiwan, comprising 117 females and 85 males.There were 100 non-English majors and 102 English majors, whose age ranged from 18 to 23.All subjects had studied English for at least six years up to secondary school level and had taken additional English coursework upon entering the university.Based on their score in English at the college entrance examination, their English proficiency level was at the intermediate level.All participants signed a consent form before the study was conducted.
This study was conducted through paper-based survey and interview.Data were collected using a questionnaire on background information and factors affecting FL reading anxiety, the Foreign Language Reading Anxiety Scale (FLRAS), and a semi-structured interview.The questionnaire was constructed to investigate possible sources of EFL students' reading anxiety.It was divided into four parts: (a) background information; (b) factors related to reading text; (c) personal factors and (d) reading strategy use.The questionnaire was designed to obtain some basic information about the participants, such as age, gender, the year of study, whether the participants like or dislike reading English texts, and time spent on reading per week.The reliability of the scale was confirmed by the internal consistency coefficient (Cronbach's alpha = .871,n= 202).
In order to investigate the participants' English reading anxiety, Liu's (2011) version of Foreign Language Anxiety Scale (FLRAS) was adapted.The original FLRAS was developed by Saito, et al. (1999), which was modified by Liu in her PhD dissertation.The modified FLRAS consisted of 20 items, which used a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).The internal consistency coefficient of modified FLRAS for the present study was .847(Cronbach's alpha, n= 202), which indicates the reliability of the scale.
A semi-structured interview was designed to gather information required for deeper understanding of the factors affecting FL reading anxiety.On the background information questionnaire, participants could indicate whether they were willing to be interviewed.Among those who indicated their willingness to be interviewed, sixteen participants were invited for a face-to-face interview.They consist of eight English majors and eight non-English majors.The interview began with general questions about how participants felt about reading English texts and then continued to more specific questions about the difficulties they have encountered in English reading, possible sources of their reading anxiety and strategies they used to cope with reading English texts.
Concerning the procedure of data collection, the data from the English majors were collected in reading courses at different levels of the department, while the data of the non-English majors were collected in various classes of practical English course, which focused on the training of English skills including reading.With the support of the instructors, the paper-based survey was conducted during regular class time in the eleventh week of the winter semester in 2016.The participants were asked to complete the questionnaire in the class, which took about 20 minutes.The questionnaire was given to a total of 230 students, but the number of cases finally remitted for purposes of analysis was 202.The data from the students who did not complete the instruments were omitted.After the questionnaires were administered, the researchers arranged time during the semester with the interviewees individually.Each interview took about fifteen to thirty minutes, which was audio-recorded and transcribed.The data collected from the transcripts were analyzed in a qualitative way to find regularities or common features across student responses to identify possible sources of reading anxiety and reading strategies used by the participants when they encounter reading difficulty.The transcripts were first grouped into two categories, reading problems/difficulty and reading strategies, then key words and phrases related to each theme were found from the transcript.

Background Information and Factors Related to FL Reading Anxiety
This study employed descriptive statistical analysis in analyzing the data.Participants' background information to answer the first research question is presented in the following.There were 202 participants in this study, including 85 males and 117 females.As for the major, 102 participants were English majors, while 100 were non-English majors.The factors relevant to reading texts included unknown vocabulary, unfamiliar cultural background, unfamiliar topic, and text format.Table 3 shows the means and standard deviations of the items related to reading text.As displayed in Table 3, the majority of the learners responded that if they had little cultural background knowledge about the topic (M = 3.13, SD = 1.01) or they were not familiar with the topic (M = 3.10, SD = 1.10), they might feel anxious during reading.Not surprisingly, unknown vocabulary was also the main factor which caused anxiety among participants during reading (M = 3.04, SD = 1.08).Personal factors included fear of making errors, negative background experiences, and lack of self-confidence.Table 4 displays the means and standard deviations of the items related to personal factors.As shown in Table 4, the mean of this section was 2.88 (SD = 1.10).The participants seemed to be worried about the possibilities of misunderstanding English reading texts (M = 3.20, SD = 1.10).They did not have much confidence in reading difficult or long texts; however, the majority did not have negative experience in reading English texts (M = 2.59, SD = 1.07)Reading strategy use included using dictionary, translating, note-taking, highlighting, and using diagrams and contextual clues.The means and standard deviations of the items related to reading strategy use are presented in Table 5.Using dictionary to find out vocabulary meaning is the strategy most frequently used by the participants (M = 3.67, SD = 1.04), followed by highlighting, notetaking, translating and using diagrams.

Item
Mean SD When reading in English, I would highlight the parts I don't understand.
3.50 1.06 When reading in English, I usually translate English into Chinese first.
3.32 1.02 When reading in English, I use note-taking to enhance my understanding.
3.42 1.00 When reading in English, I use tables, figures, and charts in the text to increase my comprehension.

1.01
When reading in English, I use dictionary to understand the vocabulary.
3.67 1.04 Average 3.40 1.03 Liu's (2011) Foreign Language Reading Anxiety Scale (FLRAS) was utilized to measure students' English reading anxiety and to answer two research questions in the present study.According to Zhou (2017), a scale of 3.0 indicates a "neutral" response on the FL reading anxiety scale; therefore, a score above 3 reflects that participants tended to agree with an item.The mean score of average anxiety level is 3.13 (SD = .617),which suggests that the students tend to have a high level of English reading anxiety.The mean scores and standard deviations were calculated, and the items of the FLRAS were arranged in descending order.The results are illustrated in Table 6.I am worried about all the new symbols I have to learn in order to read English.

The Correlation between FLRAS and Possible Causes of EFL Reading Anxiety
To address the first research question, one-way ANOVA tests were performed between FLRAS scores and the variances of gender, age, major and time spent weekly reading English materials.Analysis of variance revealed that there is no significant difference in reading anxiety regarding gender [F (1, 200) = .418,p = .518,n.s.], age [F (5, 196) = 2.0, p = .80,n.s.] and major [F (1, 200) = .789,p = .376,n.s.].However, significant difference of anxiety levels was found in the variance of time [F(3, 198)=5.276,p= .002< .05].A correlation analysis was also conducted to assess the relationship between time spent on reading English and FL reading anxiety.Time spent on reading English was found to be negatively associated with FL reading anxiety (r = -.317,p < .05),indicating that the more time an individual spends on reading English, the lower his or her FL reading anxiety.
A Pearson correlation coefficient was computed to examine the relationship between FLRAS and reading text, personal factors, and reading strategy use.The results were presented in Table 7.As Table 7 indicates, FLRAS was positively correlated with reading text [r (202) = .802,p < .001],personal factors [r (202) = .752,p < .001]and negatively correlated with reading strategy use [r (202) = -.410,p < .001].The results showed that there is a positive relationship between FLRAS and these factors.
In addition, multiple regression analysis was utilized to test if text, personal factors and strategy use significantly predicted participants' ratings of FL reading anxiety.The results of the regression indicated the three predictors explained 72.3% of the variance (R2 =.723, F(3,198)=172.434,p<.001).It was found that text significantly predicted reading anxiety (β= .802,p<.001), as did personal factors (β= .752,p<.001) and strategy use (β = .410,p<0.001).

Results of the Interview
The interview questions mainly focused on possible sources of reading anxiety and reading strategies used by the participants when they encounter reading difficulty.Based on the face-to-face interview data, the major sources of reading anxiety were unknown vocabulary and complex structure, both of which were concerned with text feature.The participants responded that if the texts contain too much unknown vocabulary and complex sentence structures, they could not understand the texts and might feel frustrated.Sometimes they might stop reading the text.As one of the students responded in the interview: "I had difficulty during reading.I can't get the meaning about the text because my vocabulary size isn't large enough in order to understand the texts" (Subject 1).Subject 2 also explained his reading difficulties by saying, "Most of the time, vocabulary and complex sentence structure are the main problems, which stresses me out".Similar experience was expressed by Subject 3: "The difficulty I faced is unknown vocabulary.If the texts contain a lot of unknown vocabulary, I won't understand the texts.As a result, I may end up giving up on reading".
Concerning the strategies used by the participants during reading, the results showed that the most frequently used strategies were cognitive strategies, such as using dictionary, translation, highlighting and note-taking.However, they seldom used tables, figures, charts or context clues to help them solve the reading problems.Several participants used a dictionary or Google translation to help them understand the meaning of sentences and vocabulary (Subjects 1 & 3).As for highlighting and note-taking, participants stated that it would be easy for them to understand the main ideas of the text (Subject 11).Also, when using highlighting and note-taking, they could increase their reading comprehension (Subject 7).

Discussion
This study was undertaken to address issues concerning possible causes of EFL reading anxiety among Taiwanese university students.The major findings are discussed in the following.

The Relationship between Background Variables and FL Reading Anxiety
The results showed that in general, EFL learners in Taiwan experienced an above average level of reading anxiety (M = 3.13), which is slightly higher than the mean reported in Zhou's (2017) study (M = 2.80) among learners of Chinese in the United States.Among the four background variables, the overall results of the analyses showed there was no significant difference in reading anxiety pertaining to age, gender and major.Consistent with Jafarigohar & Behrooznia (2012), our results suggest that age is not a relevant predictor of FL reading anxiety.Concerning gender, the finding of this study supports the research conducted by Capan and Karaca (2013), who found that gender did not play an important role in reading anxiety.However, the results somewhat contradict the results of other studies.Jafarigohar & Behrooznia's study (2012) suggested that female participants felt more anxious and worried than male participants did, while Zheng's study (2008) revealed that male students suffered from higher level of anxiety than the female students.As for the variable of major, which has hardly been mentioned in the literature, there is no significant difference in FL reading anxiety between English majors and non-English majors, indicating that major is not an obvious predictor of FL reading anxiety.
It was found that EFL reading anxiety was dependent on the time spent on reading.The results of the present study reveal that learners who had spent more time on reading English experienced lower levels of reading anxiety compared to those who had spent less time on reading.It seems that reading anxiety is related to how much time a learner has spent on reading English texts.Proficient reading in English requires devotion of time.With more experience with English and an increasing proficiency level, learners may find reading in English less intimidating, which would explain why learners who spent more time on reading English had significantly lower levels of reading anxiety than those who spent less time on reading English.Along the same line, Zhou's (2017) study suggests that reading anxiety is related to how long a learner has been learning the target language.

Main Sources of FL Reading Anxiety
Another major finding of the study concerned the sources of reading anxiety.Text feature, personal factors and reading strategy use were identified as major sources of EFL reading anxiety.
As for the relationship between reading text and reading anxiety levels, there is a significant correlation between text feature and FLRAS.Most of the participants agreed that they feel anxious when they encounter unknown vocabulary, unknown topics, and long English texts, which suggests that factors related to text feature would evoke learners' reading anxiety.Our results are similar to the findings in previous studies (Al-Shboul, et al., 2013;Huang, 2012;Kuru-Gonen, 2005).
With regard to personal factors, the significant correlation indicates that personal factors are related to learners' reading anxiety level.However, there are variances among the factors.Take lack of self-confidence for example, it is assumed that lack of self-confidence would make EFL learners feel anxious when they are reading long English texts.However, about 39.1% of participants indicated SD (Strongly Disagree), compared with 35.1% of participants who circled SA (Strongly Agree).Most of the participants disagreed that they had negative experience in reading English texts (M = 2.59).On the other hand, they were indeed worried about misunderstanding the English text (M = 3.20), which supports previous research that fear of making mistakes would cause learners' FL reading anxiety (Al-Shboul et al., 2013;Celebi, 2009;Cubukcu, 2007;Liu, 2007).Our results were slightly different from those in Huang's (2012) study, in which personal factors, such as faith and attitude in reading, or cultural background knowledge, were identified as major barriers preventing learners from understanding English texts correctly.
As for the relationship between reading strategy use and reading anxiety levels, the results show that reading strategy use is negatively correlated with reading anxiety, indicating that the more strategies the learners use, the lower anxiety level they have.This finding is consistent with that reported in Lien' (2011) study.The strategies more frequently employed were: using a dictionary, highlighting, note-taking, and translating, compared to using diagrams or context clues.Therefore, it is confirmed that reading strategy use is related to reading anxiety negatively; in other words, learners who use reading strategies less frequently would experience a higher reading anxiety level.

CONCLUSIONS
The present study was conducted to investigate the sources of reading anxiety among EFL students.This study furthers our understanding of factors associated with reading anxiety revealed in previous studies (Güvendir, 2014;Huang, 2012;Joo & Damron, 2015;Juyandegan, 2016;Kuru-Gonen, 2009;Zhou, 2017).The results yielded no connection between FL reading anxiety and gender, age or academic major.However, time spent on reading English played a considerable role in increasing reading comprehension.The results of Pearson correlation showed that overall reading anxiety was significantly and positively linked to text features and personal factors.Students reported that they feel anxious because of unknown vocabulary, and long or complicated text structure.Personal factors are a less relevant indicator of reading anxiety.Although some researchers mentioned that lack of self-esteem was a major source of reading anxiety (Güvendir, 2014;Juyandegan, 2016), in this study, fear of making mistakes was more likely to evoke reading anxiety.As the last cause of reading anxiety, reading strategy use was found to be negatively correlated with reading anxiety, indicating that learners' anxiety levels decreased as they used more reading strategies.
The findings of this study imply that one possible way to reduce reading anxiety caused by text features is to offer reading strategy instruction.Since it is inevitable that learners be required to read texts with challenging features, it is suggested that instructors provide explanation of unknown vocabulary or analysis of sentence structures.However, EFL instructors should not solely focus on teaching reading strategies, but should also create a friendly and positive learning atmosphere in order to lower the anxiety that students bring with them, thus increasing students' confidence and the pleasure of reading.Moreover, EFL learners should raise their awareness that the more time they spend on reading, the better they will know how to deal with their English reading anxiety.
To sum up, as Saito el al., (1999) stated, FL reading anxiety varies depending on the target language and writing system.Although this study sheds considerable light on the causes of FL reading anxiety, some other factors may play a role as well.Hence, more research on EFL reading anxiety is needed to address further factors that are associated with FL reading anxiety.

Table 1 . Age Distribution of the Participants
Table 1 displays the age distribution of the participants.Concerning the time spent on reading English, half of them indicated less than five hours, as illustrated in Table 2.