This chapter reviews literature related to the three learner variables explored in this study: anxiety, self-efficacy, and metacognitive awareness. In each section, the researcher first discusses main conceptualizations of each learner variable and its role in second/foreign language learning and then reviews literature on its relationship to second/foreign listening performance.
Anxiety Conceptualizations
Within psychological domain, psychologists view anxiety as a tension state, an apprehension which has some vague and indirect association with an object (as cited in Scovel, 1978). According to Spielberger (1983), anxiety is regarded as an
individual’s subjective feeling of tension, uneasiness, apprehension, nervousness, disquiet, and worry. It is this subjective feeling that alerts an individual’s autonomic nervous system (Spielberger, 1983). Because anxiety itself is a multi-faceted concept, it has been categorized into some different types, including trait anxiety, state anxiety, achievement anxiety, and facilitative-debilitative anxiety (Horwitz, 2010).
In the field of SL/FL acquisition, anxiety is viewed as the state of tension and apprehension that exclusively related to FL learning such as speaking and listening (MacIntyre & Gardner, 1994, p.284). This multi-faceted feeling has been
distinguished into different categories with various terminologies, with facilitative–
debilitative anxiety (Scovel, 1978) and state-trait anxiety (Spileberger, 1983) being two popular dichotomies.
The former categorization is based on the idea that anxiety is not always harmful
12
to learning outcomes, and in some cases it can even impose a positive influence on individual learning outcome. Hence, facilitating anxiety has positive effects on
performance while debilitating anxiety is more likely to impede one’s learning. As for the latter dichotomy, state anxiety, also recognized as transient anxiety (Horwitz, 2001), refers to some temporary and passing feeling of anxiety aroused by anxiety-provoking situations and may fluctuate and diminish over time (Papi, 2010; Sıla, 2010). Trait anxiety, on the other hand, is identified as a personality trait which is a rather stable anxiety felt by individuals regardless of threats generated by their environment (Papi, 2010; Sıla, 2010).
Further, researchers also distinguish among different kinds of anxiety associated with specific situations, that is, situation-specific anxiety. Different from trait anxiety, situation-specific anxiety may vary with situations. Language anxiety is one such situation-specific anxiety. Horwitz, Horwitz, and Cope (1986) claimed that language anxiety was associated with three kinds of performance anxiety: communication apprehension, test anxiety, and fear of negative evaluation. Communication apprehension concerns a type of fear and shyness felt by individuals when communicating with others; test anxiety, as the name suggested, is a type of
performance anxiety rooted in the fear of failure in academic evaluation situations;
fear of negative evaluation, which stems from an individual’s dread of being
evaluated negatively, may lead the individual to avoid appraisal situations (Horwitz et al., 1986, Papi, 2010; Sıla, 2010). However, Horwitz and her associates (1986) do not view foreign language anxiety as simply the sum of these three kinds of performance anxiety. Rather, they describe the construct as complicated self-perceptions or
behaviors specifically associated with foreign or second language learning (Horwitz et al., 1986, p.128). More specifically, language anxiety may be activated when one
13
learns and uses second or foreign languages (MacIntyre & Gardner, 1989).
Anxiety in Second/Foreign Language Learning
Regarding the effect of foreign language anxiety on language achievement, many studies showed a similar negative correlation. MacIntyre and Gardner’s (1989) study revealed significant and negative correlation between French (L2) performance and language anxiety (French class anxiety). Aida’s (1994) finding echoed MacIntyre and Gardner’s (1989) finding that there was a significant and negative relationship
between foreign language classroom anxiety scale (FLCAS) scores and final course grades among American second-year Japanese students. Consistent with the previous studies, Coulombe’s (2000) study revealed a negative correlation between foreign language anxiety and final course grade in French classes ranging from beginning to advanced levels.
Furthermore, Kim’s (1998) study in an Asian EFL context not only revealed significant and negative correlations between foreign language anxiety and
achievement but also indicated differences in the relationship between a conversation class and a traditional reading-based class. She found that learners were
comparatively at ease and not that nervous in the reading class than in the
conversation class. The result of the study corresponds to people’s common belief that speaking classes are more likely to arouse learners’ anxiety compared with traditional reading-based classes.
Anxiety and Second/Foreign Listening Performance
Gardner and MacIntyre (1993) claimed that anxiety is the most negative factor to language performance. A large number of studies about anxiety have laid emphasis on the speaking skill since it is considered as the most anxiety-producing skill in FL learning for many learners. However, Krashen (in Young, 1992) acknowledges that of
14
the four language skills, although speaking is often considered and mentioned as the most anxiety-provoking, listening skill may also result in high degrees of anxiety especially under the situation that listeners cannot comprehend the text. Vogely (1998) further explains the importance of managing listening anxiety: In order to maintain a conversation, the interlocutors must first have mutual understanding of each other’s speech, which could be hindered by listening comprehension anxiety (Vogely, 1998).
The need to differentiate language anxieties associated with different language skills is most clearly articulated in Cheng, Horwitz, and Schallert (1999). They posit that the differentiation helps educators or researchers to identify more precisely the source of anxiety that negatively impacts learning and/or using of a specific language skill. Pae (2013) further examined the intra-relations and inter-relations between the four skill-based SL anxieties and general foreign language anxiety with 285 Korean university EFL students. The study proved that each of the four skill-specific anxieties was unique and different from each other statistically. Moreover, each of the skill-specific anxieties served as an independent variable contributing to general classroom anxiety. These studies justify investigations of listening anxiety as an independent construct.
In comparison with research on general foreign language anxiety, studies
concerning SL/FL listening anxiety are rather limited (Elkhafaifi, 2005; Zhang, 2013).
Elkhafaifi (2005) was the first one to investigate the relationship between general FL anxiety and FL listening anxiety and to associate both anxieties with overall FL achievement and listening comprehension performance respectively. To be more specific, he recruited 233 postsecondary students of Arabic as a FL from six US universities, ranging from juniors to seniors and even to graduates. All the participants were non-Arabic majors. The measures of their overall FL performance and listening
15
comprehension achievement were the final course scores and listening comprehension grades respectively provided by the instructors. As for the assessment of general anxiety and listening anxiety, the FLCAS designed by Horwitz et al. (1986) served as the measurement instrument for the former anxiety and the Foreign Language
Listening Anxiety Scale (FLLAS) the latter. The FLLAS was adapted from Foreign Language Reading Anxiety Scale (FLRAS) developed by Saito, Garza, and Horwitz (1999). The results revealed that FL anxiety and FL listening anxiety were two separate constructs but were in positive correlation with each other (r = .66, p < 0.1).
That is, students with higher degrees of FL listening anxiety were prone to have higher degrees of FL anxiety and vice versa. In regard to the relationship between anxiety and performance, the data evidenced that both types of anxiety (i.e., FL anxiety and FL listening anxiety) were significantly correlated with final course grade and listening comprehension score. To be clearer, FL anxiety was significantly and negatively correlated with overall performance (r = -.54, p < .01) and listening comprehension achievement (r = -.53, p < .01), indicating that students with higher levels of FL anxiety had lower grades on overall achievement and listening
performance than those with lower FL anxiety. FL listening anxiety was also negatively correlated with general performance (r = -.65, p < .01) and listening comprehension achievement (r = -.70, p < .01). Students who experienced higher levels of FL listening anxiety had lower final course grades and listening
comprehension scores than those with lower degrees of FL listening anxiety. Apart from the finding regarding anxiety and achievement (general and listening
performance), Elkhafaifi’s (2005) study also revealed a significant and negative relationship between FL listening anxiety and students’ year of study in school. That is, the older/more advanced students experienced lower levels of FL listening anxiety.
16
However, gender and FL listening anxiety were not correlated with each other statistically. In other words, gender was not a factor affecting one’s level of listening anxiety (Elkhafaifi, 2005).
Similar finding on the relationship between FL listening anxiety and listening performance can be found in Zhang’s (2013) research. Specifically, Zhang (2013) further addressed the causal effect on FL listening anxiety and listening performance.
In the study, Zhang included 300 first year English majors in a university in China.
All the participants filled in the FLLAS and took the IELTS listening test twice. The FLLAS was a replication of the FLLAS by Elkhafaifi (2005) but the version tapped into English listening classrooms rather than Arabic classrooms. To ensure that each item in the scale was clear enough to each participant, all the content of FLLAS was translated into Chinese by Zhang. The result indicated that FL listening anxiety was negatively related to FL listening performance. Moreover, the causal relationship found in the study showed that FL listening performance was statistically affected by FL listening anxiety, but not vice versa. Zhang (2013) further claimed that the
combination of self-efficacy and FL listening anxiety might lead to more powerful prediction on listening comprehension performance.
Likewise in Taiwan, studies on FL listening anxiety and FL listening
performance evidence that FL listening anxiety was in a significant and negative relationship with listening achievement. Cheng’s (2006) research on the effect of listening anxiety on EFL senior high school students’ listening performance on different question types (i.e., global / local questions), text types (i.e., statements, conversations, and short talks), and strategy use revealed results consistent with the previous studies. In the study, Cheng (2006) recruited 132 senior high school students, who were divided into high/low anxiety groups and high/low proficiency
17
groups based on their responses to a modified FLLAS from Elkhafaifi (2005) and a listening test. Apart from the use of the FLLAS questionnaire, Cheng further adopted an interview from different groups of participants to elicit more information. The results revealed that listening anxiety was negatively correlated with the participants’
listening performance significantly (r = -.29, p < .01), which was in line with Elkhafaifi’s (2005) and Zhang’s (2013) findings. In addition, the low-anxiety group outscored high-anxiety group on overall listening performance (t = -3.56, p < .01) and global questions (t = -4.20, p < .01), but not on local questions (t = -1.87, p
> .05). Regarding the text types, students in low anxiety group significantly
outperformed the participants in high anxiety group on statements (t = -3.43, p < .01) and short talks (t = -2.18, p < .05), but not on conversations (t = -1.73, p > .05). With respect to the listening strategy, it was found that the low-anxiety group tended to use all the strategy types (i.e., memory, cognitive, metacognitive, compensation,
affective, and social) more frequently than the high-anxiety group, except for
cognitive and affective strategies. Besides, the more advanced listeners were prone to apply more strategies to the listening task than their counterparts.
Another study conducted by Su (2007) revealed similar results on FL listening anxiety and FL listening comprehension attainment. The study aimed to explore factors which affect listening comprehension anxiety and the effects of listening anxiety on Taiwanese EFL college students’ listening comprehension. Su (2007) involved 1772 non-English-major freshmen, from various departments. The results not only showed that the participants’ listening comprehension anxiety was in significant negative relation to listening comprehension performance (r = -.54, p
< .01), but also evidenced that listening comprehension anxiety was the most dominant factor affecting participants’ listening comprehension performance
18
compared with the other two factors (i.e., participants’ personality and linguistic intelligence).
Cheng’s (2005) research was in line with the aforementioned studies. The study investigated the sources of and change in EFL leaners’ listening comprehension ability over one semester. A class of 23 English-major freshmen and sophomores at a university in northern Taiwan were recruited to participate in the study. An open-ended questionnaire and two one-item rating scales (one for retrieving participants’
anxiety and one for assessing self-efficacy belief) were used as instruments in the study. The participants were asked to fill out the open-ended questionnaire right after every class meeting throughout the whole semester. The other two one-item scales were given to the participants at the beginning, the middle, and the end of the semester. Apart from the negative association between final course grades and listening anxiety reported at three points of time throughout the semester (r = -.14, r
= -.53, r = -.31), the results further revealed an even stronger association between listening anxiety and listening self-efficacy (r = -.22, r = -.64, r = .61). Moreover, it also showed a decline of learners’ listening anxiety levels with time passing, but the difference failed to reach statistically significant level.
Compared to the large amount of research on general language anxiety and overall L2 performance, studies probing into the relationship between FL listening anxiety and FL listening comprehension are rather insufficient. Although in Taiwan Cheng (2006), Su (2007) and Cheng (2005) have shed some light on this area, their participants only came from a senior high school and two universities respectively.
Studies concerning participants from junior high school are rare, let alone studies conducting on both junior high and senior high levels. Hence, this study examined the relationship between listening comprehension and listening anxiety across the two
19
education levels.
Self-efficacy Conceptualizations
Self-efficacy is a construct grounded in socio cognitive theory (Bandura, 1986), which studies human cognition, actions, motivation and emotion, assuming people’s ability to self-reflect and self-regulate. The core of social cognitive theory lies in
“triadic reciprocality” the interaction among environmental, personal, and behavioral effects (Bandura, 1986). Based on this theory, one’s behavior is the outcome from the interaction of the three dimensions. And this outcome is used to form expectations of one’s future behavior. Individuals then form a self-system to control over what they think, believe, and feel, through which one interpret and decode the outcomes of their behavior (Bandura, 1986; Pajares, 1997). This interpretation affects their self-beliefs, which, in turn, influences their subsequent behaviors. Accordingly, individuals’
behaviors are under the impact of the beliefs they hold about themselves.
Self-efficacy is defined by Bandura (1986) as people’s own perception of their abilities to frame and perform courses of action required to achieve designated types of performance. Further, it is not related to the skills individuals have but to the perception and judgments of what individuals can do with the skills they possess.
Therefore, the foundation of personal efficacy beliefs lies in what people believe they might be accomplished with their possessed skills rather than in their actual abilities.
According to Bandura’s (1986) social cognitive theory, self-efficacy serves as the dominant element in human behavior. It imposes great influences on the decision individuals make, on the effort they are willing to pay in given endeavors, the persistence they hold against obstacles, and the proportion of tension or confidence they will hold in the face of challenges (Bandura, 1986). Further, Bandura (1997)
20
posited that individuals’ personal efficacy mainly formed the dominant factor to human agency, without which one may have little motivation to perform the courses of action. It is the different degrees of perceived personal belief one possesses that distinguishes people’s behaviors even when they are equipped with similar skills and abilities (Rahimi & Abedini, 2009).
Moreover, according to Bandura (1986), self-efficacy affects human behavior in four ways. First, it influences choice of behavior. Bandura (1994) claimed that
people’s evaluation of their capability corresponds to the performance they expect and to the abilities they seek and require. Second, self-efficacy determines how much effort and persistence individuals will give and hold on each activity. For example, a low self-efficacy student has no confidence in himself/ herself might give up
immediately in the confrontation of obstacles and refuse to take challenges. Hence, a higher degree of self-efficacy contributes to a greater endeavor, expenditure, and endurance. Third, self-efficacy has great impact on individuals’ emotional reactions and thought patterns. People with a low sense of self-efficacy, for instance, might consider things more difficult and tougher than they really are. As a result, this thought generates stress and might make individuals ascribe their loss or failure in challenging tasks to insufficient ability instead of to deficient efforts. The last point self-efficacy affects human behavior is by recognizing humans as producers rather than simply foretellers of behavior.
Many studies have provided evidence that self-efficacy is a strong predictor to academic performance. Lent, Brown, and Larkin’s (1986) study revealed that self-efficacy has a significant unique contribution to the prediction of final grades and learners’ perseverance within science field. Meece, Wigfield, and Eccles (1990) conducted a study on affect variables contributing to math achievement among
21
adolescent learners. The results indicated that those young learners’ self-efficacy had great impact on their math performance expectancy, which then influences their math performance. Intriguingly, math anxiety was proved to be not directly related to math attainment. Instead, learners’ perceived self-efficacy belief served as the mediator to lower down the impact of anxiety on math achievement.
Self-efficacy in Second/Foreign Language Learning
In SL/FL learning field, several researches have shed light on the role of self-efficacy in language learning (Cheng, 2002; Mills, Pajares & Herron, 2007). Cheng’s (2002) study revealed that individual’s self-perceptions of confidence in English writing explained the largest amount of variance in SL writing anxiety (34%). Mills, Pajares, and Herron’s study (2007) also probed into the effect of self-efficacy and other motivational factors on the achievements of college students who learned French as a second language. The results evidenced that compared to French anxiety, self-efficacy contributed more to language achievement. To be more specific, their study indicated that self-efficacy for self-regulation contributed more to French language achievement than French learning self-concept, French anxiety in reading and listening, and self-efficacy to obtain grades in French. Moreover, learners who considered themselves with a stronger sense of deploying metacognitive strategies to supervise and monitor their academic work were in higher possibility to succeed in intermediate French.
Though much research has revealed that anxiety is in significantly negative relation to language learning and performance (Horwitz, 2001) and is one of the leading predictors of SL acquisition (Horwitz, Horwitz, & Cope, 1986), social cognitive theorists (Bandura, 1997) stress the dominant role self-efficacy plays in performance compared with other motivational variables. In social cognitive theory,
22
self-efficacy belief is viewed as the main factor to determine learners’ anxiety. The extent to which learners suffer from anxiety largely rely on their self-efficacy beliefs (as cited in Mills et al., 2006). Bandura (1997) described anxiety as the state of tension over the potential detrimental outcomes. The arousal of anxiety mainly results from the imbalance of people’s perceived self-efficacy belief and the possible harmful happenings (Bandura, 1991, as cited in Chen, 2007). Apprehensive thinking occurs when learners possess a weak sense of self-efficacy, resulting in the arousal of
anxiety. In contrast, learners who possess with more self-confidence are more likely to hold more optimistic expectations in the face of difficult situations, which helps to lower down the anxiety level. Thus, learners with a stronger sense of self-efficacy belief often change the difficult situations into satisfying outcomes. In other words, anxiety arousal occurs only when SL/FL leaners’ perceived values of themselves are
anxiety. In contrast, learners who possess with more self-confidence are more likely to hold more optimistic expectations in the face of difficult situations, which helps to lower down the anxiety level. Thus, learners with a stronger sense of self-efficacy belief often change the difficult situations into satisfying outcomes. In other words, anxiety arousal occurs only when SL/FL leaners’ perceived values of themselves are