Effects of Affective Variables and Willingness to Communicate on Students’ English-Speaking Performance in Thailand

Recently, Willingness to Communicate (WTC) has been explored and proposed to be one of the key concepts for identifying when one decides to initiate communication in the target language. The attention, nonetheless, has been concentrated on how WTC interacts with other learners’ variables. This study, on the other hand, intends to offer empirical evidence on the direct roles of affective variables (self-confidence, anxiety, motivation, and grit) and WTC variables (WTC inside the classroom, WTC outside the classroom, and WTC in a digital setting) in Thai students’ English-speaking performance. The data were obtained from 35 undergraduate students (17.1% males; 82.9% females) using a survey questionnaire. Then, a speaking test was taken from an English Communication Skills course and was analysed using correlation and hierarchical regression. The results disclosed positive correlations among self-confidence, motivation, grit, WTC in a digital setting, WTC inside the classroom, WTC outside the classroom, and speaking performance, and negative correlations among self-confidence, anxiety, and grit. However, motivation was the only significant predictor of speaking performance. The findings offer some implications for English teachers in improving students’ English-speaking performance.


INTRODUCTION
Studies exploring factors influencing the degree of proficiency attained by individuals learning a foreign language have been conducted since the 1970s, the findings of which generally suggest two influential factors in L2 (second language) or foreign language learning: cognitive and affective variables. Cognitive variables involve aptitude and intelligence, while affective variables include individual characteristics such as attitude and motivation (Bialystok & Frohlich, 1978). According to Schumann (1975), cognitive variables are commonly stable and operate independently and naturally, but affective variables are dynamic and can be influenced by the language learning environment or setting, which offers fruitful insights into the investigations of successful L2/foreign language acquisition. Regardless, it has been found that cognitive variables have as much influence on learning as do affective variables (Chastain, 1975). The difference lies in the stage of learning development, i.e., childhood or adulthood, because affective variables will tend to play more significant roles in adult L2/foreign language learning (Taylor, 1974). Hence, given the relevance of the subjects, the investigation of the present study is centred on the roles of affective variables rather than cognitive ones.
In the past two decades, there has been a growing interest in exploring the relationships between affective variables and willingness to communicate (WTC) in L2/foreign language (e.g., Lee & Drajati, 2019;Lee & Hsieh, 2019). The assumption is that WTC is a vital variable in the interpersonal communication process as it conceptualizes the probability of initiating conversation when met with opportunities to do so (McCroskey & Baer, 1985). Moreover, exploring variables affecting learners' WTC can potentially help improve learners' WTC, which will probably increase the learners' chance of using L2/foreign language in oral communication (MacIntyre, 1994). Studies on WTC can potentially help orient theory and research to attain the ultimate goal of language learning: being able to engage in authentic communication with people from different languages and cultural backgrounds (MacIntyre et al., 1998;Waluyo, 2019).
These days, being able to communicate in English in either oral or written forms is important for both education and career. Therefore, studies examining variables that have significant correlations and can potentially become direct predictors of communicative competence and performance would be of great significance. Several studies have investigated the relationships between affective variables and learners' WTC, which have generated encouraging results for teaching and learning practices. However, there are still a few empirical studies examining the direct roles of affective variables and learners' WTC in influencing learners' English-speaking performance. The present study, hence, continues the exploration of the interplay between affective variables and WTC with the primary focus on the examination of correlations and predictive roles of affective variables and WTC in learners' English-speaking performance in Thailand. It addresses the following research questions: 1. What are the relationships between affective variables (self-confidence, L2 anxiety, motivation, and grit), willingness to communicate (inside the classroom, outside the classroom, and in a digital setting), and students' English-speaking performance in Thailand? 2. What predictive roles do affective variables and willingness to communicate play in students' English-speaking performance in Thailand?

Affective Variables
The concept of affective variables underlies those emotionally relevant characteristics possessed by learners which affect how they will respond to situations occurring in their L2/foreign language learning (Gardner & MacIntyre, 1993). The first affective variable is L2 communication self-confidence, which results from the outcomes of self-rating proficiency and the level of L2 use anxiety (Clément & Kruidenier, 1985). Self-confidence in L2 communication will only develop if one has high self-perceptions of L2 competence and low anxiety (Clément et al., 1994). Early studies have confirmed positive correlations between self-confidence and L2 language proficiency (Clément et al., 1994;Labrie & Clement, 1986). Positive correlations have also been observed among self-confidence and English-speaking skills (Gürler, 2015) and fluency (D'Amico, 2012). However, the latest study by Waluyo and Rofiah (2021) disclosed that students' self-confidence did not predict their English-speaking performance. Furthermore, there has been an increasing number of studies investigating the correlations between self-confidence and WTC. Some of the key findings indicate that L2 communication self-confidence predicts learners' L2 WTC (Fallah, 2014;Khajavy et al., 2016), positively correlates with learners' oral achievement (Pyun et al., 2014), and plays a crucial role in making learners willing to communicate in the target language (de Saint Léger & Storch, 2009;Yashima et al., 2004). Teachers who can enhance learners' L2 communication confidence will likely trigger active learner-learner interaction in a group work setting (Fushino, 2010;Ohata, 2016). Learners with high self-confidence and risk-taking features are likely to have high levels of L2 WTC outside the classroom and be willing to be engaged in frequent intercultural experiences in digital settings .
The second variable is L2/foreign language anxiety, which 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 has been considered as one contributing factor to successful speaking performance (Kasap, 2019), effects of which can be different across genders (Hwa & Peck, 2017). A low level of anxiety will likely increase learners' willingness to communicate in the target language (MacIntyre et al., 2003), and the latest meta-analysis study by Shirvan et al. (2019) has confirmed the moderating effects of language anxiety on WTC. However, in a digital learning setting, Lee and Drajati (2019), who examined the relationships between language speaking anxiety and WTC, found positive correlations, but the anxiety was not a significant predictor. Additionally, in a previous study by Alemi and Pahmforoosh (2012), learners' WTC was found to have positive correlations with language proficiency, but not with anxiety.
Motivation is the third variable, which is a multifaceted construct in which the interpretation requires specific contexts; it can be interpreted as, for instance, an effect, personality trait, set of beliefs, stimulus appraisal, etc. (Dörnyei, 1998). In the present study, the concept of motivation refers to learners' personal beliefs about their English learning and it adopts Gardner's (2007) interpretation, according to which, the term 'motivated learners' describes individuals who are motivated to learn L2/foreign language, have a genuine interest in using L2/foreign language for communication either in oral or written forms and possess a favourable attitude toward the language learning environment. The importance of motivation lies in the impact that it has on learners to stay in the learning process, despite facing obstacles and failures, until the learning goal is achieved. Nonetheless, there have been various results about the effects of motivation on learners' L2 WTC. While several studies have indicated that motivation is one of the significant predictors of learners' L2 WTC (Fallah, 2014;Lin, 2019;Shirvan et al., 2019), others have found that motivation influences learners' L2 WTC through other variables such as confidence (Peng & Woodrow, 2010), learning behaviour and resultant competence (Yashima et al., 2004). Hashimoto (2002), who examined affective variables and L2 use in classrooms of Japanese ESL students, found that motivation has a significant influence on L2 communication frequency in classrooms. In a digital learning setting, motivation can significantly predict learners' WTC (Lee & Drajati, 2019).
The last variable is grit, which is defined as perseverance and passion in the process of attaining long-term goals and is influential in the outcomes of educational attainment (Duckworth et al., 2007). In L2/foreign language learning, studies around grit contribute to the discussion of what makes learners successful and unsuccessful; in short, the grittier the learners are, the better outcomes they will achieve (Keegan, 2017). A recent study by Wei et al. (2019), who explored the relationships between grit and foreign-language performance among middle school students in China, unveils the positive effects of grit on students' performance in a foreign language; they concluded that grit not only positively supports foreign language performance but also has a crucial role in fostering a classroom environment that can potentially shape better performance in a foreign language. Previous studies have also indicated that the positive influence of grit on students' foreign language achievement will get stronger with high levels of care, support, and control from teachers (Banse & Palacios, 2018;Yoon et al., 2020). Concerning WTC, the examination of grit is often combined with other affective variables such as motivation, anxiety, and self-confidence. Empirical findings suggest that high levels of grit positively correlate with high L2 WTC inside the classroom , also that grit is one of the significant predictors of WTC (Lee & Drajati, 2019). Figure

Affective Variables
Anxiety: distinct complex of selfperceptions, beliefs, feelings, and behaviours.
Self-confidence: the outcomes of self-rating proficiency and the level of L2 use anxiety Motivation: learners' personal beliefs on their English learning.
Grit: perseverance and passion in the process of attaining long-term goals.

Willingness to Communicate (WTC)
The growing attention to the investigation of learners' WTC is closely tied to the constant search for determining variables that hold significant influences on making learners communicate orally across contexts and settings. WTC is considered a personality construct that can be affected by situational constraints, leading to the outcome of whether a person is willing to communicate with another person (McCroskey & Baer, 1985). To date, there are two areas of investigation explored by previous studies on WTC. The first area is the influence of affective variables on learners' WTC in L2/foreign language learning. Some of the key findings from previous studies have disclosed that communication apprehension and perceived competence determine an individual's WTC (Burroughs et al., 2003;MacIntyre, 1994), and four affective variables including self-confidence, motivation, anxiety, and grit, have been identified to hold significant influences on an individual's level of WTC (Lee & Drajati, 2019;Lee & Hsieh, 2019;. The other area that has been explored is the correlation between learners' WTC, frequency of communication in L2/foreign language, and reported use of L2/foreign language. MacIntyre and Charos (1996), for instance, examining affective variables as predictors of success in L2/foreign language learning communication, realized significant paths from WTC to the frequency of communication. Similarly, Hashimoto (2002) found WTC as a significant predictor of reported L2 use. Comparing these two areas, the number of studies exploring the influence of affective variables on WTC is much higher, implying that much more studies are needed to examine the direct effects of WTC on learners' use of L/foreign language in communication.
In terms of communication settings, previous studies have also focused on investigating learners' WTC in L2/foreign language inside and outside the classroom. The studies were foundationally driven by the question of why some students actively look for chances to use their L2/foreign language for interactions in class while others are reluctant to communicate, regardless of their proficiency levels; the findings were expected to help teachers facilitate WTC in the L2/foreign language classroom (Zarrinabadi et al., 2014). Some factors that have been confirmed to have significant effects on learners' WTC in the classroom encompass self-confidence, perceived opportunity to communicate, and classroom conditions including task, topic, interlocutor, teacher, and class size, coupled with linguistic factors (Cao & Philp, 2006;Peng & Woodrow, 2010). Meanwhile, high levels of L2 self-confidence and risktaking in communication have been reported to be positively correlated with higher L2 WTC outside the classroom . Similarly, Guntzviller et al. (2016) examined the role of foreign language communication anxiety in WTC outside the classroom and found a negative association between social anxiety and WTC. Exploring factors influencing learners' WTC both inside and outside the classroom seems to have attracted a considerable amount of interest from scholars as the number of studies in this area is increasing in the literature.
Nowadays, social media and online games are inseparable from student life and are starting to show their usefulness in L/foreign language learning through creative methods of teaching and learning, WTC in a digital setting is the next area of exploration apart from the inside and outside the classroom. The types of digital settings that have been explored by recent studies include online games, e-mail, and online chat (Reinders & Wattana, 2015). The findings sustain the argument that learners' WTC can be influenced in a digital setting and will likely lead to the use of the target language in real communication. Thus, Lee and Drajati (2019) developed the scales for assessing learners' WTC in digital and non-digital EFL contexts, offering a comprehensive measurement of learners' WTC in all related settings. Given the fact that most language learners are familiar with various social media, play online games, and even use online websites and applications for learning L2/foreign languages, more empirical studies investigating learners' WTC and how it links to learners' use of the target language in real communication are needed.

Research Design
This study employed a quantitative research design that involved survey and correlational analysis to address the research questions (Cook & Cook, 2008). The research was specifically designed to delve into the roles of affective variables and WTC on Thai EFL learners' English-speaking performance. The design is illustrated in Figure 2.

Participants
This study was conducted in an academic English course, English Communication Skills, aimed at enhancing both receptive and productive skills through integrated methods involving various conversation settings through role-play activities. The participants were first-year Thai EFL learners at Walailak University, Thailand. They consisted of 6 males (17.1%) and 29 females (82.9%), who came from non-English disciplines such as Nursing, Pharmacy, Architecture, and Design. The mean age of the learners was 18.54 years, varying from 18 to 20 years. The learners had more than five years of experience in learning English and when this study was conducted, it was the first time they began learning English at a university level. Learning English at the university level could give a different experience from that at elementary or high school levels in Thailand. One of the apparent differences was that at this university, the learners were taught by foreign English lecturers, including  Language and Education, 9(1), 45-61, 2022 native and non-natives; in this instance, students had no chance of communicating in their first language with the lecturers, encouraging the use of English in communication. In terms of English proficiency, most of the students were at A2 while the others were at B1 in the Common European Framework of References (CEFR).

Survey questionnaire
This study used a survey questionnaire to collect data on the students' affective variables and WTC in learning English. It adopted the questionnaire employed by Lee and Hsieh (2019) who examined the interplay between affective variables and WTC in class, outside class, and digital contexts. The affective variables consisted of four constructs, namely L2 communication self-confidence (6 items), L2 anxiety (6 items), motivation (4 items), and grit (5 items), while learners' WTC was divided into three categories, namely WTC inside the classroom (4 items), outside the classroom (4 items), and in digital settings (4 items). The responses ranged from strongly disagree (1) to strongly agree (5). The reliability of these constructs has been examined by Lee and Hsieh (2019) and several previous studies (e.g., Pyun et al., 2014), and the results showed high internal consistency among the items; thus, these constructs were utilized in this study. However, this study ran its own reliability analysis, which also served as a means of checking the consistency of the reliability results.  Table 1, the results of the factor and reliability analyses were slightly different from what was observed in the mentioned previous studies. The noticeable differences are in L2 motivation and grit in which poor and very poor internal consistencies were obtained. Initially, the α coefficients were even lower before certain items were eliminated. In L2 motivation, items 2 and 4 were eliminated; in grit, items 2, 4, and 5 were eliminated because they resulted in coefficients of less than.70 -the criterion for good internal reliability. The elimination of items also occurred in WTC inside the classroom, in which items 1 and 2 were omitted to get a higher alpha coefficient. Meanwhile, the other constructs maintained the same items from the beginning. Dunn et al. (2014) argue that fulfilling the alpha requirements has been a challenging task for constructs with a few numbers of items and reaching the minimum alpha of .70 has involved item deletion; this argument seems to be appropriate to explain what happened to motivation, grit, and WTC inside the classroom. The constructs were then continued to the next stage of the data analysis, excluding the deleted items.

English-speaking performance
The speaking test was conducted in weeks 11 and 12 of the course. The test applied an interview format consisting of three parts: 1) self-introduction, 2) talking about three-week of an independent learning experience, 3) question-answer. The teacher could also ask other questions related to the materials that students had learned in the course. One speaking test lasted 5-10 minutes. An assessment rubric, developed based on CEFR and IELTS, was used to assess fluency and coherence, lexical resource, grammatical range and accuracy, and pronunciation, encompassing pre-A1 (score 1) to C2 (score 10) levels. Descriptively, the mean of the students' scores was 6.4 at B2 level (SD = .65, Skew = -.625, Kurt = -.523). An independent t-test was performed and no significant difference between male and female students was observed (p = .686, SE = .295).

Data Analysis
The data analysis began with the identification of the learners' profiles of affective variables and WTC descriptively. Then, to answer the first question, correlational analysis was performed to seek the association between affective variables and WTC. Afterward, hierarchical linear regression with 2 and 3 models ( Table 2) was run to unveil the predictive roles of affective variables and WTC concerning the student's speaking performance. Affective variables and WTC were entered separately. Hierarchical linear regression enables us to determine if an exogenous variable's effect on a dependent variable, is 'unique' to this exogenous variable with another predictor variable (Lindenberger & Pötter, 1998). The hierarchical models followed the following orders: Table 2. The sequence of the variables in 2 and 3 models.

Students' Profiles of Affective Variables and WTC
The constructs were arranged on a five-point Likert scale from strongly disagree (1) to strongly agree (5). The interpretations were divided into three levels based on mean scores: Low (1-2.49), Moderate (2.5-3.49), and High ( . Independent t-tests were then performed, and it was observed that there were no significant differences across gender for all these variables. Figure 2 illustrates the means of affective and WTC variables.

Hierarchical Regression
All the models were run in sequence as displayed in Table 2. In two models, the hierarchical regression unveiled that motivation at stage one had a significant contribution to the regression model (F (1.33) = 6.62, p = .01), which accounted for 17% of the variance in speaking performance. This reflected that motivation was influential in the students' English-speaking performance. However, when the other three variables (self-confidence, anxiety, and grit) were added, the regression model did not reflect statistical significance (F (1.30) = 1.66, p = .19) despite having a higher R square (R 2 = .18). In fact, self-confidence, anxiety, and grit might have reduced the influence of motivation on the students' English-speaking performance. Then, in three models, at stage one, motivation still emerged as a significant predictor of speaking performance (F (1.33) = 6.62, p = .02), explaining 17% of the variance in the outcome variable. Nonetheless, non-significant results were observed in stages two and three involving the other three variables (self-confidence, anxiety, and grit), except for motivation, as seen in Table 4. The hierarchical regression also showed that none of the models (Table 2) involving WTC variables showed significant results on speaking performance.

DISCUSSION
The primary goal of this study is to explore the correlations and predictive roles of affective variables and WTC on English-speaking performance. The outcomes of the correlation analysis suggest three positive conditions. First, students who develop high levels of L2 communication self-confidence will likely be more willing to communicate in English outside the classroom. It follows the findings of recent studies (e.g., Fallah, 2014;Khajavy, et al., 2016;; further, with regards to speaking performance, an increase in self-confidence can potentially lead to student progress in English speaking skills (Gürler, 2015), fluency (D'Amico, 2012) and oral achievement (Pyun et al., 2014). Second, students who have high levels of motivation for learning English will likely make more attempts, which can lead to advancement in their English-speaking performance. In this case, the findings of the previous studies have been divided into two categories: the ones suggesting positive correlations and the ones indicating mediation. The present study endorses positive correlations between motivation and grit and motivation and speaking performance, as confirmed by Fallah (2014), Lin (2019), and Shirvan et al. (2019).
Third, students who make more attempts at learning English will likely be more willing to communicate in English in digital settings, such as on Facebook and in online games. Keegan (2017) argues that successful L2/foreign language learners possess high levels of grit, as the grittier the learners are, the better outcomes they will achieve. Wei et al. (2019), examining the interplay between grit and foreign-language performance among students in China, claim that grit positively influences foreignlanguage performance and the classroom environment; in this instance, the findings of the present study add the positive influence of grit on student communication in digital settings. The last suggestion is that students who are more willing to communicate in English inside the classroom will likely be more willing to use the target language outside the classroom. As students actively look for chances to use their L2/foreign language for interactions in class, they are likely to do so outside the classroom (Zarrinabadi et al., 2014).
Additionally, the outcomes of the correlation analysis disclose that students with low levels of anxiety in learning English will be inclined to possess high levels of selfconfidence in using English for communication; they will also tend to put more effort into their English learning, regardless of their gender. About this outcome, an early study by Clément et al. (1994) indicated a negative association between selfconfidence and anxiety, and this seems to follow human psychological common sense, but this study adds one fresh understanding of the negative relationship between anxiety and grit. At this point, students' levels of anxiety need to be kept low during the whole L2/foreign language learning process, and the previous studies by Banse and Palacios (2018) and Yoon et al. (2020) advise that the positive influence of grit on students' foreign language achievement will get stronger with high levels of care, support, and control from teachers.
Furthermore, the hierarchical regression results indicate motivation as the only significant predictor for students' English-speaking performance when selfconfidence, anxiety, and grit are involved in the regression model. In the present study, none of the WTC variables were considered as significant predictors of speaking performance. It is natural to assume that despite the claims from previous studies about the correlations between affective variables and WTC (Lee & Drajati, 2019;Lee & Hsieh, 2019; and between WTC, frequency of communication in L2/foreign language, and reported use of L2/foreign language (Hashimoto, 2002;MacIntyre & Charos, 1996), it is the only motivation that has a significant, predictive role in students' English-speaking performance. Such assumption defines motivated students as individuals who are motivated to learn L2/foreign language, have a genuine interest in using L2/foreign language for communication either in oral or written forms and possess a favourable attitude toward the language learning environment, as explained by Gardner (2007).

CONCLUSION
The present study examined the correlations and predictive roles of affective and WTC variables on speaking performance. The results disclosed five positive relationships (self-confidence and WTC outside the classroom; motivation and grit; motivation and speaking performance; grit and WTC in a digital setting; WTC inside the classroom and WTC outside the classroom) and two negative relationships (selfconfidence and anxiety; anxiety and grit). Meanwhile, motivation was the only significant predictor of speaking performance. These results sustain the findings from previous studies as discussed in the earlier section, yet it is also important to mention that these results do not support the argument that affective and WTC variables can significantly predict students' English-speaking performance.
The results of this study can also be considered as a sign of inconsistency in the studies around affective variables, WTC variables, and communication competence. Peng (2013) has brought such inconsistent results to attention by claiming, "measuring L2 WTC in EFL contexts appears to remain a challenging task" (p. 288). Peng notes that it is challenging because the developed WTC scales available in the literature apply to certain specific learning tasks and are dominated by western-originated items. Items should be scrutinized and examined before they are used to assess WTC in daily encounters in EFL contexts. Together with the growing number of empirical studies around WTC, one can notice that there are inconsistencies among the reported findings. Thus, the present study encourages more empirical studies investigating the direct effects of affective and WTC variables on L2/foreign language speaking or communication performance as it will supervise teachers specifically towards useful specific speaking tasks; one has been done by this study.
The empirical evidence obtained in this study implies that to enhance students' speaking performance, L2/foreign language teachers should include materials that can potentially enhance both students' self-confidence and motivation in English learning because escalations in these two affective variables can contribute to the enhancement of students' WTC outside the classroom and digital setting as well as encourage students to put more effort into their language learning. Besides, motivation can significantly predict students' speaking performance in English.
Given the fact that the participants in the present study are originally from Thailand, it is important to learn the findings from Vongsila and Reinders's (2016) study that suggest a vital role of teachers in encouraging WTC to make Asian learners talk. The vital role of teachers is also evident in a class full of high-performing students since teachers can create a positive emotional classroom climate that helps decrease students' anxiety levels, thereby increasing their WTC in English (Dewaele & Dewaele, 2018;Waluyo, 2020).
Regarding the limitations, it should be noted that the findings of this study may or may not represent students in another context although some results may remain consistent given the statistically proven reliability level in this study. Then, this study specifically examines students' English-speaking performance in an individual interview setting; different types of speaking tasks may or may not produce similar results. Therefore, it is recommended that future studies investigate the predictive roles of affective and WTC variables in students' speaking performance in various speaking tasks and disclose any significant differences as it will offer valuable insights for L2/foreign English teachers in attaining one of the ultimate goals of English learning, i.e., making students use English communication across contexts or settings.