Among the main language skills, the listening skill had been overlooked for years; but in the 1970s, it gradually became important (Holden, 2004). According to Chastain (1988), listening had been overlooked due to the failure of the students and teachers in understanding the “need for developing functional listening comprehension skills” (p. 192). Vandergrift (2007) also believes that this failure may be connected to the implicit nature of listening, the ephemerality of acoustic input, and the inaccessibility of listening processes. Today, however, the role of listening in EFL/ESL programs is already well recognized (Goh, 2000; Lotfi, 2012).

In the late 19th century, the importance of oral skills was emphasized by reformists, which led to the emergence of new teaching methods such as the audio-lingual method which was later completed by Communicative Language Teaching (CLT) (Goh, 2008). Later, the mechanical process of language acquisition was challenged by a new method that addressed the mental and cognitive processes (Rahimi and Katal, 2013). Some studies (e.g., Rubin, 1994; Lynch, 1998; Vandergrift, 2004; Macaro et al., 2007) shifted to new approaches influenced by developments in linguistic and cognitive psychology. Cognitive psychology emphasizes listening over speaking by highlighting the comprehensible input in listening (e.g., Krashen and Terrell, 1983).

Recently, the role of social context in listening comprehension instruction has been highlighted. According to these models, the focus should chiefly be on interactive listening in which listeners have an active role as they interact with the interlocutor. This has caused an increasing interest in investigating the role of strategy use in listening comprehension development and consequently established the theoretical foundation for strategy-based classroom instruction (Rahimi and Katal, 2013).

The second half of the twentieth century heralded a new era of focus on individual learner differences (cognitive and metacognitive differences) and their role in language learning and teaching (Kök, 2013; Whitehead, 2020). Early research was mostly based on Krashen’s idea that mere exposure to comprehensible input would increase listening skills and promote language acquisition. However, this focus has been replaced by an emphasis on how input is processed by the learners (Vandergrift, 1999). Consequently, knowing what strategies learners use and what predicaments they encounter has become a fundamental part of listening research. Listening experts argue that learners will perform better if they learn more about cognitive and metacognitive processes in listening comprehension (Flowerdew and Miller, 2005; Coşkun, 2010). In addition, some researchers believe that equipping learners with proper and effective listening strategies, such as planning, monitoring and evaluating will help them to become more efficient listeners (Vandergrift, 2002, 2004; Panggabean and Triassanti, 2020).

In the 1960s, research into Language Learning Strategies (LLS) began. This was due to the development in cognitive psychology (Williams and Burden, 1997). The primary focus was on the reports of efficient language learners on how they learn a language. These successful learners were put under observation (Wenden and Rubin, 1987). The first article published about learner strategies was in 1966 by Aaron Carton, entitled “The Method of Inference in Foreign Language Study.” Afterwards, numerous studies on learning strategies followed (e.g., O’Malley et al., 1985, 1989; O’Malley and Chamot, 1988, 1990, 2001; Goh, 1997, 2002b; Annevirta and Vauras, 2003; Chamot, 2004, 2005a,b; Habte-Gabr, 2006; Macaro, 2006; Cross, 2011; Abdul Malik et al., 2013).

The term metacognition was first introduced by Flavell (1979). He identified three types of metacognitive knowledge: person, task, and strategy knowledge. According to Wenden (1991), person knowledge is the general knowledge learners have about how learning happens and how various elements such as age and learning styles can influence language learning. In addition, it is the knowledge learners have about themselves and their beliefs about success or failure. Task knowledge refers to what learners know about the aims, demands, the nature of the tasks, and the knowledge of the procedures to complete these tasks. Strategic knowledge is about learners’ knowledge of the strategy used in order to be successful in achieving learning goals. Thus, it can be said it is the knowledge of knowing how best to approach language learning.

Metacognitive strategies are generally divided into three types: planning, monitoring, and evaluating; therefore, they help learners to understand both what they can learn (knowledge of cognition) and how they can learn it (regulation of cognition) (Eftekhary and Gharib, 2013). These strategies can be parallel to the pre-task, on-task, and post-task activities featured in many texts (Holden, 2004). Anderson (2002) emphasizes the importance of metacognitive strategies among cognitive and socio-affective strategies in developing learners’ skills. Various researchers have pointed out the importance of metacognitive strategies and their direct influence on the learning process and outcome (Pressley et al., 1987; Paris and Winograd, 1990; Victori and Lockhart, 1995; Winne, 1995; Schoonen et al., 1998; Boekaerts et al., 2000; Zimmerman and Schunk, 2001; Mokhtari and Reichard, 2002; Pintrich, 2002; Bolitho et al., 2003; Eilam and Aharon, 2003).

According to Goh (2008), listening strategy instruction can help enhance learner motivation, decrease anxiety, and benefit weak listeners. Yet studies show that what and when and how much to use strategies are related to learners’ learning styles (Hsueh-Jui, 2008) and to listening task types (Chang, 2008), as well as to learners’ level of English language proficiency (Fewell, 2010), listening ability (Bidabadi and Yamat, 2011), attitudes toward the effectiveness of strategies (Zhang and Goh, 2006), and listening anxiety (Golchi, 2012).

A lot of empirical evidence in the literature implies that employing metacognitive strategies leads to more productive listening in various contexts (O’Malley and Chamot, 1990; Thompson and Rubin, 1996; Vandergrift, 2003; Wilson, 2003; Zhao, 2013; Hariri, 2014). Related research findings also suggest that listening strategy awareness is related to learners’ level of education (Rahimi and Katal, 2012a), motivation (Vandergrift, 2005; Kassaian and Ghadiri, 2011), listening test performance (Zhang and Liu, 2008), technology use (Rahimi and Katal, 2012b), and listening self-efficacy (Rahimi and Abedi, 2014). Vandergrift (2003), for example, taught some metacognitive strategies to some beginner elementary school and university students in France. Students in both groups gained the benefits of training and they improved their listening performance.

Some related studies have resulted in contradictory and mixed findings A number of studies have shown that metacognitive instruction enhances students’ metacognitive knowledge about listening (Vandergrift, 2004; Liu and Goh, 2006); furthermore, it has a positive effect on students’ listening performance and accomplishments (Vandergrift, 2007; Vandergrift and Tafaghodtari, 2010). It also raises learners’ awareness of the nature and requirements of listening tasks along with an increase in their self-confidence while working on listening tasks (Goh and Taib, 2006). Meanwhile, other studies have indicated no significant change in listening performance (Thompson and Rubin, 1996; Seo, 2002), and in strategy employment and awareness (Chen and Huang, 2011). Vandergrift and Tafaghodtari (2010) propose that by considering the newness of the experiments in this area, more studies should be conducted in different languages, ages, genders, and learning contexts in order to settle the extant contentions.

As an attempt to suggest a model for strategy-based training, the Cognitive Academic Language Learning Approach (CALLA) was developed by Chamot and O’Malley (1986). This metacognitive strategy training model helps instructors to amalgamate language, content, and strategies in a meticulously planned lesson. In the CALLA model, the main principles are learners’ prior knowledge and their constant evaluation of their learning process. Their model consists of five phases: preparation, presentation, practice, evaluation, and expansion (Chamot and O‘Malley, 1994). The first phase is about preparing learners about their own prior strategy knowledge and the benefits of the use of particular strategies (e.g., setting goals and objectives, identifying the purpose of a language task, overviewing and linking with already known materials). The second phase is the instructor’s presenting and explaining new strategies as well as their implementation (e.g., explaining the importance of the strategy, asking students when they use the strategy, etc.). The third phase is to practice the strategies in the classroom context (e.g., asking questions, cooperating with others, seeking practice opportunities). The fourth phase includes learners’ self-evaluation of their strategy use and the results of their works (e.g., self-monitoring, self-evaluating, and evaluating their learning). The last phase is expansion, in which learners expand their strategy knowledge into new learning situations (e.g., arranging and planning their learning).

Holden (2004) explained this matter from a different perspective. He introduced a cyclical approach including three stages for a listening task: pre-listening, task-listening, and post-listening. Such an orientation, which has been used for at least two decades, is pedagogically well-founded and reliable and leads students to make consistent use of special strategies at proper times in the listening process.

Rotter (1954) introduced the social learning theory suggesting that a person’s behavior is influenced by environmental factors and social context. Rotter also emphasized that the expected results of a behavior influence the person’s motivation to engage in that behavior. The concept of locus of control or LoC was derived from Rotter’s social learning theory underlying the idea of the control over life events (Williams and Burden, 1997), and was first used by Phares (1976), then organized by Rotter (1966) (Mutlu et al., 2010).

A fundamental supposition of Rotter’s theory is that people’s behavior is determined both by nature and by the prominence of their aims or reinforcements as well as by their expectation of these aims to happen (Siri et al., 2007). According to this theory, a learner’s expectancy of a result anticipates future behavior in a given situation. Developing a locus of control idea is a good way to describe these anticipations (Siri et al., 2007). Rotter (1975) explains that LoC is a bridge to intermingle the behaviorist approach to cognitive psychology.

Based on the earlier theories introduced, people are categorized across the two ends of a continuum: internalizers and externalizers. The first group tend to attribute their outcomes and results to internal factors, to themselves, and feel personally in charge of their lives. The second group believe events happen out of their control by external factors such as fate and luck (Williams and Burden, 1997; Slavin, 2003). A great deal of study has been done on LoC as well as its connection with perceived success/failure in life and educational achievement.

According to recent studies, locus of control can be effective in a vast area of social and psychological sciences. Pavalache-Ilie and Unianu (2012) examined the relationship between LoC and the pro-environmental attitudes of undergraduate students. Results showed that internalizers have higher biocentric concerns for interventionist conservation policies. In another study carried out in Romania, Mihaela et al. (2013) concluded that LoC, teachers’ personality structure, and creative attitudes were interconnected and they could be regarded as predictors of didactic competencies.

Various researches have been conducted to explore the advantages/disadvantages of being at each end of the I-E scale. Internalizers display significant progress in areas like academic achievement, social maturity, and motivation (Lefcourt, 1992). They also are achievement-oriented, better information processors, and establish a productive relationship with the environment (Phares, 1976). In addition to the above, internal LoC leads to greater success in reading, math, and self-esteem (Nowicki and Strickland, 1973; Bulus, 2011). Externalizers, on the other hand, prefer extreme risks, are low persistent, and have unconventional changes in their attitude regarding educational, occupational, and cognitive circumstances (DuCette and Wolk, 1972). They usually are more anxious and depressed and have fewer adaptability skills (Benassi et al., 1988; Parkes, 1991). Another area to mention is the connection between LoC and perceiving others or themselves. Research shows that internalizers enjoy high self-esteem and competing spirit, while externalizers face poorer self-concept and lower self-evaluation; they also believe that they are insecure, unlucky, and inadequate (Bellack, 1975; Chandler, 1976; Burns, 1979). Furthermore, students with internal LoC are proud of their accomplishments yet shameful of their failure, and learners with external LoC show little emotional altercation in either success or failure (Mearns, 2006).

Concerning the role of LoC in second/foreign language learning, some studies have investigated it in relation to L2 learners’ reading (Ghonsooly and Elahi Shirvan, 2011; Naser and Ghabanchi, 2014; Mesgar and Tafazoli, 2018), writing (Ghonsooly and Elahi Shirvan, 2011) and listening skills (Sotoudenama and Hosseini Otaghsaraee, 2012) and academic achievement (Yazdanpanah et al., 2010). In these studies, the findings were in favor of the internal LoC regarding its role in L2 learning.

As understood from the literature reviewed above, almost no research was conducted on the role of LoC in second/foreign language learning. More specifically, its interplay with metacognitive strategies in developing the L2 listening skill is a concrete gap in the literature. Thus, this study can be considered an attempt to partially fill this gap. To this end, the following research questions were posed: (1) Does metacognitive listening strategy instruction (MLSI) make any significant difference in listening comprehension for Iranian EFL learners with internal LoC? (2) Does MLSI make any significant difference in listening comprehension for Iranian EFL learners with external LoC?

The participants of the study consisted of intermediate Iranian EFL learners (both male and female) at the University of Zanjan and Imam Khomeini International University of Qazvin, Iran. Out of 151 subjects who initially took part in the study, based on the proficiency test results, 100 intermediate learners were selected as the final participants of the study.

Two sources were used in the present study. One was language learning strategies: what every teacher should know by Oxford (1990), and strategies for success: a practical guide to learning English by Brown (2002). Oxford explains how these strategies can be directly and indirectly applied to the four language skills. Brown also provides full descriptions of specific strategies and introduces some questionnaires, which are quite useful in learning about individual characteristics of learners.

In this study, a mixture of O’Malley et al.’s (1986) CALLA (the CALLA) and Holden’s (2004) cyclical approach was used. The CALLA consists of five stages: preparation, presentation, practice, evaluation, and expansion. Holden’s cyclical approach on the other hand includes three stages: pre-listening, task-listening, and post-listening. The researchers extracted materials from Brown and Oxford’s strategy training books and then altered them in terms of students’ needs and interests. Teacher-prepared materials were taught to the experimental group in a ten-week period based on the CALLA and the cyclical approach.

To make sure that the main assumptions hold for conducting the ANCOVA analyses, the researchers ran three tests to testify the normality of the data distributions for the dependent variable, the homogeneity of variances and the homogeneity of regression. The results confirmed these assumptions holding for the data before the ANCOVA tests were run. The results appear in Tables 1, 2, 3, 4 below.

As indicated in Table 1, the data in the pre-test and post-test scores was normally distributed for both control and experimental internal LoC and external LoC groups (p-values > 0.05). Also, based on the results obtained from Levene’s tests of homogeneity of variances, this assumption held for all the data in the pre-test and post-test scores of the control and experimental internal LoC and external LoC groups (p-values > 0.05) (Table 2). Finally, as shown in Table 4, no interaction was found between the dependent variables and the covariates, indicating no violation of the regression homogeneity assumption.

The first research question was meant to investigate the effect of MLSI on the listening comprehension ability of the internal LoC intermediate EFL learners. To answer this question and compare the control and experimental group’s mean scores on the pre-test and post-test through eliminating the effect of the pre-test, a univariate ANCOVA analysis was conducted. As shown in Table 3 below, the results indicated that MLSI had a significant effect on the participants’ listening comprehension ability (F = 40.20, p < 0.05). Moreover, based on the general benchmarks suggested by Plonsky (2015) as the ratings for interpreting different effect sizes in L2 research, the effect of MLSI on this group’s listening comprehension performance turned out to be small (eta squared = 0.35 < 0.40).

The second research question was posed to study the effect of MLSI on the listening comprehension ability of the external LoC intermediate EFL learners. To answer this question and compare the control and experimental group’s mean scores on the pre-test and post-test through eliminating the effect of the pre-test, another univariate ANCOVA analysis was conducted. According to Table 4, the results indicated no significant effect for MLSI on the participants’ listening comprehension ability (F = 0.59, p = 0.45).