Introduction

There is a substantial body of research that indicates beliefs about how individuals approach learning a language lead students to different paths of motivation, engagement, or academic outcomes (see Bai and Wang, 2023; Dörnyei and Ushioda, 2011; Fathi et al., 2024; Lou and Noels, 2017; Sadoughi et al., 2023; Wang and Peverly, 2018, etc.). In the last decade, considerable research has focused on language mindsets, beliefs about the malleability of personal qualities related to language learning (Lou and Noels, 2017), as they have been thought to be one of the influential beliefs held by students. This concept was originally taken from implicit theories in psychology as a vital framework to examine individuals’ motivation and achievement in an academic setting and reconceptualized in the context of language learning. Lou and Noels (2017) constructed The Language Mindset Inventory, and studies employing it have produced meaningful results that shed light on the mechanisms involved in second or foreign language (L2) learning.

Several studies have explored how language mindsets and self-efficacy, defined as individuals’ judgments of their capabilities to complete specific tasks (Bandura, 1997), influence the language learning process, primarily by examining their effects in isolation (e.g., Bai and Wang, 2023; Cho et al, 2021; Fathi et al., 2024; Zarrinabadi et al., 2022). Although these studies offer invaluable information to understand L2 learning mechanisms, the potential effects of their interplay and synergistic influence remain largely unexplored. Similarly, the study by Bai and Wang (2023) underscores the need for future investigation to delineate the distinct roles of a growth mindset and self-efficacy. They highlight the necessity for further exploration into the unique characteristics and contributions of these factors to the process of acquiring L2.

Therefore, this study attempts to delve deeper into the interplay between language mindsets and self-efficacy with a special focus on the motivational aspects of these variables in the L2 learning process, along with engagement and perceived proficiency (PP). Engagement reflects the level of effort and enthusiasm that students invest in their learning process. In particular, this study employs PP, which refers to students’ subjective judgment of their proficiency in English. While previous studies have provided important insights into the impacts of language mindsets and self-efficacy on actual language achievement scores (e.g., Bai and Wang, 2023; Fathi et al., 2024), this research aims to go beyond these insights by focusing on the relationship between these variables and PP. This shift in focus allows us to explore the motivational dynamics in language learning from a different perspective, as PP can significantly influence students’ intrinsic motivation and behavior more than actual language proficiency scores do (Agawa and Takeuchi, 2016; Tanaka, 2023). PP can potentially serve as a driving force for their actions irrespective of the actual reality (Du, 2015). Additionally, Mercer and Ryan (2010) suggest that PP should be considered when examining language mindsets. Using PP as a dependent variable in this study is crucial for understanding how students’ perceptions of their proficiency are shaped by their mindsets, self-efficacy, and engagement, offering a novel perspective on the motivational dynamics in language learning.

Overall, investigating the impact of key language learning variables, such as language mindsets and self-efficacy—especially when considered together—on PP, may provide deeper insights into the intricate dynamics of the L2 learning process. This approach not only enhances our understanding but also guides the development of more effective and strategic pedagogical implications.

Literature review

Language mindsets

Mindsets have been widely studied in psychological and other academic contexts (Zarrinabadi and Lou, 2022). According to Mercer and Ryan (2010), mindsets represent beliefs that individuals have about the possibility of development or change in personal qualities such as aptitude, intelligence, and personality. In general, mindsets are viewed in two ways: a fixed mindset and a growth mindset. Those with a fixed mindset may perceive these qualities as being static or fixed and having limited potential for improvement. This may lead to them not actively committing to opportunities for change. On the other hand, those with a growth mindset believe that people possess an inherent capacity to alter and modify personal characteristics for growth and development, which may assist them in putting effort into growth. Many studies support this two-factor structure of mindsets in language learning (Khajavy et al., 2021, 2022; Lou and Noels, 2017; Papi et al., 2021; Yao et al., 2021; Zarrinabadi et al., 2022., etc.). To examine language learners’ mindsets in detail, Lou and Noel (2017, p. 215–216) constructed a language mindsets inventory based on three major categories of beliefs for L2 learning, that is, general language intelligence, language aptitude, and age sensitivity for use. Khajavy et al. (2022) emphasize considering both aspects when examining the effects of language mindsets to gain a more profound understanding. Using this instrument, many studies have examined how language mindsets are related to important variables in the learning process (Khajavy et al., 2021; Lou et al., 2021; Lou and Noels, 2017; Yao et al., 2021, etc.). For example, individuals are likely to choose different sets of goals and strategies based on their mindsets. Those who view language learning through the lens of a growth mindset tend to show reactions such as persistence in challenging situations. On the other hand, those holding a fixed mindset are likely to exhibit reactions such as abandoning additional endeavors or avoiding challenges (see Dweck, 2012; Lou and Noels, 2017). In this perspective, language mindsets have been used as a motivational variable in several studies (Bai and Guo, 2019; Bai and Wang, 2023; Papi et al., 2021). Papi et al. (2021) perceive language mindsets as part of motivation theory in consideration of their impact on the kinds of a goal pursued, the amount of investment and effort placed, and the way to perceive the outcomes of these pursuits. Lou et al. (2021) viewed mindsets as ‘deep-seated ontological and epistemological beliefs’ and claimed that they need to be examined as a system along with other related variables.

An individual’s mindsets are seen as domain-specific (Burnette et al., 2013; Cho et al., 2021; Khajavy et al., 2022; Mercer and Ryan, 2010; Lou and Noels, 2019). In a cross-sectional study by Khajavy et al. (2022), the roles of language mindsets were examined within an EFL Iranian context. They found that L2 reading achievement was predicted by both an L2 growth reading mindset and a domain-general growth mindset. Conversely, Cho et al. (2021), which focused on sixth-grade students in the United States over a year, did not find a significant link between a reading growth mindset and reading achievement. However, they found that a domain-general growth mindset was a significant predictor of reading achievement. While there are differences in the duration of the studies and the student populations examined, the results imply the significant impact of a domain-general growth mindset on language learning outcomes. Notably, this type of mindset consistently showed more robust results across different contexts.

Such consistency of a domain-general growth mindset is supported by a few studies. For example, the findings of Khajavy et al. (2021) revealed that a domain-general growth mindset was a weak but significant predictor of L2 achievement. Similarly, Lou et al. (2021) found that students with a growth profile, characterized by a growth mindset and learning-related variables achieved higher than those with mixed and fixed profiles. The study by Shirvan et al. (2024) offers a comprehensive analysis of the effect of a growth mindset on actual language achievements. Employing a multilevel meta-analysis of 22 studies and 50 effect sizes, it revealed the patterns between language mindsets and L2 achievements that were measured across various domains, including overall performance, grammar, pragmatics, and specific skills in reading, speaking, and writing. Their findings suggest that a growth mindset is positively and significantly associated with actual language outcomes, exhibiting a moderate effect size (r = 0.26). Conversely, a fixed mindset was linked to them negatively, with a relatively small effect size (r = −0.19). These findings highlight the pivotal role of a growth mindset in improving actual language performance.

To sum up, it appears that language mindsets undoubtedly shape students’ language learning paths, indicating their close link with self-regulatory processes. Therefore, to gain a more comprehensive understanding of language learning, further investigation into how language mindsets interact with self-efficacy is critical. Such an investigation will broaden our insights into the effects of their interplay within the language-learning process. Additionally, this study aims to investigate the effects of a domain-general language mindset on the English learning of Korean university students. While language mindsets are typically domain-specific, focusing on domain-general mindsets allows us to grasp broader motivational dynamics that influence various aspects of language learning, not just specific skills like reading or writing. Domain-general mindsets provide a more holistic view of students’ overall attitudes toward learning and their belief in the malleability of their language abilities.

Self-efficacy

While language mindsets are the beliefs held by individuals regarding the development potential (fixed, or malleable by effort), self-efficacy refers to “judgments of personal capabilities” to accomplish given tasks or goals (Bandura, 1997, p. 11). It involves individuals’ expectations and convictions in their abilities (Bong and Skaalvik, 2003), which does not necessarily reflect mindsets (Cho et al., 2021) and is a distinct construct from mindsets (Lou and Noels, 2019). There appears to be a consensus that self-efficacy regulates motivation and thought patterns, which results in specific behavioral outcomes. Many studies have focused on the self-regulatory capabilities of self-efficacy beliefs in language learning, revealing how closely they are related to motivations and achievements (Anam and Stracke, 2016; Hsieh and Schallert, 2008; Kim et al., 2015). Students’ self-efficacy beliefs in their capabilities to regulate their own learning can possibly offer valuable information about their motivations, behaviors, and future choices (Usher and Pajares, 2009). This indicates that self-efficacy plays an important role in effectively facilitating language learning. Notably, the study by Fathi et al. (2024) identified significant direct links from growth mindset to L2 achievement (measured by IELTS scores) and from self-efficacy to L2 achievement. The results show that the connection between self-efficacy and L2 achievement was stronger than the connection between a growth mindset and L2 achievement, suggesting that self-efficacy is a more reliable indicator of L2 achievement than a growth mindset.

Importantly, language mindsets may influence one’s level of self-efficacy. Although not in the context of language learning, Jourden et al. (1991) reported that those with a fixed mindset were not related to growth in self-efficacy throughout the different learning stages, while those with a growth mindset showed enhancement in self-efficacy. Supporting this, the language mindsets study by Zarrinabadi et al. (2022) reported that there is a direct impact of having a growth mindset on self-efficacy. In the study by Bai and Wang (2023), they categorized a growth mindset and self-efficacy as motivating variables and examined how they were related to learning strategy use, such as monitoring and effort regulation; both a growth mindset and self-efficacy as independent variables strongly predicted the learning strategy use. These studies provide important insights into these variables. However, they tend to either focus on the isolated effects of a growth mindset and self-efficacy or overlook the synergistic effects between them. Notably, Jourden et al. (1991) argued that self-efficacy may mediate the effect of mindsets in the areas of endeavor or learning outcomes, given its regulating functions in learning. However, the mediating role of self-efficacy in the relationships among language mindsets, engagement, and PP is not yet well understood. Therefore, this study aims to explore these interplays, considering self-efficacy as a potential mediator.

Engagement and self-perceived proficiency

For meaningful learning, engagement is critical in that students who are highly engaged actively participate in and are dedicated to their own path of learning (Hiver et al., 2024). Engagement refers to the degree of effort and enthusiasm students put into their learning process, which is indicated by “energy, dedication, and absorption” (Salmela‐Aro and Upadyaya 2014, p. 139). In L2 learning, Lou et al. (2021) found that engagement had a relationship with mindsets and various ‘language mindset-related constructs’ within self-regulatory domains. Interestingly, the degree of engagement varied depending on the student mindset profiles (in the order of growth, mixed, and fixed), which indicates its close relationship with language mindsets. Similarly, the study by Sadoughi et al. (2023) was carried out in an attempt to explore the underlying processes and conditions through which a growth mindset influences engagement and learning outcomes, focusing on motivational and situational factors as key mediators and moderators. Their findings suggested that a growth mindset influenced engagement through motivating factors. Additionally, positive learning experiences moderated the impact of a growth mindset on engagement. Lou et al. (2021) maintain that it is important to analyze how engagement interacts with other significant factors in concert, rather than examining its relationship with a single variable in language learning. In other words, it is essential to examine not only the relationship between language mindsets and engagement but also how they interact with other crucial variables in language learning.

Regarding L2 proficiency, Dörnyei and Ushioda (2011) indicate that students’ proficiency has a close relationship with achievement. As Du (2015) argues, the perception of their L2 proficiency is true for students and appears to serve as a driving force for their language learning behaviors despite its weak correlations with actual achievement. The findings of several studies reveal that students’ PP is closely linked to their motivation to learn a target language. Particularly, a few studies examined if PP is related to motivational indicators of self-determination theory (Agawa and Takeuchi, 2016; Bureau et al., 2022; Noels et al., 2020; Tanaka, 2023). Noels et al. (2000) found that PP is correlated with intrinsic motivation, a more self-determined type of motivation, which is in line with a meta-analysis study by Bureau et al. (2022). Tanaka’s (2023) study which examined the data produced by Chinese students studying Japanese also showed that PP was a stronger predictor of intrinsic motivation than actual L2 achievement. Such studies inform that students’ perception of their proficiency plays a crucial role in L2 learning and serves as a meaningful and motivating variable. The choice of PP as a dependent variable is based on existing research highlighting its significant role in influencing learners’ motivation and behaviors (Agawa and Takeuchi, 2016; Tanaka, 2023).

When considering language mindsets, Mercer and Ryan (2010) suggest that perceived proficiency (PP) is likely to influence how individuals perceive their inherent qualities related to language learning. Lou and Noels (2017) argue that students with a growth mindset tend to put in more effort in the learning process, view failure as an opportunity to learn, and persist in their efforts. For example, motivation and resilience are enhanced for those who exhibit a growth mindset in the course of learning (Lou and Noels, 2019). Conversely, students with a fixed mindset interpret failure as a sign of their lack of ability, underestimate the effects of effort, and struggle to achieve their learning goals. The type of mindset students hold ultimately leads to different learning paths. As a result, individuals with a growth mindset tend to have a more positive perception of their abilities, while those with a fixed mindset may have a negative perception of their capabilities. Despite the apparent close relationship, information on the relationship between language mindsets and perceived proficiency remains scarce. Although there are some findings report a relatively positive link between a growth mindset and actual L2 achievements, the relationship with PP remains unclear. When considering that PP is likely to serve as a stronger intrinsic motivator than actual achievement (Tanaka, 2023), this gap signifies an important area to explore how a growth mindset influences students’ perceptions of their language skills.

Taken together, it is urgent to understand the effects and how the aforementioned crucial factors interact with one another. Therefore, this study aims to explore language learning variables such as language mindsets, self-efficacy, engagement, and PP and explore their relationships and effects on L2 learning.

The present study

The current study seeks to explore the interplay and effects on language mindsets, self-efficacy, engagement, and PP shown by Korean EFL university students. In order to achieve the research objectives, this study first examines whether there are any direct relationships between the variables. Then, it investigates the inter-relationship of language mindsets, self-efficacy, engagement, and self-PP. The study involves the following two research questions.

RQ1: How are language mindsets related to self-efficacy, engagement, and PP?

RQ2: What indirect effects do language mindsets have on engagement and PP through self-efficacy in language learning?

Methods

Participants

The data for this study were collected by a convenient sampling and the survey was conducted in three universities located in Seoul, Gangnueng, and Kyungbuk. English was a required course for these universities. A total of 257 participants from a variety of majors (male = 118, female = 139) joined the study voluntarily. They learned of the objectives and content of the study and the guidelines about how to respond to the online survey from their instructors. Participants were informed that they could quit answering the survey at any time if they felt uneasy. The survey was administered when their classes were over; ~10 min after the class. Those who participated in the study received a small gift.

The majority of them were freshmen who were taking English courses such as English reading, writing, conversation, or TOEIC. The average age of the participants was 19.81 years (SD = 1.86). According to their responses to PP in the four English skills, they could be regarded in the low-intermediate level of English: reading (M = 3.62, SD = 1.23), Writing (M = 2.98, SD = 1.10), Listening (M = 3.55, SD = 1.26), Speaking (M = 2.88, SD = 1.18). As is typical of EFL students, the average scores for receptive areas like reading and listening hovered around mid-point 3.5, whereas the scores for the productive areas like writing and speaking were below 3.0. The average score of the overall four skills was 3.26 (SD = 0.97).

Instruments

This study employed the following instruments, and the average scores from each tool were analyzed.

  • Perceived proficiency: The participants’ perceived English proficiency was assessed in the four skill areas mentioned by asking questions like “How would you assess your English reading proficiency?” They responded on a six-point Likert scale from 1 (none at all) to 6 (highly proficient) based on their judgment.

  • Language mindsets: The Language Mindsets Inventory (Lou and Noels, 2017) was employed with nine items each that correspond to fixed and growth mindsets. These items measure three major categories of beliefs for L2 learning such as general language intelligence, language aptitude, and age sensitivity for use. The average scores associated with fixed and growth mindsets were used to identify the features and characteristics of these mindsets.

  • Self-efficacy: Five self-efficacy items were adapted from the study by Bong (2001). These items reflect their subjective beliefs about individuals’ capabilities of performing a relevant task in an English course (I can master even the hardest material in English if I try. I’m certain that I can do an excellent job on the problems and tasks assigned for English class., etc.).

  • Engagement: The Foreign Language Course Engagement was adapted from the study by Salmela‐Aro and Upadyaya (2014). Overall, nine items were categorized into energy, dedication, and absorption. It measures students’ level of commitment and enthusiasm for the learning process to acquire English language proficiency. Participants were instructed to answer questionnaire items for language mindsets, self-efficacy, and engagement on a six-point Likert scale from one (never agree) to six (strongly agree).

Procedures

The purpose and outline of the study were presented to four instructors, including the author of the study, from three universities, all of whom were teaching English courses. They agreed to participate in the study and received the detailed guidelines and information required to carry out data collection within their classrooms. All the survey items were translated into Korean and reviewed meticulously by the instructors prior to the administration. The instructors encouraged students in their classes, ultimately overseeing the data collection process.

Data analysis and the hypothesized model

Correlations, descriptive statistics, and internal consistencies were reported using SPSS version 25. Exploratory factor analyses (EFA) were conducted to verify the construct validity of the questionnaires used in the study. In order to explore the direct and indirect effects of the variables, path analysis was performed, employing LISREL 8.80.

Figure 1 shows the theoretical model of the study. The model to be tested shows that self-efficacy and engagement mediate the links between language mindsets and PP. The path that connected a fixed mindset and PP was removed because it did not support casualty in a few studies (e.g., Khajavy et al., 2021, 2022).

Fig. 1
figure 1

A hypothesized path diagram for language mindsets, self-efficacy, engagement, and perceived proficiency.

Full size image

Results

Preliminary analyses

  • Exploratory factor analyses for construct validity: To identify how the items of language mindsets clustered, EFA using principal axis factoring and direct oblimin rotation was performed. It resulted in a two-factor solution with eigenvalues greater than 1.0 (loading value greater than .40), accounting for 52.50% of the total variance. All nine items for a growth mindset loaded on Factor 1, which explained 40.90%. Eight items for a fixed mindset loaded on Factor 2 except for one item (“How well a person speaks a foreign language depends on how early in life he/she learned it.”), which explained 11.60% of the total variance. They were labeled as a growth mindset and a fixed mindset because the attributes to which each group of the items belonged remained intact.

Another EFA using principal axis factoring and direct oblimin rotation with self-efficacy and engagement resulted in a two-factor solution with eigenvalues greater than 1.0 (loading value higher than 0.40), accounting for 57.55% of the total variance. All nine items of engagement loaded on Factor 1, which explained 46.92%, and all five self-efficacy items loaded on Factor 2 which explained 10.63%. These two factors were also labeled as engagement and self-efficacy due to their intact attributes. Regarding PP, EFA using principal axis factoring and direct oblimin rotation with listening, speaking, reading, and writing resulted in a single-factor solution, accounting for 55.62% of the total variance.

  • Descriptive statistics, reliability, and correlation analyses: The descriptive statistics are shown in Table 1. The mean score for a growth mindset was a lot higher (M = 4.69, SD = 0.92) than the mid-point (3.5), while the score for a fixed mindset was relatively low (M = 2.81, SD = 0.88). The score for self-efficacy was higher than the mid-point (M = 4.09, SD = 0.95), followed by that of engagement (M = 3.55, SD = 1.04). The participants appeared to believe that they had a relatively moderate level of perceived English competence (M = 3.26, SD = 0.97). All variables showed an internal consistency of 0.82 or higher, which is considered high (see Table 1).

    Table 1 Descriptive statistics (n = 257).
    Full size table

The correlation and covariance matrix are offered in Table 2. It appears that a growth mindset was negatively associated with a fixed mindset (r = −0.46, p < 0.01). While a growth mindset was positively related to all the other three variables, a fixed mindset had a negative relationship with self-efficacy (r = −0.23, p < 0.01) and did not have any significant correlations with the other variables. In particular, self-efficacy showed high correlations with a growth mindset, engagement, and PP (r = 0.55, 0.50, and 0.47, respectively, all p < 0.01). In addition, self-efficacy exhibited the highest correlations with PP (r = 0.47) than the other two variables, a growth mindset (r = 0.25) and engagement (r = 0.35).

Table 2 Covariance and correlation matrix (n = 257).
Full size table

Path analyses

The univariate normality of each variable was checked with the mean values of Skewness and Kurtosis (see Table 1). Since the values of Skewness and Kurtosis were less than ±1.0, the normality assumption was considered supported (Curran et al., 1996). Thus, parameters for the path analysis were estimated using the Maximum likelihood method. The overall goodness of the structural model was evaluated using the fit indices suggested by Byrne (2001): χ2/df testing, comparative fit index (CFI), root mean square error of approximation (RMSEA), and standardized root square residual (SRMR) were used for the evaluation of the structural model. χ2/df below 5 is considered an acceptable fit and below 3 is an excellent fit. CFI above 0.90 is an acceptable fit and above 0.95 is an excellent fit. A fit is considered acceptable if RMSEA and SRMR are less than 0.08, while a fit is considered excellent if they are less than 0.05.

RQ1 addressed how language mindsets are related to self-efficacy, engagement, and PP. Path analysis was performed to examine the paths of the hypothesized model (see Fig. 1). As seen in Table 3, the result shows that the model fits the observed data well at an excellent level. The path model with standardized estimates is demonstrated in Fig. 2. Statistically insignificant pathways are presented with dotted lines.

Table 3 The model fit indices.
Full size table
Fig. 2
figure 2

The path model with the causal relationships between language mindsets, self-efficacy, engagement, and perceived proficiency.

Full size image

A growth mindset significantly predicted self-efficacy (γ = 0.56, t = 9.43***), while a fixed mindset did not (see the direct effects in Table 4). This indicates that a total of 30% of the variance in self-efficacy was explained by a growth mindset. For engagement, both growth and fixed mindsets predicted engagement to a similar degree (γ = 0.15, t = 2.15*; γ = 0.15, t = 2.55*, respectively), and self-efficacy predicted engagement (ß = 0.46, t = 7.15***). A total of 28% of the variance in engagement was accounted for by a growth mindset, a fixed mindset, and self-efficacy. Concerning PP, only self-efficacy (ß = 0.40, t = 5.63***) and engagement (ß = 0.15, t = 2.42*) were found to significantly predict PP. A total of 26% of the variance in PP was accounted for by self-efficacy and engagement. Notably, a growth mindset was not a significant predictor of proficiency. Taken together, these results show that while a growth mindset strongly predicted self-efficacy, it was not a significant predictor of PP. Additionally, self-efficacy had relatively strong causal relationships with engagement and PP.

Table 4 Causal direct effects between the variables.
Full size table

RQ2 focused on the indirect effects in the relationships between language mindsets, self-efficacy, engagement, and proficiency. As shown in Table 5, the relationship between self-efficacy and PP was significantly mediated by engagement (ß = 0.07, t = 2.3*). It is worth noting that when examining the impact of a growth mindset on this relationship, the link between a growth mindset and PP was mediated by self-efficacy and engagement to a much greater degree (ß = 0.29, t = 6.01***). This highlights the importance of taking into account a growth mindset when examining the connection between self-efficacy, engagement, and proficiency. The LISREL output did not offer information about the relationship between a growth mindset, self-efficacy, and proficiency. Thus, the indirect effect of this relationship was calculated by the Sobel test.Footnote 1 The result suggests that self-efficacy significantly and fully mediated the relationship between a growth mindset and proficiency (z = 0.501, p < 0.001). On the other hand, it was found that the relationship between a fixed mindset and proficiency was not mediated by self-efficacy and engagement (see Table 5).

Table 5 Indirect effects of the path diagram.
Full size table

Discussion

This study aimed to investigate how language mindsets, self-efficacy, engagement, and PP are related to each other. In order to fulfill the research objectives, a model incorporating these variables was proposed (Fig. 1) and examined through path analysis as shown in Fig. 2. Based on the results identified, the two major findings that emerged are as follows.

The roles of a growth mindset in initiating L2 learning

This study supports the findings of previous studies that indicate the crucial role of a domain-general growth mindset in language learning. In line with Zarrinabadi et al. (2022) and Jourden et al. (1991), a growth mindset predicted self-efficacy with the highest direct impact (γ = 0.56, t = 9.43***) among those observed variables. This suggests that the more individuals view L2 learning through the lens of a growth mindset, the more they are likely to perceive that they are capable of performing tasks and reaching the goals they have set. Additionally, a growth mindset was found to significantly influence the degrees of engagement, which can be considered as a motivational outcome that informs individuals’ effort and dedication toward language learning (Hiver et al., 2024). Interestingly, however, a growth mindset did not significantly influence PP. This is discussed further in the next section (see “Self-efficacy and its mediating roles”).

Concerning a fixed mindset, the results indicate that it did not predict self-efficacy. Individuals who possess a fixed mindset tend to perceive themselves as having a predetermined capacity for learning L2. As a result, they are less likely to encounter situations that offer them the opportunities to have adequate successful experiences or positive feedback which are resources to enhance self-efficacy in L2 learning. On the other hand, a fixed mindset was a significant predictor of engagement, almost with identical strength to a growth mindset (γ = 0.15, t = 2.15*; γ = 0.15, t = 2.55*, respectively). These results should be interpreted with special care. For example, several studies (e.g., Burnette et al., 2013; Lou and Noels, 2017) inform us that individuals with a growth mindset tend to engage in L2 learning to develop their own abilities and skills, whereas those with a fixed mindset may be motivated by external factors such as proving themselves to others. Lou and Noels’ (2017) L2 learning model of the mindsets-goals-responses supports this claim. Their findings indicate that individuals exhibit varying goals and responses based on their language mindsets when engaging in challenging tasks. These differences can potentially lead to diverse outcomes, including differences in individuals’ engagement or persistence over a longer period. Therefore, when interpreting the effect of language mindsets on engagement, it is important to interpret them with caution due to the cross-sectional nature of the study.

In the present study, the indirect effect of self-efficacy on PP through engagement was found to be significant (ß = 0.07, t = 2.3*). Engagement partially mediated the link between self-efficacy and proficiency. However, when a growth mindset was incorporated into this link as a variable that initiates L2 learning, a growth mindset was found to significantly affect PP through the mediation of self-efficacy and engagement to quite a larger degree (ß = 0.29, t = 6.01***) as compared to the prior link (see Table 5). This finding brings our attention to the pivotal role of a growth mindset in L2 learning. Incorporating a growth mindset as the initial stage is crucial to boost the effectiveness of English learning. Lou et al. (2021, p. 6) view language mindsets as “deeply-seated ontological and epistemological beliefs” about human beings as having an innate language ability, and the nature of language acquisition. The findings of this study support their claims, suggesting that language mindsets, particularly a growth mindset, are the fundamental beliefs that shape individuals’ positive learning behaviors within a language learning context. Thus, it is important to foster more sophisticated epistemological beliefs as a means to enhance L2 learning.

Self-efficacy and its mediating roles

The results of the direct effects highlight the critical roles of self-efficacy in L2 learning. Self-efficacy was found as a significant and strong predictor of engagement and PP in this study. Its impact on engagement was much greater (ß = 0.46, t = 7.15***) than that of the growth mindset (γ = 0.15, t = 2.15*). It also should be noted that self-efficacy significantly and directly predicted students’ PP (ß = 0.40, t = 5.63***). Interestingly, however, a growth mindset did not directly serve as a significant variable to affect students’ self-PP. This contrasts with findings from Lou and Noels (2017, 2019) and Mercer and Ryan (2010), who reported a significant relationship between language mindsets and PP. Several studies have also documented the positive impact of a growth mindset on actual L2 achievement (Khajavy et al., 2021, 2022; Fathi et al., 2024; Shirvan et al., 2024). It needs to be reminded that the PP employed in this study differs from actual L2 achievements. Specifically, an individual’s subjective judgment of their proficiency is more closely related to their intrinsic motivation than to L2 achievement (Tanaka, 2023) and is likely influenced by their experiences, comparisons with peers, feedback from teachers, or personal expectations. Given this, it appears that a growth mindset alone cannot account for the formation of PP due to its multifaceted nature. This distinction is important to acknowledge, as it aligns with the complexity of PP and its dependence on various contextual and personal factors beyond the influence of a growth mindset alone.

Accordingly, the absence of a direct and significant link between a growth mindset and PP in the study highlights the need to consider a complex interplay of factors such as self-efficacy and engagement. Of particular importance, this study draws special attention to the mediating role of self-efficacy in the relationship between a growth mindset and academic outcomes. As the results of the study show, self-efficacy partially mediated the link between a growth mindset and engagement. In addition, while a growth mindset did not directly affect PP, the finding of this study revealed that self-efficacy served as a full mediator in the relationship between a growth mindset and PP (see Fig. 2). Its indirect effect was found to be powerful (z = 0.501, p < 0.001). This indicates that the presence of a growth mindset does not directly enhance an individual’s self-assessment of their L2 proficiency; rather, its impact is channeled through self-efficacy. This suggests that self-efficacy appears to play an even more essential role in mediating the influence of a growth mindset on students’ subjective evaluation of their language skills. In other words, the more a growth mindset individuals have toward L2 learning, the more they tend to believe they are capable of accomplishing tasks in challenging situations, and the more likely they “go through the drudgery of perfecting skills that enable them to perform at extraordinary levels.” (Bandura, 1997, p. 119). Such active engagement in tasks likely leads individuals to perceive their proficiency more positively.

To sum up, the findings of this study highlight the crucial role of self-efficacy as a mediating and facilitating variable in language learning as a self-regulatory variable that motivates and facilitates language learning (Bai and Wang, 2023). A sense of self-efficacy plays the role of an indicator of motivation, behaviors, and future decision-making (Usher and Pajares, 2009). This study underscores the importance of considering an individual’s self-efficacy beliefs in conjunction with a growth mindset to result in better academic and motivational outcomes in L2 learning.

Conclusion

The purpose of this study was to investigate the relationships between language mindsets, self-efficacy, engagement, and PP. This study found that, first, students need to engage in L2 learning that is grounded in a growth mindset because it is likely to make the L2 learning process more effective. This is especially true when considering self-efficacy, engagement, and PP. Secondly, self-efficacy serves as a powerful mediator, linking the impact of a growth mindset to L2 learning outcomes such as engagement and PP.

Although the major findings in this study make a valuable contribution to the existing literature related to L2 learning mechanisms, it is important to note some limitations. First, the findings of this study cannot be generalized. The number of participants who joined this study was 257 from universities in three regions. As such, the sample cannot be seen to adequately represent Korean university students, and caution is required when understanding and applying the findings to a specific educational field. Second, it was observed that a growth mindset did not emerge as a predictor of PP, underscoring the distinction between participants’ self-reported PP and actual L2 achievement. This difference may be attributed to various factors that should be considered, as mentioned earlier. It is worth noting that the measurement of PP in this study primarily relied on quantitative data. While offering valuable insights, integrating qualitative data, such as interviews or retrospective reports, could enrich our understanding of PP. For future research, incorporating a multifaceted approach could provide a richer, more comprehensive understanding of PP. Third, this study focused on the relationship between domain-general language mindsets and important language learning variables. In future studies, exploring how a domain-specific growth mindset affects PP in particular skill areas would provide valuable insights into the distinct impact of mindsets on language learning. Finally, it would be valuable to explore how interventions aimed at fostering a growth mindset affect student engagement in L2 learning over an extended period.

The findings revealed in this study offer crucial implications for L2 learning classrooms. It is vital to design and implement an intervention program along with procedural strategies tailored to a particular L2 context to nurture students’ mindsets toward growth. In addition, teachers should pay particular attention to organizing their courses for students to have sufficient opportunities to enhance their self-efficacy beliefs. Most importantly, teachers and students need to understand the considerable impacts of a growth mindset and self-efficacy on the effectiveness of L2 learning, especially when considered in combination.