Introduction

Growth mindset (GM) framework theorises the perception that one’s ability can be developed with effort (Dweck, 1999). As one of the socio-cognitive beliefs (also motivational factors) that promotes learning results, GM has garnered increasing attention over the past two decades (e.g. Boncquet et al., 2023; Chen & Wong, 2015; Lou & Noels, 2017; Yeager & Dweck, 2012). Regarding domain specificity, GM is conductive to writing performance and development as well (Bai et al., 2020). Although GM is supposed to facilitate learning outcome, it generally exerts a rather weak direct effect (r = 0.10) in learning according to Sisk et al.’s (2018) meta-analysis. Alternatively, GM has an indirect positive effect on learning outcome via the mediation of many other non-cognitive factors such as learning motivation (Cheong et al., 2023; Bedford, 2017; Ng, 2018; Zhao et al., 2018), and the utilization of strategy use (SU) during the learning process (Bai & Guo, 2018; Bai & Wang, 2021; Doron et al., 2009; Law, 2009), which in turn contributes to learning results. Particularly in Chinese learning contexts, where hard-working is traditionally nurtured, GM is considered an essential and important mentality for learning outcomes (Bai et al., 2021).

Given that GM is related to motivational and strategic factors, and writing is one of the most challenging domains in language learning, we are interested in exploring the extent to which GM synergises with learning motivation and SU—two effective facilitators for learning, in a writing scenario. Above all, motivation within the framework of self-determination theory (hereafter ‘SDT’) in learning contexts conceptualises an active autonomous orientation to learn (Ryan & Deci, 2017). Typically classified as autonomous and controlled, motivation has gained increasing attention in contemporary educational and psychological research over the past decades (e.g. Ryan & Connell, 1989; Ryan & Deci, 2017, 2020). Both types of motivational beliefs contribute to self-regulated SU in language learning (Bai & Wang, 2021), as they do in writing (Bai & Guo, 2018; Bai et al., 2021). Second, SU, as a major processing index, also plays a significant role in writing performance. Particularly in highly competitive learning and examination contexts such as Hong Kong, SU tends to be integral to learning since it can effectively serve examination purposes (Bai & Wang, 2021). Hence, the importance of SU manifests itself clearly in writing—a multifaceted language production task.

While there are a variety of writing genres designed with different purposes, one of the challenging writing tasks is integrated writing (IW). IW not only entails writers to transfer their own thinking into written verbal form (Flower & Hayes, 1984), but also necessitates writers to integrate and evaluate multiple sources of information (Cumming et al., 2016). Mastering writing strategy is particularly beneficial for improving the performance of young writers, as it enables them to develop essential self-regulation skills crucial for enhancing their future writing achievements.

There is an urgent need to examine a joint examination of SDT-based motivation (Ryan & Deci, 2020) and Dweck’s mindset (Dweck & Leggett, 1988) as these two complement each other in learning achievement (Boncquet et al., 2023). In learning scenarios, the former centres on learners’ motivational foundation (i.e. the source of learning) while the latter places more emphasis on learners’ social-cognitive layer (i.e. individual beliefs about learning). Accordingly, the primary impetus of this present study is to disentangle the intricate relationships between these two different layers of motivational beliefs (GM and motivation) and SU in an L1 Chinese learning context—Hong Kong. We aim to propose a more comprehensive motivational framework that accounts for Hong Kong secondary students’ learning psychology situated in an IW task, which as mentioned earlier, requires high self-regulation.

Literature review

GM and its role in learning

Dweck’s GM framework posits that one’s intelligence can be improved through effort from a social-cognitive perspective (Dweck & Leggett, 1988). It is often used interchangeably with the ‘incremental theory of intelligence’ in literature reviews. Like other positive learning beliefs, GM is expected to enhance the learning experience and be conducive to learning (Chen & Wong, 2015). Research on GM is important for educators and practitioners to understand students’ learning behaviours. Findings suggest that individuals with a growth mindset tend to boost their academic performance more rapidly and with less effort compared to those lacking a growth mindset (e.g. Bai & Wang, 2021; Claro et al., 2016; Chen & Wong, 2015; Lou & Noels, 2017; Waller & Papi, 2017; Yeager & Dweck, 2012). This is because GM fosters greater academic engagement, leading to more enjoyable learning experiences (Lou & Noels, 2016; Yeager & Dweck, 2012), and cultivates a persistent and resilient mentality towards learning. For instance, Chinese university students with a higher level of L2 writing GM maintained greater learning motivation over a 15-week longitudinal study (Yao et al., 2021).

However, emerging evidence has emphasized GM’s indirect effect on learning. As summarized in Sisk et al.’s (2018) meta-analysis, the average positive correlation between GM and academic achievement worldwide was unexpectedly weak (r = 0.10, k = 273, N = 365,915, screened studies from 1989 to 2016). This suggests that GM tends to have a limited direct effect on learning outcomes in general. Alternatively, GM often plays an indirect facilitative role in learning through the enhancement of motivational beliefs. Specifically, GM enhances autonomous learning behaviours by promoting positive learning experiences, including goal orientation (Chen & Wong, 2015; Lou & Noels, 2016, 2017), grit (Park et al., 2020; Zhang et al., 2022), enjoyment (Cheong et al., 2023), and motivation (Boncquet et al., 2023; Han & Hiver, 2018), thereby effectively promoting learning achievement.

The positive indirect effect of GM applies to various socio-cultural learning contexts around the world, such as Chile (Claro et al., 2016), Canada (Lou & Noels, 2017), Iran (Khajavy et al., 2021), Singapore (Ng, 2018), mainland China (Cheong et al., 2023) and Hong Kong (Chen & Wong, 2015). Significantly, GM appears to be particularly important in socio-cultural contexts that highly value hard work, such as in Chinese culture (Bai et al., 2021; Cheong et al., 2023). For example, GM contributed to Chinese students’ English learning in Hong Kong. Bai and Guo (2018) reported that higher achievers endorse a higher level of GM, which leads to self-regulated learning strategy (β = 0.29) and self-efficacy (β = 0.44). Similarly, in Bai and Wang (2021), GM significantly predicted monitoring (β = 0.59), effort regulation (β = 0.54), and goal setting and planning (β = 0.41), respectively.

In terms of domain specificity, GM has been identified as a facilitator of writing outcomes, both directly and indirectly. Regarding its direct effect, Xu and Wang (2022) observed that GM was associated with increased strategy use in English L2 writing among Chinese undergraduate students. Similarly, a recent intervention study demonstrated that bolstering GM significantly improved the writing performance of Portuguese sixth graders, including enhancing text quality and length (Camacho et al., 2023). An example of GM’s indirect effect is illustrated in the findings of Cheong et al. (2023) among Chinese high school students. They identified two pathways between GM and emotions. Firstly, GM was found to mitigate learning anxiety among students at a moderate level (β =  − 0.36) and therefore promoted writing performance. Secondly, GM was associated with increased enjoyment of learning among students at lower levels (β = 0.38), which also benefited writing outcomes. These studies together underscored GM’s positive impact on learners’ writing performance.

Taken together, previous studies collectively demonstrate that the concept of GM (i.e. improving oneself with effort) is deeply rooted in Chinese societal and learning contexts, and GM can foster positive learning psychology (Bai & Guo, 2018; Bai & Wang, 2021) and contribute to learning (Yao et al., 2021). However, less attention has been given to the combined effect of GM and motivations under SDT. As argued by Boncquet et al. (2023), it is crucially meaningful to research these two well-established motivational theories together, as they often interact with each other and reflect different aspects of motivational beliefs, shedding insights into the learning process and achievement.

Motivation and their roles in writing learning

Motivation signifies the inclination to work towards a specific goal or outcome, and in the realm of education, it is widely regarded as a pivotal and direct factor influencing learning behaviours and outcomes. SDT stands out as one of the most influential motivation theories in contemporary research (Vasconcellos et al., 2020). SDT approaches learning motivation from a perspective grounded in humanistic psychology. Unlike previous frameworks that simply categorized motivation into intrinsic and extrinsic, SDT constructs a comprehensive framework that encompasses the factors underlying motivation and behaviours (Ryan & Deci, 2017). Within the SDT framework, motivations are classified into autonomous motivation (AM) and controlled motivation (CM). AM entails engaging in an activity based on the belief that it aligns with personal goals or desired outcomes, originating from internal determination (Ryan & Deci, 2020). Conversely, CM involves participating in activities driven by external incentives, such as seeking rewards, approval from others or avoiding punishment or guilt (Ryan & Deci, 2020).

The conceptualization of SDT is central to Ryan and Connell (1989)’s motivation theory. Ryan and Connell (1989) were among the first to validate an academic motivation questionnaire, where motivation is delineated into four nuanced types: external (compliance with rules and avoidance of punishment), introjection (seeking approval and avoiding disapproval), identification (valued goals and personal importance) and intrinsic (enjoyment and fun). According to SDT, the former two fall within the realm of AM, while the latter two are categorized under CM. Recent reviews indicate that both types of motivation have been found to enhance language learning performance (e.g. Bai & Guo, 2018; De Smedt et al., 2020; Guay et al., 2010; Rasteiro & Limpo, 2023; Ryan & Deci, 2000; Wang & Guthrie, 2004). Nonetheless, it should be pointed out that unlike the relationship between intrinsic and extrinsic motivations, AM and CM are not polar opposites; rather, they often exhibit a moderately positive relationship (Wigfield & Guthrie, 1997).

AM and CM play an important but distinctive roles in writing attainment. To illustrate, both AM and CM contributed to French-Canadian bilingual grade 1–3 students’ writing performance (Guay et al., 2010). De Smedt et al. (2020) further noted two insightful observations. First is that students’ AM deceased significantly from middle primary to lower secondary grade, while there was no such change with CM. This implied the necessity of fostering AM particularly during higher school years. Second, they also reported that students with a higher level of AM in writing are more likely to endorse a higher level of AM in reading as well. These together demonstrated the importance of nurturing AM in writing as well as literacy learning.

Meanwhile, the relationship between the two types of motivation and writing performance vary when educational contexts are considered. For instance, some evidence suggested that there was no correlation between motivational factors and writing strategy use among American primary school students (Graham et al., 2017). Much to the contrary, AM was found to be significantly correlated with writing strategy use among Portuguese (Rasteiro & Limpo, 2023), French-Canadian (Guay et al., 2010) and Chinese (Bai & Guo, 2018; Bai et al., 2021; Law, 2009) primary school students. Take Chinese as an example. It was found that Chinese primary school students with higher levels of AM towards writing tend to produce more satisfactory writing performance and exhibit greater advancement in writing strategy use (Bai & Guo, 2018; Bai et al., 2021). This is because writers with AM are intrinsically motivated to invest effort in learning and employ more effective strategies accordingly.

IW and its SU

Writing is a multifaceted cognitive task that draws on various dimensions, including cognitive, metacognitive and affective aspects. At its core, writing involves problem-solving through the organized representation of ideas in written form (Flower & Hayes, 1984). IW, commonly encountered in Hong Kong secondary schools, is particularly demanding compared to independent writing. This is because IW requires additional skills such as digesting and evaluating source materials, while generating ideas and forming argument simultaneously (Cumming et al., 1989; Graham et al., 2017; Spivey & King, 1989).

In IW tasks, SU plays a pivotal role that facilities the writing performance. Spivey and King (1989) specifically proposed ‘discourse synthesis’ as an IW strategy model focusing on the integration of the source information and transforming into writing, which is characterized by three main operations: organization, selection and connection. However, this model did not explicitly address metacognitive or regulatory processes involved in writing. Building on Spivey and King’s (1989) writing model, Yang and Plakans (2012) expanded the scope of the model to include metacognitive processes and strategic thinking in writing. They proposed a new model composed by three sub-factors: self-regulatory (e.g. planning, monitoring, and evaluating one’s witing), discourse synthesis (e.g. how writers integrate and synthesise information from multiple sources) and test-wiseness (e.g. the ability to employ effective strategies tailored for test-taking situation). This model has served as a foundation for subsequent research on IW tasks (e.g. Zhu et al., 2021a, 2021b; Plakans & Gebril, 2013; Shi et al., 2020; Zhao & Liao, 2021).

Effective SU can enhance writing skills and performance, thereby contributing to successful writing activities (Cheong et al., 2019; Bai & Guo, 2018; Bai & Wang, 2021; Bai et al., 2021; Graham et al., 2017; Pitenoee et al., 2017; Smith et al., 2007). The importance of SU is particularly evident in the quality of writing among young writers (Cheong et al., 2019; Zhu et al., 2021a; Bai & Guo, 2018; Graham et al., 2017), underscoring its unique role in youth compared to adult writing. For instance, Cheong et al. (2019) found that discourse synthesis skills significantly predicted the IW performance of Hong Kong secondary four students in both their Chinese L1 (β = 0.20, p < 0.001) and L2 (β = 0.12, p < 0.01). Similar findings were later corroborated by Zhu et al. (2021b), demonstrating the importance of discourse synthesis through both traditional assessment results and eye-tracking patterns.

The relationships between GM, motivations and SU

The concept of GM stands as a relatively stable and enduring psychological factor that exerts influence on an individual’s overall learning outcomes. However, as previously discussed, the optimization of learning outcomes often occurs when GM collaborates with other contributing factors. Accumulated evidence highlighting the indirect effects of GM suggests its alignment with motivation, which, in turn, fosters learning achievement (Bedford, 2017; Dweck, 2007). Individuals who embrace a GM, believing in the power of effort to yield positive results, are more inclined to engage in SU in their writing endeavours. Furthermore, both intrinsic and extrinsic motivations, driven by the desire and interest in writing, can also stimulate the utilization of strategies.

Existing research indicates a complex interplay among GM, motivation and SU in the learning patterns of young learners. For instance, GM and intrinsic value (akin to AM) have been found to facilitate self-regulated learning and SU (Bai & Wang, 2021). Specifically within writing contexts, findings from a study involving Hong Kong primary four students revealed a reciprocal correlation between GM and motivation (self-efficacy and interest), with both factors positively predicting SU (self-monitoring and acting on feedback) (Bai & Guo, 2018). This underscores the significance of considering GM and motivation in the learning process, yet there remains a gap in understanding how the dynamics of these relationships contribute to the ultimate learning outcomes.

Research gaps and hypotheses

In sum, existing literature reviews underscored the pivotal roles of GM, motivation and SU in promoting learning outcomes, respectively. However, there are notable gaps and areas for further exploration. First, while motivational foundation has received considerable attention, its concurrent GM’s social-cognitive dimension remains relatively underexplored. This gap motivates a deeper investigation into the unique role of GM in language learning contexts. Second, although motivational factors (i.e. GM and motivation) have been studied in writing, most were disproportionately represented by younger writers such as primary school students (Bai & Guo, 2018; Bai & Wang, 2021; Camacho et al., 2023), adult writers such as undergraduate (Zhu et al., 2024) or L2 contexts (Bai et al., 2021; Xu & Wang, 2022; Zhao et al., 2018). Hence, the inconclusive findings insofar over the effects of motivation and CM further motivate this study to probe the relationship of both motivations and their interplay with GM and SU among secondary school students’ L1 writing.

Moreover, existing empirical studies have placed primary attention to SU in the scenario of L2 writing contexts (e.g. Bai & Guo, 2021; Bai et al., 2021; Golparvar & Khafi, 2021), lacking firm insights into L1 writing. Although some studies have examined writing strategies in pre-adolescent writers, i.e. grade-four (Bai & Guo, 2018; Bai & Wang, 2021; Zhao et al., 2018) or adult writers, i.e. university students, while the role of writing SU among adolescent writers remained relatively unanswered (except for Cheong et al., 2019; Smith et al., 2007). Given the importance of SU, particularly in IW tasks that require synthesizing multiple materials (Cheong et al., 2019), this study aims to examine the role of GM in both types of motivation and SU among Hong Kong Chinese secondary school students engaged in IW tasks. Building on the existing literature, our hypotheses are as follows:

  • Hypothesis 1: GM promotes IW performance via the mediation of both motivations.

  • Hypothesis 2: GM promotes IW performance via the mediation of SU.

  • Hypothesis 3: GM promotes IW performance first via the mediation of AM and then the mediation of SU.

  • Hypothesis 4: GM promotes IW performance first via the mediation of CM and then the mediation of SU.

Below is the conceptual framework of this study (Fig. 1).

Fig. 1
figure 1

The conceptual framework of this present study. GM growth mindset, AM autonomous motivation, CM controlled motivation, SU strategy use, IW integrated writing

Full size image

Methods

Sampling

A total of 465 secondary four students (Mage = 15.4, SD = 1.30, 65% male) from six public schools in Hong Kong were recruited in a larger study that investigated on reading-writing connection. All students were Chinese native speakers (Cantonese or Mandarin), but they can all speak Cantonese fluently as Cantonese is the medium of instruction in their Chinese literacy class. The schools are distributed across three major territories (Hong Kong Island, Kowloon and the New Territories). To reflect variance of academic representation, the sampling covers local bandings of top, middle, and lower levels, each representing two schools. Additionally, we sought a combination of background variables like gender and socio-economic backgrounds. Research ethical application was reviewed and approved by a prestigious University in Hong Kong. Consents from students and their parents were obtained before data collection.

Measurements

IW

The IW task was adopted from Cheong et al. (2019) that requires students to write an argumentative essay within 550 Chinese characters. Students need to express their opinions based on two sources on the topic ‘Whether the courtesy seats on metro are only reserved for the elderly?’ The first source is a journal review showing different perspectives, which contextualises the arguments over the use of courtesy seats. The second source expresses netizens’ contrasting viewpoints on the controversial issues related to this topic. After reading the two sources, students were instructed to write a response letter to the corresponding journal in the capacity of the council president of the courtesy seats committee. The metro topic is appropriate for secondary four students because they are familiar with it.

We adapted Zhu’s (2005) writing rubrics for the grading of the IW task because it has been applied to the Chinese language examination of the current HKDSE (Hong Kong Examinations and Assessment Authority, 2018) and was used in previous published studies such as Zhu et al. (2021a). To ensure content validity, we also consulted three Chinese writing experts for the appropriateness of using the rubrics, and they all confirmed so. The rubrics assessed writing from four-dimension: contextual awareness, citation and synthesis, opinion and argument, written expression and organization. Each dimension accounted 10 points, divided into five bands, with two points being an interval. So the overall IW score accumulated at 40 points.

Questionnaires

The originally English questionnaires (GM, motivation and SU) were first translated into Chinese, and then back-translated into English. The original English and back-translated versions were compared. The instructions and all the items were scrutinized and revised for accuracy, clarity and readability to fit into Chinese contexts. To keep consistent in this study, we transformed the following questionnaires’ response scale into a six-point one, ranging from 1 (strongly disagree) to 6 (strongly agree).

GM

The GM questionnaire was domain-general, adapted from Dweck (2007), and the original instrument was assessed on a five-point Likert-scale. The Cronbach’s α of GM questionnaire in this study is 0.75, comprised of eight items (e.g. ‘You can always substantially change how intelligent you are’).

Motivation

We employed the Academic Self-Regulation Questionnaire (SRQ-A) developed by Ryan and Connell (1989). We used SRQ-A for two major reasons: (1) SRQ-A is one of the most influential motivation scale that captures students’ motivational beliefs towards learning; (2) the four dimensions of SRQ-A taps into the differentiation of AM and CM in our study (Boncquet et al., 2023). The original SRQ-A concerns the reasons why children do their school work. In this study, we adapted the questions as domain-specific, into asking students why they want to learn Chinese well. There were 12 items in this study, dividing into intrinsic (six items, Cronbach’s α = 0.84, e.g. ‘Because it’s fun’) and CM (i.e. external regulation in the original SRQ-A, six items, Cronbach’s α = 0.91, e.g. ‘Because I want the teacher to say nice things about me’). Items associated in this study were anchored to overall Chinese language learning.

SU

The 23-items strategy inventory was originally developed by Yang and Plakans (2012), which is domain-specifically about writing. Our confirmation factor analysis result showed a four-dimension structure, that is, planning (e.g. ‘I thought about the type of essay I wanted to write’), self-regulatory (e.g. ‘I double-checked to see if my writing met the task requirements’), discourse synthesis (e.g. ‘I searched for connections among sentences’) and test-wiseness strategies (e.g. ‘I planned to copy good sentences from the reading in my writing’). All 23 items were included, and the Cronbach’s α of the full SU scale in this study was 0.94.

Procedure

Data collection

For the data collection of this present study, the procedure was administered in the same school day. Students first completed the IW task (45 min) along with the questionnaires (20 min). All tasks were in paper-based form.

Grading procedure

The IW tasks were scored by two raters independently. Both raters were native speakers of Chinese who were employed as full-time research assistants for this project. At the time, one has completed her PhD study in Chinese education and the other was a PhD student in education. Besides, experienced Chinese language academics were invited to convene with the authors and the raters for standardization and moderation processes.

Before formal grading procedure, the two raters received an individual 2-h training session, and they both practiced marking five random samples independently. The average of the two independent ratings determined the final mark on each writing. Significant discrepancies (more than three marks) were re-graded through discussion with a third rater (the corresponding author). The inter-rater reliability for the four dimensions ranged between 0.92 and 0.95, and the overall Cronbach’s α was 0.98.

Statistical analysis

First, confirmatory factor analysis (CFA) were conducted to validate the structure of SU and IW performance, which were latent variables in this study. Then, structural equation modelling (SEM) analysis was conducted to explore the relationship between GM, AM, CM, SU and IW performance. Furthermore, bootstrapping method was performed to test the significance of the mediating effects.

Data were analysed using Mplus 8.3 (Muthén & Muthén, 1998–2017), with a robust maximum likelihood estimator. Furthermore, given the nested structure of the data in this study, we included the ‘type = complex’ command in Mplus to adjust the standard errors for school clustering. To assess the quality of the model-data fit, the following criteria were used: comparative fit index (CFI; good > 0.95), Tucker-Lewis index (TLI; good > 0.95), root-mean-square error of approximation (RMSEA; good < 0.05) and standardized root-mean-square residual (SRMR; good < 0.05) (Hancock & Mueller, 2013).

Results

Descriptive statistics and correlations

Descriptive statistics were presented in Table 1. In accordance with Kline’s (2023) criteria, all variables’ absolute values of Skewness and Kurtosis were below 3, indicating normal distribution.

Table 1 Descriptive statistics for the four measures (N = 465)
Full size table

The correlation between the observed variables is shown in Table 2. GM was positively correlated with AM (r = 0.16, p < 0.01) and SU (r = 0.13–0.18, p < 0.01), but not with CM and IW. AM is positively correlated with all dimensions of SU (r = 0.24–0.35, p < 0.01) and the three dimensions of IW (r = 0.12–0.14, p < 0.01). Whereas CM was positively correlated with all four dimensions of SU (r = 0.21–0.38, p < 0.01), and with only one dimension of IW (r = 0.12, p < 0.05). Moreover, the contextual awareness, citation and synthesis and opinion and argument in IW showed significant correlations with all dimensions of SU (r = 0.13–0.27, p < 0.01). Hence, the correlation matrix suggested the possibility of exploring the relationships between GM, AM, CM, SU and IW performance.

Table 2 Bivariate correlations of the study variables
Full size table

The relationship between GM, AM, CM, SU and IW performance

We examined the measurement of SU and IW performance (used as latent variables in this study) using CFA. The two measurement models fit the data well (Table 3), and all factor loadings were larger than 0.40 (see Appendix Fig. 3 for details). Overall, the CFA results indicated high measurement validity of SU and IW task have.

Table 3 Fit indices of the measurement model and structural model
Full size table

To test the hypothesis, the full SEM model (see Fig. 2) was conducted to examine the relationship between GM, motivation, SU and IW performance. The model showed a good model fit (χ2 = 96.76, df = 37, p < 0.001, CFI = 0.98, TLI = 0.98, RMSEA = 0.06, SRMR = 0.04) The results showed that GM positively affected AM (β = 0.16, p < 0.01), but not CM. Moreover, although without direct effect on IW performance, GM positively impacted SU (β = 0.15, p < 0.05). Furthermore, both AM (β = 0.27, p < 0.001) and CM (β = 0.29, p < 0.001) had positive effect on SU, but did not affect IW performance. Finally, as expected, there was a significant positive correlation between SU and IW performance (β = 0.26, p < 0.001).

Fig. 2
figure 2

Standardisation coefficients of conceptual model. GM growth mindset, AM autonomous motivation, CM controlled motivation, SU strategy use, IW integrated writing. *p < 0.05; **p < 0.01; ***p < 0.001

Full size image

Mediation analysis

We further ran a mediation analysis to examine SU’s mediating effects on the relationships between GM motivation and IW performance. Our results showed that motivation or GM indirectly affect IW performance through four different pathways (Table 4; with three simple mediations—Path A, B, C and one chain multiple mediation—Path D).

Table 4 The potential pathways by which GM and motivation affect IW performance
Full size table

To summarise, hypothesis 1 was not supported—GM did not promote IW performance via either motivation. Instead, we noticed that AM (Path A: β = 0.07, 95% CI = [0.04, 0.10], p < 0.001) and CM (Path B: β = 0.08, 95% CI = [0.03, 0.13], p < 0.001) indirectly affected IW performance via SU, respectively. Hypothesis 2 was fully supported as evidenced in Path C—GM promoted IW performance via the mediation of SU (β = 0.04, 95% CI = [0.01, 0.07], p < 0.001). Hypothesis 3 was also fully supported while hypothesis 4 was not—GM only benefited IW performance via the mediation of AM and then the mediation of SU (Path D: β = 0.01, 95% CI = [0.003, 0.02], p < 0.001).

Discussion

The major purpose of this study is to explore the conjunct effect of two motivational theories (GM framework and SDT) and SU in writing performance. To do so, we examined the relationship between GM, motivations (AM, CM) and SU in writing among secondary four students in Hong Kong. We underscored three major findings. First, GM was a significant positive predictor of AM, SU, but not CM. Second, GM indirectly promoted IW through the medication of SU and AM. Third, neither AM nor CM exerted direct effect on IW performance, whereas both did so via the mediation of SU.

The role of GM on motivations, SU and IW performance

In terms of direct effect, interestingly, GM positively predicted AM, but not CM. Most existing studies researched GM and AM consistently showed a positive relation between the two (e.g. Aronson et al., 2002; Bai & Guo, 2018; Bai & Wang, 2021; Boncquet et al., 2023; Zhao et al., 2018), and this is not unexpected. Our present study expands on the current literatures by noticing that GM did not predict CM, which delineated a more nuanced picture of the relationship between GM and motivational domains. The fact that GM only promotes AM but not CM further evidenced that CM (vs. AM) is usually less concerned with belief about self-improvement (Ryan & Connell, 1989). This also resonates with Boncquet et al.’s (2023) research with seventh-grade students that GM and AM lead to positive learning psychology and outcome. Hence, theoretically, our findings adduce evidence to Boncquet et al.’s (2023) argument that the two motivational resources—GM (from a social-cognitive belief) and AM (under SDT, emphasizing the source of learning)—are intrinsically interconnected and together affected learning.

Moreover, congruent with many previous studies, GM predicted SU in writing contexts (Bai & Guo, 2018; Bai & Wang, 2021; Xu & Wang, 2022) and other learning domains such as reading (Law, 2009). These together implied that GM can effectively maneuver learning SU, whereby contributing to writing performance and achievement. We would like to elaborate on the assumption that GM and SU probably have two different routes. The former one pertains to a typical psychological layer related to affect and value (Dweck, 1999), while the latter one is tied to task processing layer (Yang & Plakans, 2012). Accordingly, bringing awareness of GM and activation of SU concurrently is most desirable for learning. Apart from that, referring back to the cognitive challenges of IW, GM holders are more likely to positively perceive their learning of writing over time (Bai & Wang, 2021), so they anticipate the developmental process of writing as a long-term, gradually improving stage.

Meanwhile, our findings converged with extant studies highlighting an indirect GM effect on learning performance. Namely, GM affected learning performance through the mediation of many non-cognitive factors (e.g. Cheong et al., 2023; Chen & Wong, 2015; Lou & Noels, 2016, 2017; Park et al., 2020; Zhang et al., 2022). Aligned with previous study, we found that GM affected IW via the mediation of AM (e.g. Han & Hiver, 2018) or SU (Bai & Wang, 2021) but not via CM. These findings unanimously stressed the importance of bearing GM in learning, since it leads to intrinsic motivation and SU on ultimate writing performance. Furthermore, we also reported a long-trajectory indirect effect from GM—> AM—> SU—> IW. This pathway is particularly meaningful in terms of teaching implications (will soon be discussed in ‘Limitations, contribution, and implications’ section). Finally, we would like to underscore that the four facilitative paths all cannot sidestep SU, which indicated the unique role of SU in IW performance. Our explanation is that the IW task, owing to its complex traits (e.g. integrating information sources), probably open a unique window to solicit SU.

The role of two motivations in the context of IW

Both AM and CM facilitated IW via the mediation of SU, but neither motivation directly influenced IW performance. Above all, our findings were in line with previous research that motivation mobilizes SU (Bai & Guo, 2018; Graham et al., 2017). Second, it should be noted that simply having motivational beliefs detached from the exertion of strategy would make it difficult to succeed in language learning (Bai & Wang, 2021), particularly in demanding task like the IW in our present study. Third, consistent with previous studies, our findings stress the concurrent effect between motivation and SU (Bai et al., 2021; Guay et al., 2010; Rasteiro & Limpo, 2023). As writing is a productive and multifaceted work, it demands higher-order cognitive activities and it needs to be represented by formal, well-organized verbal expression (Cumming, et al., 1989; Flower & Hayes, 1984; Spivey & King, 1989). Collectively, our study underscored the importance of cutting down CM and bringing up AM.

Limitations, contribution, and implications

Although with the above discussion, this study is not without limitations. First, our sample size is relatively small in the SEM analysis with four variables, even though the statistical power and model fit indicated an acceptable level (Kline, 2023). Second, our study potentially lacks data triangulation, since we only obtained self-reported questionnaire and writing performance, without being able to include qualitative sources such as interview or think-aloud data. Future studies are encouraged to compensate these two potential issues.

Regarding contributions, we draw on two motivational frameworks (GM theory and SDT) and provided a more comprehensive understanding of how motivational factors influence learning. Specifically, we zoomed in on the significant role of GM along with the investigation of AM, CM, and SU in a writing scenario. Our findings corroborated that positive social-cognitive beliefs (i.e., GM) can elicit the internal source of learning motivation (i.e., AM), but not external one (i.e., CM), which again highlighted the significant role of GM in learning contexts and also suggested the homogeneity between GM and AM—the two motivational factors. In addition, compared with the wealth of research supporting GM as a vehicle to improve learning (e.g., Wanzek et al., 2021), we further added on evidence that GM as the more general psychological property, is highly associated with one’s sustainability (Lou & Noels, 2017) because it leads to positive learning attitudes such as AM (Boncquet et al., 2023; Cheong et al., 2023). These together stressed the necessity to foster GM in the writing scenario as mentioned in published studies (Cheong et al., 2023; Bai & Wang, 2021; Waller & Papi, 2017; Yao et al., 2021), which could be further inferred to wider educational contexts (Boncquet et al., 2023).

The findings of the present study spontaneously bring fresh insights into pedagogical implications as well. Given with GM’s sustainability, we encourage the proliferation of GM in every-day classroom. As Bedford’s (2017) intervention study suggests, teachers may employ a GM and learning approach, and take advantage of the Mindset Assessment Profile (developed by Petscher et al., 2021) to cultivate GM. Language teachers are encouraged to guide students through the writing process rather than writing outcome (Xu & Wang, 2022). To do so, teachers can foster students’ GM towards their writing via in-class facilitated activities and after-class reinforcement assignments. For example, letting students report their progress of learning and organizing peer review by tapping into the Mindset Assessment Profile tool (Petscher et al., 2021).

Meanwhile, GM fits well with Chinese context (Bai & Guo, 2018), while AM is found to be low in many studies conducted in Chinese educational scenarios (Lau & Chan, 2003; Law, 2009) from East Asia. Hence, we encourage educators to consider paying more attention to the promotion of AM with GM, and via SU to improve learning performance. This may require teachers to weaken CM (e.g., toning down competition among students), to arrange interesting writing materials such as selecting topics of interest to the learner groups (Cheong et al., 2023; Camacho et al., 2023), and to advance teaching approaches such as employing multimodal teaching or the inclusion of AI for assistance to activate AM. As such, students are likely to enhance AM and performance concurrently, which would gradually form a long-term positive circulation for learning as well as self-development.

As our study evidenced that SU synergizes with GM and CM, we would suggest the training of SU as part of the scaffolding in tandem with the development of learning motivation and GM. As proposed by Zhu et al. (2024), teachers may organize well-designed classroom learning activities like arranging regular peer writing feedback sessions to enhance learning interest and therefore promoting students’ writing strategy and skills. Another possibility is to implement classroom intervention for a purposeful cultivation of students’ self-regulated learning strategy as discussed in Xu and Wang (2024). Taken together, our study highlighted the necessity to foster effective GM, motivations, and strategy use in a Chinese L1 writing context. Our findings and pedagogical implications could facilitate secondary school writers in their journey of becoming resilient and confident writers.