Abstract
This study examined the impact of school leadership on teacher autonomy (TA), with teacher professional collaboration (TPC) and teacher self-efficacy (TSE) playing the mediating role. Distributed leadership (DL) and instructional leadership (IL) are the most employed leadership styles. Data collected from teachers and principals nested within schools in Vietnam TALIS 2018 were analyzed using hierarchical linear modeling to explore the relationships between the above variables (DL with TPC, TSE, TA; IL with TPC, TSE, TA; TPC with TA; TSE with TA; TPC with TSE). The findings showed that DL directly influenced TA and indirectly affected TA through TSE. However, IL had no direct impact on TA but affected it through TSE. While IL did not affect TPC, DL did. TPC had no direct impact on TA and so did not mediate the relationship of TA with DL and IL. The results also uncovered that TPC directly impacted TSE. These results suggest that DL and TSE play crucial roles in fostering an autonomous teaching environment, while IL and TPC did not effectively promote TA. However, TALIS survey iterations may want to include cultural items to help interpret research results better. Such adjustments can enhance TALIS’ relevance in diverse cultural contexts and results, providing policymakers with more contextualized insights to improve TA and school professional development.
Keywords
Introduction
According to the Organization for Economic Co-operation and Development (OECD), “Successful education systems in the 21st century will do whatever it takes to develop teachers’ ownership over professional practice” (OECD, 2019, p. 4). Hence, teacher autonomy (TA) can be viewed as “a personal sense of freedom from interference or in terms of teachers’ exercise of control over school matters” (Wilches, 2007, p. 245) and a win-win for teachers, students and the school. With classroom autonomy, teachers can choose how to design and implement instructional strategies to meet their students’ needs (Hsieh, Song, & Li, 2024; Jang et al., 2023). Additionally, TA contributes to teachers’ self-efficacy (TSE) (S. Liu, Keeley, et al., 2021), fosters a value for collective innovativeness (D. Nguyen et al., 2021), and supports professional development (Choi & Mao, 2021). Furthermore, TA benefits students by significantly promoting their academic engagement, a result that motivates and commits educators more to their roles (Q. Ma, 2021; Onrubia et al., 2022) and improving learning outcomes ( M.Wang & Hu, 2022).
In 2018, the Vietnam Ministry of Education and Training (MOET) issued Circular No. 32/2018/TT-BGDT with a new general education curriculum to be implemented in state schools from primary to upper secondary education (Luat Vietnam, 2023). This Circular also granted teachers greater autonomy to develop their own school curricula (Do et al., 2021). However, this study aims to explore reality whether teachers have voice in their teaching career because TA has been reported to decline globally (Jeong & Luschei, 2018), and such decline would have different impacts depending on educational policies, economic status, and type of school (Salokangas et al., 2020). With this interest, the study further explores different factors that may influence TA.
Previous studies showed that teachers could become professionally autonomous when undertaking a joint teaching project, such as collaborating with colleagues to plan lessons, engaging in lesson studies, and developing teaching materials (Bush & Grotjohann, 2020; Li et al., 2022; Xu, 2015) or collaborating to develop a school-based curriculum (Pan et al., 2017). Teacher professional collaboration (TPC) can prompt teachers to reflect more on their own behaviors, fostering a greater willingness to make positive changes (Pan et al., 2024; X. Yang, 2023). In addition, teachers can also collaborate with principals to share their opinions and beliefs about teaching. When collaborating, teachers can work independently through self-learning, mentoring, and professional discussions (H. N. Tran et al., 2022).
Prior research revealed that teachers perceiving high self-efficacy feel more professionally competent and are willing to take initiative when they need autonomy as they consider themselves qualified for the job (Hsieh et al., 2023; Sokmen & Kilic, 2019). TSE represents teachers’ beliefs about their teaching competence (Keller et al., 2024). Efficacious individuals tend to cope better with hindrances and can plan the necessary steps for gaining autonomy (Federici, 2013). With these views, this study examines the relationships of TPC and TSE with TA.
Instructional leadership (IL) and distributed leadership (DL) have been extensively studied within educational research (Gumus et al., 2018) and are central constructs in the Teaching and Learning International Survey (TALIS) 2018 (OECD, 2018). Given their prominence, it is essential to explore the effects of IL and DL on teacher outcomes, particularly TA, across various cultural contexts and educational systems (Cihat & Abdulvehap, 2022). This approach has been advocated by scholars such as Y. Liu, Bellibaş, and Gümüş (2021).
IL is defined as leadership that directly influences the quality of teaching and learning (Gumus et al., 2018). In other words, IL is the active involvement of school principals in enhancing teaching and learning efforts (Ng, 2023). Over the past three decades, numerous frameworks have been developed to enhance IL, demonstrating positive school outcomes in both Western and Eastern contexts (Goddard et al., 2015). As a result, IL continues to be one of the most widely adopted models in school leadership, largely due to its demonstrated impact on school performance and student outcomes (Harris et al., 2017; Leithwood et al., 2020). Despite these advancements, Dorukbaşi and Cansoy (2024) and Y. Liu (2020) emphasize the need for more empirical studies on the impact of leadership on teacher outcomes in non-Western countries.
DL, on the other hand, emphasizes group decision-making involving staff at various levels rather than centralized decision-making by a single individual, typically the principal (Harris, 2009). In other words, school leadership no longer refers solely to the principal’s leadership (Makgato & Mudzanani, 2019). DL is frequently used interchangeably with shared, collaborative, delegated, and dispersed leadership (Spillane, 2005). According to Mayrowetz (2008), DL offers several benefits, including promoting democracy, enhancing efficiency and effectiveness, and fostering human capacity development. Many studies have explored the impacts of DL on various school outcomes, such as TA, teacher instructional quality, and teacher innovativeness (Hsieh, Song, & Li, 2024), TSE (Hsieh, Tai, & Li, 2024), and collective teacher innovativeness and TPC (Hsieh et al., 2023). In this study, the proposed conceptual model was empirically tested using the TALIS 2018 data collected in Vietnam, where traditional Confucian culture is blended with Western ideas and values (Truong et al., 2017), to examine the relationships between leadership styles and TA with TPC and TSE as mediators. Findings thus obtained would contribute to enrich the related literature with more comprehensive evidence of the abovementioned relationships and provide valuable references for principals and countries with contexts like Vietnam in promoting TA to achieve educational reform and improvement.
Literature Review and Hypotheses
Research Context
The concept of contextual leadership, rooted in the understanding that “one size does not fit all” (Harris & Jones, 2022; Leithwood et al., 2020; Noman & Gurr, 2020), is particularly fascinating when applied to Vietnam’s unique education system. Therefore, to account for the potential influence of context on participants’ responses, the study includes a detailed overview of Vietnam’s education system, culture, and school structure. Historically dominated by Confucianism, Vietnam reflects values such as collectivism, respect for elders and authority, and community harmony (Ho, 2024; London & Duong, 2023). These values, central to Vietnamese society, are infused into the organization of the education system and the practices of school leaders (Hallinger et al., 2023; Q. H. N. Tran, Jaakson, & Vadi, 2023; Truong et al., 2017). The concept of losing face also reflects the collective feature of this culture (Hallinger et al., 2018).
Vietnam, which adopted a socialist political system in 1945, has an education ministry that still reflects many Soviet-era characteristics from the 1950s and 1960s, characterized by bureaucratic norms of command and control (Hallinger & Truong, 2016). Recognizing the significance of globalization, Vietnam introduced the Đổi Mới (renovation) and Mở cửa (Open door) policies in 1986 to adapt to global changes. This shift involved incorporating Western models to make education more responsive to labor market needs while maintaining a socialist ideology (Duggan, 2001; Ho, 2024). As a result, Vietnam’s education system reflects a blend of Confucian traditions and contemporary social norms and values (Truong et al., 2017), a blend referred to by Q. H. N. Tran, Jaakson, and Vadi (2023) as the Confucian socialist context. In terms of school structure, every state school in Vietnam operates under the directives and supervision of the Party Standing Board. Members of this board are selected by the School Party Committee and are responsible for decision-making at the school level. Typically, the school’s principal also holds the position of School Party Secretary, thereby leading both the political and administrative functions of the school (Hallinger et al., 2018; Truong et al., 2017). Thus, Vietnam’s educational context reflects three interconnected influences: political dynamics, institutional bureaucracy, and Confucian culture (Hallinger et al., 2018). Research indicates that maintaining relationships with the Communist Party Committees that oversee the school and managing bureaucratic relations with entities like the Department of Education and Training is crucial for school leaders’ success (H. T. Nguyen et al., 2018; Truong et al., 2017). Besides, a “có uy” [prestigious] leader can combine legitimate authority with moral authority to achieve obedience, trust, respect, commitment, and emulation (Hallinger & Truong, 2014; H. T. Nguyen et al., 2023). In short, the intertwining of the Soviet political ideology and Confucian culture has reinforced a bureaucratic structure in which school leaders often hold political and administrative authority, enabling them to represent all school personnel (C. Nguyen, 2023;Truong & Hallinger, 2017).
Conceptual Framework
Wilches (2007) defines TA as personal freedom from external control regarding school matters. Mausethagen and Mølstad (2015) describe TA as teachers having pedagogical freedom, the capacity to justify practices, and local responsibility. Vangrieken et al. (2017) see TA as encompassing self-governance, interdependence, personal choice processes, and the promotion of connectedness. X. N. C. M. Nguyen and Walkinshaw (2018) regard TA as the degree of professional freedom a teacher has over classroom practice, including deciding on teaching materials and curriculum development. All these definitions share a common emphasis on professional freedom. The present study defines TA as conceptualized in the TALIS 2018 data, including determining course content, teaching methods, and assignment amounts, disciplining students, and assessing students’ learning.
With autonomy, teachers can adjust their instructional content and approach to meet students’ needs and expectations and are also responsible for their practices and results (Mausethagen & Mølstad, 2015; D. Yang et al., 2022). Teachers perceive teaching positively when unnecessary controlling elements are removed, and work-related autonomy increases (Martinek, 2019). Choi and Mao (2021), Collie et al. (2018), and Mausethagen and Mølstad (2015) highlight that classroom autonomy helps teachers build self-efficacy and adapt practices to students’ needs. Hsieh, Song, and Li (2024) further confirm that autonomy enhances instructional quality. Supporting TA requires school reforms and decentralization (Vangrieken et al., 2017; Wilches, 2007). Teachers possess autonomy when they can handle all matters independently (Jumani & Malik, 2017). X. N. C. M. Nguyen and Walkinshaw (2018) found that the Vietnamese culture of seniority and harmony affects TA. For instance, teachers with overseas professional training felt hesitant to improve the established syllabus out of respect for seniors and reluctance to jeopardize relationships within their school. A principal’s top-down command regarding the curriculum and methodology also negatively impacts TA (Karimpour et al., 2023).
School leadership styles like DL and IL can promote a quality learning environment (Bellibas et al., 2016). Leadership is crucial for teachers’ professional development (Hulpia et al., 2009), improving TSE ( V. D.Tran et al., 2024), and instructional quality (Bellibaş et al., 2021), ultimately leading to improved student achievement (Robinson & Gray, 2019). IL boosts shared teachers’ classroom practices (Bellibaş et al., 2022). Instructional principals can improve TA by discussing the school’s goals and incentives with teachers and ensuring visibility (Cihat & Abdulvehap, 2022). DL allows multi-level staff to participate in leadership and management in their community (Leithwood et al., 2008; Makgato & Mudzanani, 2019). When teachers have a voice in decision-making, they can find ways to maximize student learning (Skaalvik & Skaalvik, 2021). Leaders should also provide sustainable autonomy by allowing teachers to select teaching methods and formulate curriculum plans (S. Liu, Keeley, et al., 2021).
Xu (2015) found that TPC promotes autonomy among teachers. However, to effectively boost TA, TPC needs to be genuine (Zheng, Yin, & Li, 2019), respectful (Q. Wang & Zhang, 2014), well-regulated (Dobber et al., 2014), suited to teachers’ capabilities and desires (Xu, 2015), and instead of sorting teachers’ errors, collaboration should aim to support them (K. M. Tran, Dang, et al., 2023). Moreover, for teachers to be autonomous, they must have confidence in themselves as professionals and be recognized by others as experts in their field (Dierking & Fox, 2013). Teachers with a sense of self-efficacy also experience a sense of autonomy (Sokmen & Kilic, 2019).
Figure 1 presents the conceptual framework of the current study, where DL and IL are the independent variables directly affecting TPC, TSE, and TA. Contrary to prior research (Lin, 2022; S. Liu, Keeley, et al., 2021) that examined its mediating role, this study treats TA as the outcome variable with TPC and TSE as its predictors. The relationships among DL, IL, TPC, TSE, and TA have not been examined in the same study, so this investigation explores these relationships at multiple levels to fill this knowledge gap.

Conceptual framework.
Relationship of IL and DL With TA
Previous empirical research has revealed that IL positively affects TA (Cihat & Abdulvehap, 2022; Saylik & Sahin, 2021). In this study, IL is defined as the actions a principal takes to improve the quality of teaching and learning. Schools’ instructional leaders define school missions, coordinate curriculums, supervise and evaluate instruction, monitor students’ progress, and provide incentives for teachers, learners, and professionals to improve school learning (Hallinger et al., 2015), instrumental to enhancing school performance (Ng, 2023). Instructional principals can help enhance TA by cultivating a supportive learning environment that encourages teachers to develop their professional skills and classroom environments (Cihat & Abdulvehap, 2022). IL involves principals observing teachers’ instructional practices in the classroom, fostering collaboration among teachers to develop new teaching methods, and ensuring that teachers take responsibility for student learning outcomes (Bellibaş et al., 2022), so instructional leaders should give teachers autonomy to exercise more control over their instructional practices. Based on these views, this study proposes the following hypothesis (
Prior studies have also revealed a positive relationship between DL and TA (Lin, 2022; S. Liu, Keeley, et al., 2021; Özdemir et al., 2023). In this study, as conceptualized in TALIS 2018, DL is defined as involving staff, parents/guardians, and students in making decisions for the school. Principals adopting DL tend to involve staff at multiple levels in decision-making (Harris, 2009; Makgato & Mudzanani, 2019). There are several reasons for principals to delegate more responsibilities to teachers and give them greater autonomy. First, teachers are regarded as experts in their disciplines and the curriculum, which enables them to create a supportive learning environment and produce positive learning outcomes (OECD, 2019). Second, when principals distribute leadership to teachers, it fosters a sense of recognition of their capabilities (Dierking & Fox, 2013). Third, when leadership is distributed, staff will feel a sense of workplace democracy, where decision-making is shared, and everyone’s voice is heard. Granting professional autonomy is encouraged in the 21st century (OECD, 2019). Of importance, this leadership style values individuals’ capabilities, and offering TA is one method of acknowledging these abilities (Xu, 2015). Based on these views, this study proposes the following hypothesis:
Relationship of TPC and TSE With TA
In the study by Q. Wang and Zhang (2014), most teachers reported that they became more autonomous professionals and learned how to conduct research on their daily teaching through a joint action research project. Engaging in action research enables teachers to develop autonomy by critically reflecting on and improving their teaching practices (Dikilitaş & Griffiths, 2017). Similarly, Schipper et al. (2020) found that teachers perceived increased autonomy when collaborating on a study lesson project, as they had a say in the process. Collaboration led to higher autonomy for teachers due to social interdependence (Gülşen & Atay, 2022). In this study, as conceptualized in TALIS 2018, TPC is related to teachers’ activities related to joint instruction, observation and feedback, and collaborative professional learning. Professional collaboration can be understood as collaborative interaction among professionals in a team to achieve a common goal (Brouwer et al., 2012). H. N. Tran et al. (2022) found that collaboration includes sharing ideas, cooperating, providing peer feedback, collegiality, open relationships, peer interaction, and collective engagement. A shared learning culture should ensure awareness and respect for the local context (Q. Wang & Zhang, 2014). Lomos et al. (2011) described collaborative activity as a hands-on approach related to real-time practice and targeted outcomes, such as shaping instructional programs and teaching materials. In addition, genuine collaboration should ensure equal voice and contributions while reaching a shared goal, making collective decisions about whom to assign to what functions, and holding those responsible for the outcomes (McClellan, 2016; D. Nguyen & Ng, 2020). Based on these views, the current study proposes the following hypothesis:
TSE has been found to positively affect TA (Lauermann & Berger, 2021; Sokmen & Kilic, 2019; Valckx et al., 2020). TSE reflects teachers’ belief in their instructional skills, classroom management skills, and capacity to resolve teaching and learning challenges (Li et al., 2022). In the TALIS 2018 survey, TSE includes self-efficacy in instructional practices, classroom management, and student engagement (OECD, 2019). This study conceptualizes TSE, as proposed by Tschannen-Moran and Hoy (2001), as teachers’ judgment of their perceived abilities to execute the three dimensions to attain desired outcomes in student engagement and learning. Fradkin-Hayslip (2021) found that teachers with higher self-efficacy reported greater autonomy in their instructional methods and decision-making processes. Based on these views, the current study proposes the following hypothesis:
Relationship of DL With TPC
Previous studies have revealed that DL encourages teacher cooperation and fosters a school environment where teachers share information and share responsibility for teaching and learning (Halverson et al., 2014; Harris & Jones, 2010; Hsieh et al., 2023). Other recent studies (Bellibaş et al., 2021; Lin, 2022; S. Liu, Keeley, et al., 2021; Yakut Özek & Büyükgöze, 2023) also demonstrated a direct effect of DL on TPC. According to Kools and Stoll (2016), schools that develop a collaborative culture see significant improvements in teaching quality and student outcomes. Zheng, Yin, and Liu (2019) and Bryk and Schneider (2002) emphasize that trust forms the basis for cooperation and pedagogical development, while Hargreaves and Fullan (2012) highlight the importance of professional capital, which includes trust and respect, in transforming teaching practices.
Principals adopting a DL approach encourage collaboration and sharing, creating a school culture where more people can participate in decision-making (Spillane, 2005). Research shows that distributing leadership, rather than confining it to one person, improves school performance, teacher professional development, and student achievement (Leithwood et al., 2008; Makgato & Mudzanani, 2019). As the purpose of DL is to distribute power and responsibility across school members to empower individuals at all levels (Gumus et al., 2018), any beneficial collaboration should be encouraged. Based on these insights, this study proposes the following hypothesis:
Relationship of IL With TPC
Previous studies have shown that IL positively influences TPC (Çoban et al., 2023; Goddard et al., 2015; Hsieh et al., 2023; Supovitz et al., 2010), and frequency of TPC (Mora-Ruano et al., 2021). A quantitative study by Zheng, Yin, and Li (2019) identified a direct effect of IL on professional learning communities. Principals actively monitor instruction and provide robust instructional guidance to help improve teachers’ collective work; hence, principal IL should be considered a key recruitment criterion for school principals (Goddard et al., 2015). Leaders adopting IL make efforts to improve learning and teaching in their schools (Ng, 2023), and hence, TPC can be considered to enforce such efforts. Additionally, to enhance the principal’s instructional influence on TPC, it is essential to foster genuine collaboration among teachers (Zheng, Yin, & Li, 2019). Additionally, the effects of trust in IL on instruction must be ensured among teachers (Çoban et al., 2023). However, this relationship should be further investigated within a single society or across different countries to understand whether and how the practices defining effective IL vary among diverse societies (Çoban et al., 2023; Walker & Hallinger, 2015). IL involves principals observing classroom practices, encouraging collaboration, and holding teachers accountable for student outcomes (Bellibaş et al., 2022), so based on these characteristics of IL, this study proposes the following hypothesis:
Relationship of DL With TSE
Zheng, Yin, and Liu (2019b) found that trust mediates the relationship between DL and TSE. Additionally, Sun and Xia (2018) identified a direct impact of DL on TSE, reinforcing the idea that distributing leadership roles among staff enhances their confidence and effectiveness. Choi (2023) and J. Liu et al. (2023) provide evidence that DL can enhance TSE by involving teachers in decision-making processes and fostering a sense of ownership and responsibility. Principals distributing their leadership responsibilities benefit from improved instructional practices and overall school performance (Spillane et al., 2015). To enhance TSE, it is crucial for school administrators to encourage teachers’ involvement in school improvement initiatives and establish a collaborative feedback network, allowing teachers to receive input from various sources (Choi, 2023). By enhancing teachers’ sense of responsibility in their instructional roles and offering opportunities for involvement in decision-making, DL contributes to the improvement of TSE. Considering these views, this study proposes the following hypothesis:
Relationship of IL With TSE
Although Zheng, Yin, and Li (2019) found only an indirect association between IL and teacher TSE, fully mediated by professional learning communities, many other studies have unveiled a positive IL-TSE relationship (Alanoglu, 2022; Çoban et al., 2023; Y. Liu et al., 2022). Educational policymakers should emphasize training on IL practices for enhancing TSE, which in turn improves teaching quality and student outcomes. Alanoglu (2022) highlights that IL practices foster a supportive environment where teachers feel empowered and capable, directly impacting their confidence and effectiveness in the classroom. Hoy and Woolfolk (1993) found that when teachers receive clear and constructive feedback, their sense of efficacy increases. Moreover, Fackler and Malmberg (2016) demonstrated that principals who support teachers, especially those working with less competent students, significantly enhance teacher self-efficacy. IL behaviors include fostering a positive learning environment, observing classroom instruction, and offering feedback (X. Ma & Marion, 2021), which can lead to enhanced TSE. With these views, the current study proposes the following hypothesis:
Relationship of TPC With TSE
Chong and Kong (2012) investigated how TSE can be influenced by lesson study, a method promoting teachers’ inquiry mindset, sharing knowledge and experiences, and collaborative development of classroom materials. Teachers reported that this collaborative work improved self-efficacy. Collaboration practices like teamwork and participation in learning communities can impact TSE, job satisfaction, and enthusiasm (Skaalvik & Skaalvik, 2017). Similarly, Schipper et al. (2020) found a positive effect of lesson study on TSE, which aligns with the positive impact of TPC on TSE observed by Çoban et al. (2023). Productive collaboration allows teachers to exchange knowledge about their theories, methods, and teaching processes, and learn how to improve classroom instruction delivery (Chong & Kong, 2012). A collaborative environment promotes continuous improvement and career-long learning, which is crucial for improving instructional capacity (H. N. Tran et al., 2022). The more teachers collaborate on teaching matters, the more efficacious they perceive themselves to be (Goddard et al., 2015). Of note, for effective TPC, it must be authentic, considerate, well-structured, aligned with teachers’ competencies and aspirations, and focused on supporting rather than correcting teachers (Dobber et al., 2014; K. M. Tran, Dang, et al., 2023; Q. Wang & Zhang, 2014; Xu, 2015; Zheng, Yin, & Li, 2019). Based on these views, the current study proposes the following hypothesis:
Method
Data Source
This study performed a secondary analysis of the TALIS 2018 survey data provided by the OECD, which is publicly available. Vietnam participated in this survey between March and May 2018. The OECD applied a two-stage stratified sampling approach: the first stage involved selecting 200 schools, and the second stage involved selecting 20 teachers from these schools. This study utilized the International Standard Classification of Education (ISCED) level-2 data for Vietnam for many compelling reasons: The hierarchical structure of the data, with teachers nested within schools, enabled the examination of variables at both teacher and school levels (Sun & Xia, 2018). TALIS provided extensive and comprehensive coverage through well-established measures of critical variables, including DL, IL, TPC, TSE, and TA. TALIS employed rigorous methodology and standardized procedures, ensuring high data quality and reliability. The strong response rates, particularly among principals (97%), indicated a good representation of the target population. Vietnam’s unique educational context, which blends Confucian traditions with socialist principles, provided a unique context for investigating these relationships. As a result of this combination of factors, TALIS 2018 was a good data source for investigating the relationship between school leadership and teacher outcomes in Vietnamese schools. The study included responses from 196 lower secondary school principals (98% response rate) and 3,825 teachers (95.6% response rate). After data cleaning, 1,835 valid responses from teachers (48% response rate) and 191 from principals (97% response rate) remained. There were more male teachers (69.9%) than female teachers (30.1%). Most teacher-respondents (79.7%) held bachelor’s degrees, while some (17.2%) had received short-cycle tertiary education (ISCED level 5). Their average teaching tenure was 15.80 years. Most principals were male (71.9%), with 85.2% holding bachelor’s degrees and 10.2% holding master’s degrees. Their average work tenure was 9.13 years.
Variables
DL
The study used three items from the teacher questionnaire to measure DL, as previously studied by Bellibaş et al. (2021) and Hsieh et al. (2023). Teachers responded on a 4-point Likert scale ranging from “strongly disagree” to “strongly agree.” The items included: “This school provides staff with opportunities to actively participate in school decisions,”“This school provides parents or guardians with opportunities to actively participate in school decisions,” and “This school provides students with opportunities to actively participate in school decisions.”
IL
The study used five items from the principal survey to measure IL, as previously studied by Bellibaş et al. (2021) and Hsieh et al. (2023). Principals responded on a 4-point Likert scale from “never/rarely” to “very often.” The items included: “I observed instruction in the classroom,”“I took actions to support cooperation among teachers to develop new teaching practices,”“I took actions to ensure that teachers take responsibility for improving their teaching skills,”“I took actions to ensure that teachers feel responsible for their students’ learning outcomes,” and “I provided parents or guardians with information on the school and student performance.”
TPC
The study used four items from the teacher questionnaire to measure TPC, as previously investigated by Hsieh et al. (2023) and Torres (2019). Teachers responded on a 6-point Likert scale from “never” to “once a week or more.” The items included: “Teach jointly as a team in the same class,”“Observe other teachers’ classes and provide feedback,”“Engage in joint activities across different classes and age groups (e.g., projects),” and “Take part in collaborative professional learning.”
TSE
The study used 12 items across three dimensions from the teacher questionnaire to measure TSE, as previously explored by Burić and Kim (2021) and Hsieh et al. (2023). Teachers responded on a 4-point Likert scale from “not at all” to “a lot.” The first dimension, efficacy in instructional practices, included items such as: “Craft good questions for students,”“Use a variety of assessment strategies,”“Provide an alternative explanation when students are confused,” and “Vary instructional strategies in my classroom.” The second dimension, efficacy in classroom management, included items like: “Control disruptive behavior in the classroom,”“Make my expectations about student behavior clear,”“Get students to follow classroom rules,” and “Calm a student who is disruptive or noisy.” The third dimension, efficacy in student engagement, included: “Get students to believe they can do well in schoolwork,”“Help students value learning,”“Motivate students who show low interest in schoolwork,” and “Help students think critically.”
TA
The study used five items from the teacher questionnaire to measure TA, as previously studied by Lin (2022). Teachers responded on a 4-point Likert scale from “strongly disagree” to “strongly agree.” The selected items included: “Determining course content,”“Selecting teaching methods,”“Assessing students’ learning,”“Disciplining students,” and “Determining the amount of homework to be assigned.”
Control Variables
For the principals, the control variables are gender, educational qualification, years of experience, school location, private and public schools, and socioeconomic status (SES). For the teachers, the control variables are gender, educational qualification, and years of experience (Hsieh et al., 2023).
Analytical Strategy
The mean scores of variables were estimated for analysis. For each individual response, the score of a specific latent variable was calculated as the mean score of the items designed to measure that variable. The mean score of the latent variable was then obtained by dividing the total sum of these scores by the number of valid samples. This method was applied to the mean scores of TPC, TSE, and TA at the teacher level, and IL at the school level. For DL at the school level, the DL scores of teachers within the same school were averaged to obtain the school-level DL score. The mean score of school-level DL was then calculated by dividing the total sum of these school-level scores by the number of schools. The analysis followed these steps: The appropriateness of multilevel analysis was examined. The intraclass correlation (ICC) was calculated to determine if there were significant variations among TPC, TSE, and TA at the school level. To test the hypotheses, multilevel regression analysis was conducted to examine the relationships between IL and TA, DL and TA, TPC and TA, TSE and TA, DL and TPC, IL and TPC, DL and TSE, IL and TSE, and TPC and TSE. Hierarchical linear modeling (HLM) version 7 (Raudenbush et al., 2017) was used for the analysis. HLM is suitable for analyzing data with teachers nested within schools (X. Ma & Marion, 2021) and helps identify relationships between predictor and outcome variables at both the school and teacher levels (Hofmann, 1997).
Results
Relationships Between Variables
Table 1 presents descriptive statistics and reliability coefficients for two levels of participants: teachers (Level 1, n = 1,835) and principals (Level 2, n = 191). At level 1, TPC had a mean score (M) of 3.22, a standard deviation (SD) of 0.87, and Cronbach’s α coefficients (α) = .73; TSE had M = 3.43, SD = 0.39, α = .88; and TA had M = 3.14, SD = 0.49, and α = .86, indicating good internal consistency. The correlation coefficients of the three range from .15 to .23 (p < .001), indicating a significantly positive correlation. At Level 2, DL had M = 2.87, SD = 0.24, α = .75; and IL had M = 3.02, SD = 0.36, α = .67, indicating good internal consistency. The correlation coefficient between the two was .15 (p < .05), showing a significantly positive correlation.
M, SD, and Correlations Among Variables.
Note. Cronbach’s α are given in parentheses.
p < .05. ***p < .001.
Null Model
Null model analysis provides information about whether outcome variables of interest vary among schools. As shown in Table 2, the within-group component (σ2) of TPC, TSE, and TA respectively, .001), 13.0% (p < .001) and 16.1% (p < .001), respectively; while the between-group component (τ00) was 20.6% (p < .001), 2.5% (p < .001) and 7 % (p < .001), respectively; indicating variations in TPC, TSE, and TA among different schools. The intraclass correlation coefficients (ICC) of TPC, TSE, and TA were 27.9%, 16.0% and 30.2%, respectively, above the high correlation threshold of 13.8% (Cohen, 1988), suggesting that school-level variations cannot be ignored, thereby justifying subsequent multilevel analyses (Raudenbush & Bryk, 2002).
Results of Null Model.
Note. ICC = τ00/(τ00 + σ2).
p < .001.
Hypothesis Testing
This study performed Bonferroni correction for multiple comparisons with an adjusted significance level. The Bonferroni corrected p value was calculated as the original p value divided by the number of hypotheses. That is, the Bonferroni corrected p value was estimated to be .006 (= .05/9) with a pre-determined significant p value of .05 divided by the number of hypotheses. Table 3 lists the HLM results of hypothesis testing and Figure 2 depicts the study model with parameter estimates. As shown in Table 3, the path coefficient of DL and TPC in Model 1 (β = .67, p < .001) was significant, thus supporting H5; while the effect of IL on TPC was insignificant (β = .06, p < .589), leaving H6 unsupported. Results of Model 2 showed a positive association of DL with TSE (β = .23, p < .001), supporting H7; a positive association of IL with TSE (β = .07, p < .034), supporting H8; and a positive association of TPC with TSE (β = .09, p < .001), supporting H9. Model 3 results indicated that DL is significantly and positively related to TA (β = .20, p < .039), supporting H2, and that TSE is significantly and positively associated with TA (β = .22, p < .001), supporting H4. In contrast, the relationships of IL with TA (β = .04, p < .410) and of TPC with TA (β = .02, p < .090) were insignificant, leaving H1 and H3 unsupported. Table 4 summarizes the results of hypothesis testing.
Results of HLM.
Note. p Value is given in parentheses.
p < .006. **p < .01. ***p < .001.

Study model diagram with parameter estimates.
Results of Hypothesis Testing.
Discussion
The study attempts to use the proposed framework to foster TA in an educational setting. The significant effect of DL on TA was empirically supported as principals and teachers work in the same environment where they share the same culture with Confucian influence and are under the standard guidance and supervision of the School Party Committee in Vietnam (Truong et al., 2017). Contrary to previous results evidencing the positive relationship between IL and TA (Cihat & Abdulvehap, 2022; Saylik & Sahin, 2021), the current findings do not support such an association (H1). That is, IL practices in participating schools did not work well to promote teachers’ professional autonomy. A possible precondition for TA enhancement is reciprocal professional trust between principals and teachers (Kools & Stoll, 2016). In addition, trust in principals can derive from moral leadership. Hallinger et al. (2018) use the metaphor “We need an iron hand, but the iron hand must be a clean hand” to represent the balance between “autocratic leadership” and “moral leadership” in Vietnamese school leadership. Ethical or prestigious leadership has increasingly been recognized as a critical component of educational leadership, particularly in fostering moral values within school communities (Gumus et al., 2018). By adopting a prestige-based or moral leadership style, a principal can foster trust with stakeholders, as this approach encourages subordinates to feel trust, show respect, and aspire to emulate their leader (H. T. Nguyen et al., 2023). Another explanation for this unassociated relationship can be that Vietnamese school leadership practices are shaped by the country’s political and cultural orientation (Trinh et al., 2019), which can suggest the complexities of Vietnamese school leadership. In addition, the Confucian culture remains a solid foundation in Vietnamese schools, and the norms and high power-distance culture may make teachers hesitant to voice opinions or question authority (Ho, 2024; London & Duong, 2023; Truong et al., 2017). Besides, in the absence of such a relationship, school leaders who deemed teachers not professionally ready might be reluctant to grant teachers greater autonomy (OECD, 2019). The Vietnam TALIS 2018 found that only 57% of lower secondary school principals had completed an instructional program prior to taking up their position as principals (OECD, 2019), which might suggest insufficient IL training. Rodrigues and Ávila de Lima (2024) found that principals can be constrained by the national and regional policy context and the educational system’s organization to become instructional leaders. For example, several system centralizations have been identified as a hindrance to promoting IL in many circumstances, such as China (Walker & Qian, 2022), Iran (Hallinger & Hosseingholizadeh, 2020), Portugal (Espuny et al., 2020), and Vietnam (Hallinger et al., 2017). Hence, principals’ lack of confidence as instructional leaders and teachers’ lack of trust in principals’ IL may also account for the insignificant relationship between IL and TA among schools in Vietnam.
In contrast, DL was found to have direct influence on TA (H2). The present findings echo those of Lin (2022) and Özdemir et al. (2023), that highlighted the importance of shared leadership in allowing teachers greater autonomy. Moreover, the results evidenced that principals in Vietnam, viewing and valuing teachers as front-liners in the teaching profession and in direct contact with students and parents, are ready to share leadership with them and involve them in decision-making.
The absence of a direct relationship between TPC and TA (H3) observed in this study contrasts with the findings of Q. Wang and Zhang (2014) and Schipper et al. (2020) that joint projects enhanced TA. Though collaboration is indispensable and of value, the possible lack of reciprocal trust among teachers (Bryk & Schneider, 2002; Kools & Stoll, 2016; V. D. Tran et al., 2024; Zheng, Yin, & Liu, 2019) as well as the type and level of collaborative work (Xu, 2015) may not contribute to TA. In addition, the practice of Confucian values and hierarchical structures in Vietnam might also affect TA. Less experienced teachers tend to accept and follow their senior counterparts’ ideas to maintain harmony when collaborating. Or the more experienced might take over the collaboration process; senior teachers, who often have greater authority and experience, may only collaborate if their status is maintained, while younger teachers are more likely to be afraid of challenging authority (Ho, 2024; London & Duong, 2023; Truong et al., 2017), leaving the less experienced little room to voice their thoughts (Schipper et al., 2020). Hence, open discussion with respect and equal participation might need further promotion while collaborating in this context (McClellan, 2016; D. Nguyen & Ng, 2020; Q. Wang & Zhang, 2014), particularly considering that collectivist cultures place a high value on preserving face for all participants (Hallinger et al., 2018).
The positive effect of TSE on TA (H4) evidenced in this study is in line with those found by Sokmen and Kilic (2019), Valckx et al. (2020), and Lauermann and Berger (2021). That is, teachers can be expected to become more professionally autonomous when they are competent in their teaching (Sokmen & Kilic, 2019). Hence, in these schools, teachers can be said to have professional autonomy if they are professionally efficacious.
The direct influence of DL on TPC (H5) echoes the findings of Halverson et al. (2014) and Yakut Özek and Büyükgöze (2023), evidencing the significant impact of leadership on school collaborative culture, where school leaders are unlikely to abuse their powerful commands (Hallinger & Truong, 2016). In contrast, the absence of direct influence of IL on TPC (H6) disagrees with the results observed by Goddard et al. (2015), Zheng, Yin, and Li (2019), and Çoban et al. (2023). The current findings indicated that instructional leaders could not promote professional collaboration. A possible reason can be that principals’ school administrative routine work could have taken up most of their time (H. T. Nguyen et al., 2018; Truong et al., 2017). Moreover, as mentioned above, just slightly over half of the principals in Vietnam had received instructional training (OECD, 2019). The remaining principals, and even those trained ones, might find it difficult to delve in depth into instructional practice. Within a collaborative culture, ensuring reciprocal trust among related parties is crucial (Bryk & Schneider, 2002; Çoban et al., 2023; Hargreaves & Fullan, 2012; Kools & Stoll, 2016; Zheng, Yin, & Liu, 2019). Additionally, cultural factors such as Confucianism can hinder the adoption of new ways of thinking, being, and doing grounded in the principles of collaboration and equality among members in the community (McClellan, 2016; D. Nguyen & Ng, 2020; K. M. Tran, Dang, et al., 2023). Finally, school leaders are often required to perform political and bureaucratic functions throughout their principalship, leading many to prioritize political and managerial responsibilities over instructional leadership (Truong & Hallinger, 2017; H. T. Nguyen et al., 2018; Rodrigues & Ávila de Lima, 2024; Truong et al., 2017).
In line with previous studies (Choi, 2023; J. Liu et al., 2023), DL had a significant and positive impact on TSE (H7). With principals willing to share leadership, teachers would perceive themselves as more capable of their profession and trusted. Similarly, IL also had a positive impact on TSE (H8), echoing the findings of Alanoglu (2022), Çoban et al. (2023), and Thien (2023). Thus, both DL and IL contribute to TSE enhancement.
Finally, the results showed a positive association between TPC and TSE (H9), which agrees with the findings of Schipper et al. (2020) and Çoban et al. (2023). That is, in professional collaboration or quality joint activities, teachers perceive a greater sense of self-efficacy and find themselves playing a more decisive role, thus driving them to take necessary actions for further enhancing self-efficacy.
Implications
Most principals were found to prioritize administration and management rather than student learning; They were not responsible for leading in instructional areas (Truong & Hallinger, 2017; H. T. Nguyen et al., 2018; Rodrigues & Ávila de Lima, 2024; Truong et al., 2017). This study’s finding is no exception. Therefore, for IL to have an impact on TA, principals should actively take courses in instructional training to equip themselves with the necessary expert knowledge and work alongside teachers for teaching effectiveness and school improvement (Cihat & Abdulvehap, 2022). In addition, as teachers work directly with students and curricula, principals should confidently give teachers more professional autonomy and eliminate obstacles to professional autonomy (Martinek, 2019; OECD, 2019) so that teachers can adjust their methods and curricula to meet students’ needs and improve students’ learning outcomes (Dierking & Fox, 2013; Hsieh Song, & Li, 2024; Jang et al., 2023; Mausethagen & Mølstad, 2015; D. Yang et al., 2022).
Considering the significant and positive relationships between TPC and TSE, it is essential to involve teachers more in decision-making, curriculum planning, and classroom practices. This involvement fosters their sense of autonomy and respect for their professional expertise (Dierking & Fox, 2013; X. N. C. M. Nguyen & Walkinshaw, 2018; OECD, 2019). Promoting DL in schools (Harris, 2009; Makgato & Mudzanani, 2019) and fostering professional collaboration can cultivate and optimize teachers’ strengths in a respectful, sharing, and supportive work environment. The current findings indicate that the principals adopted DL and gave teachers a say in their professional matters, suggesting a shift away from traditional top-down commands (Hallinger & Truong, 2016) and they also promote TPC and TSE.
This study found no effect of IL on TPC. Theoretically, if a principal adopts IL, he or she should manage instructional domains at their best for teaching effectiveness and school improvement (Goddard et al., 2015; Rodrigues & Ávila de Lima; 2024). Doing so, they may try to promote TPC (Zheng, Yin, & Li, 2019), provide quality instructional guidance to promote teachers’ collective work (Goddard et al., 2015), and establish IL trust in teachers (Çoban et al., 2023). Regarding promoting trust, principals may use trust-building strategies by enhancing their prestigious images (Gumus et al., 2018; Hallinger & Truong, 2014; Hallinger et al., 2018; H. T. Nguyen et al., 2023), and therefore improving their instructional ability is one way to reinforce their prestige. Regarding collective work, Vietnam’s Communist Party and Confucian philosophy view collective effort as leverage to achieve the school’s goals (Cihat & Abdulvehap, 2022; Hallinger et al., 2017). However, this current result does not seem to enhance this nuanced relationship. Hence, for IL to have an impact on TPC, principals should promote genuine collaboration among teachers, create IL trust, and be instructional pioneers by taking more proactive actions to encourage teachers to collaborate. In addition, principals should prioritize IL over the administration, as schools are teaching and learning centers.
The positive effect of IL on TSE suggests that principals adopting IL can significantly boost TSE through active involvement in teaching and learning activities (Thien, 2023). TSE is influenced not only by IL and DL styles but also by TPC. Therefore, principals should foster quality TPC, enabling teachers to share their teaching experiences and practices, which, in turn, enhances TSE. Creating a school climate that includes protecting instructional time, maintaining high visibility, offering teacher incentives, promoting professional development, and encouraging learning is crucial for promoting TSE (Cihat & Abdulvehap, 2022; Rodrigues & Ávila de Lima, 2024).
This study also observes that TPC does not enhance TA. Theoretically, this may be because both TPC and TA operate at the teacher level, meaning teachers cannot grant themselves professional autonomy through collaboration without principals’ involvement. Another possible explanation is the influence of Confucian culture and hierarchical structures, where younger teachers are viewed as less experienced and tend to follow the decisions of their older colleagues. This dominance of experienced teachers in collaboration can negatively impact TA (Q. Wang & Zhang, 2014). To address this, teachers should be encouraged to openly discuss approaches to achieve their school’s visions and goals, regardless of age or seniority. As K. M. Tran, Dang, et al. (2023), D. Nguyen and Ng (2020), McClellan (2016), and Q. Wang and Zhang (2014) suggested, collaboration must ensure equal, respectful participation, clear roles and responsibilities for outcomes. Furthermore, recognizing everyone’s expertise and strengths fosters a culture of mutual trust and respect (Çoban et al., 2023; Q. Wang & Zhang, 2014; Zheng, Yin, & Li, 2019).
Limitations and Future Research Directions
TALIS is a comprehensive international survey with a well-established conceptual framework developed from related literature. However, these established constructs may not fully capture context-specific components, such as the cultural dimensions that may influence teacher outcomes in Vietnam (Hallinger et al., 2023). Future iterations of TALIS could benefit from more explicitly cultural items in the survey, such as questions about cultural norms, values, and beliefs, particularly in countries with unique cultural traditions and school structures that shape education systems. In addition, Spillane (2005) noted that leadership often involves multiple leaders, both formal and informal. However, in TALIS 2018, DL includes only school staff, parents/guardians, and students in decision-making processes.
The results indicate inconsistencies within this framework, with IL positively affecting TSE but not significantly impacting TPC or TA. This might be due to the dimension of managing the instructional program not exhibiting optimal measurement properties in each country (Veletić & Olsen, 2021). The lack of impacts of IL on TA and TPC and TPC on TA may be linked to Confucian values and hierarchical structures enhanced by the bureaucratic norms of command-and-control influence of Soviet-era characteristics, which create social power distance. To deepen the interpretation of quantitative findings, future research should involve semi-structured interviews and observations with teachers and principals to explore their perspectives of cultural values and school structure that may affect TA, including respect for hierarchy and decision-making authority. Theoretically, IL’s focus on teaching effectiveness and school improvement should enhance TPC and TA, suggesting a need for further research in similar contexts. Future studies should also consider longitudinal and experimental designs (Kılınç et al., 2023; OECD, 2019) to better understand the causality among these currently investigated variables.
Conclusions
In summary, regarding the influence of leadership on TA, only DL has a direct impact, while IL does not. As for the mediating roles of TPC and TSE, only TSE mediates the relationship of DL with TA and of IL with TA, while TPC does not. With this result, DL both directly and indirectly impacts TA through TSE. IL does not affect TPC, whereas DL does. TPC has no direct influence on TA but can promote TSE. The results suggest that IL may be less effective in the Vietnamese context, potentially influenced by Confucian culture, school structure, and educational heritage. The findings contribute valuable insights to the relatively limited research on educational leadership in Vietnam (Hallinger et al., 2017). Future research should examine these relationships longitudinally to strengthen educational leadership practices and evaluate the effectiveness of enhanced IL training for principals. Further, although TPC had no direct effect on TA, it positively affected TSE, suggesting that schools should foster collaborative environments to promote TA indirectly. These insights can help educational policymakers and school leaders develop more effective approaches to enhancing TA within the Vietnamese cultural context.
Footnotes
Acknowledgements
The authors would like to acknowledge the Organisation for Economic Co-operation and Development for providing access to the TALIS 2018 data, which formed the foundation of this research.
Ethical Considerations
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
All data will be provided by the authors upon request by the editor.
