Principal self-efficacy for instructional leadership in the perspective of principal strengthening training: work engagement, job satisfaction and motivation to leave

Abstract The current study explored relations between principal self-efficacy for instructional leadership, work engagement, job satisfaction, and motivation to leave the work as a principal from the principal strengthening training perspective. Participants in the study were 125 principals in the junior school selected using a saturated sampling procedure. This study uses quantitative methods with structural equation modeling (SEM). SEM analysis revealed that principal self-efficacy for instructional leadership was positively related to work engagement and job satisfaction. The relationship between principal self-efficacy for instructional leadership and motivation to leave was not significant. Work engagement and job satisfaction are positively related to motivation to leave. This study showed that Indonesian government policies related to the training of strengthening principals improve principal self-efficacy for instructional leadership. Based on the results of this study, the successful implementation of principal strengthening training in Indonesia can be adopted by policymakers from other countries because it can increase self-efficacy for instructional leadership, work engagement, job satisfaction, and prevent motivation to leave from the principal. In particular, this principal strengthening training program uses the ORPAER intervention model (observe, reflect, plan, act, evaluation, and reflect) combined from the syntax of Theory U and the ILEAD (Introduction, Link, Enforcement, Awareness, and Development) approach. This intervention model can improve the competence of principals, the principals’ problem-solving skills, creative thinking, and emotional engagements.


PUBLIC INTEREST STATEMENT
Principal strengthening training is a program that can improve principal self-efficacy for instructional leadership, which impacts increasing job satisfaction and work engagement of principals and can even prevent principals from leaving their positions as principals. SEM analysis revealed that principal self-efficacy for instructional leadership was positively related to work engagement and job satisfaction. The relationship between principal self-efficacy for instructional leadership and motivation to leave was insignificant. Work engagement and job satisfaction are positively related to motivation to leave. This study showed that Indonesian government policies related to the training of strengthening principals improve principal self-efficacy for instructional leadership. Hence, to increase principal self-efficacy for instructional leadership, job satisfaction, and work engagement and prevent principals from motivation to leave-the principal needs to be required and facilitated in the principal strengthening training program. This program should be a core value for education policy makers and educational actors involved in all levels of the education sector.

Introduction
The success and failure of a school depend largely on how the principal performs the role in creating an idea to improve the quality of the school. Therefore, the principal as a leader in the school has a very significant and strategic role and responsibility in improving the quality of education in the school (Agasisti et al., 2020;Buske, 2018;Liebowitz & Porter, 2019;Ni et al., 2018;Satyawati & Siswoyo, 2018;Senjaya, 2019;Sunaengsih et al., 2019).
One of the leadership styles adopted by principals in carrying out their roles and responsibilities for improving the quality of education in schools is instructional leadership. Over the past 50 years, principals have gradually accepted instructional leadership in performing their roles as principals around the world (Bush, 2013;Hallinger & Wang, 2015;Walker et al., 2015). This reasoning is based on the research results that have explored the influence of instructional leadership on student learning, school quality, and school improvement (Hallinger, 2011;Leithwood et al., 2010;V.M. Robinson, 2006).
The role of the principal can be described as one of bureaucratic and management responsibilities, such as responsibility for the school economy, facilities, schedule, and personnel . In an observational study, Horng et al. (2009) found that principals spend much of their days on administrative and organizational management tasks, whereas they have little time for educational leadership. Møller and Ottesen (2011) assert that over the past few decades, the responsibilities of principals in many countries have been extended to all aspects of school management, including student learning, development of goals and vision, establishing evaluation procedures, and the development of a safe and stimulating learning environment. These situations and conditions present challenges in developing countries, both Western societies (Hallinger, 2011;Hallinger & Wang, 2015;Leithwood et al., 2010;Marks & Printy, 2003;V.M. Robinson, 2006) and Asian, African, and Latin American communities (Bellibas et al., 2016;Bush, 2013;Fromm et al., 2016;Walker et al., 2015) when the principal performs his or her role traditionally being an "administrator" is asked to be an "instructional leader" (Abdollahi et al., 2013;Bhengu et al., 2014). Thus, a study from UNESCO (2016) on school leadership policies in South and West Asia identified the need for new approaches to prepare school leaders with expanded roles. This concept of principal responsibility is a global trend that has led to a stronger focus on instructional leadership and leadership for learning (Hallinger, 2010;Liu & Hallinger, 2018).
The application of instructional leadership has constraints as found by instructional leadership policymakers. The results of research by scholars have highlighted limitations in the time, confidence, and capacity of principals to fulfill this role (Hallinger & Murphy, 2013). In addition, the function of the principal is affected by their working conditions, their stress levels, and their mastery of expectations (Agasisti et al., 2020). Recognition of these constraints has prompted the need for additional research to understand the contextual and personality factors that shape the behavior of school leaders in their efforts to improve school outcomes (Hallinger, 2011;Leithwood & Jantzi, 2008). Notably, among the personality antecedents of leadership identified by scholars is self-efficacy or the belief that "I can make a difference" (Brama, 2004;Leithwood & Jantzi, 2008;Nir et al., 2006;R.A. Federici & Skaalvik, 2011).
The problem is that, surprisingly there is still little research exploring the principal's perception of their work environment and working conditions and how it is associated with their experiences of emotional exhaustion, job satisfaction, and motivation to work as a principal (Skaalvik, 2020a). In addition, many studies focus on teacher self-efficacy-research evidence shows that teacher self-efficacy is positively related to engagement and job satisfaction and negatively related to emotional exhaustion and motivation to leave the teaching profession (Collie et al., 2012;Klassen & Chiu, 2011). Meanwhile, little research is done related to the principal's selfefficacy for instructional leadership (Skaalvik, 2020a). At the same time, leaders with stronger self-efficacy will tend to communicate and become models of higher performance expectations for teachers and students (Gareis & Tschannen-Moran, 2005;Leithwood & Jantzi, 2008;Lucas, (2003).The) and creating increasingly effective school functions (Goddard et al., 2020). Recent research conducted in Norway shows that self-efficacy for instructional leadership is predictive of the principal's wellbeing (emotional exhaustion) and motivation (in this study measured both as engagement and as motivation) to continue or leave the position of principal (Skaalvik, 2020a). Previous research ignored one important variable: job satisfaction. Previous studies have shown that principals' self-efficacy is positively related to job satisfaction (Evers et al. 2002;R. A. Federici & Skaalvik, 2012;Friedman, 1995Friedman, , 2002Skaalvik & Skaalvik, 2010). Job satisfaction is negatively related to the motivation to quit (R. A. Federici & Skaalvik, 2012). A literature search of the principal's self-efficacy for instructional leadership showed no study focused on the relationship between these variables in a similar study. Therefore, to fill this gap, the current study will add a job satisfaction variable because it becomes an essential factor in influencing the principal in making decisions whether to remain or leave a profession or position as a principal (R. A. Federici & Skaalvik, 2012;Chan et al., 2008). In addition, previous studies are based on Norwegian samples and need to be replicated in other cultures (Skaalvik, 2020a), such as in Indonesia. The results show that some principals think that instructional leadership cannot be applied to a hierarchical Indonesian education system where the principal is still dependent on instructions from the government, teachers still depend on instructions from the principal, and students still rely on instructions from teachers (Sofo et al., 2012). This study focuses on the self-efficacy of principals for instructional leadership in Indonesia. In addition, the study will explore the principal's self-efficacy and how the principal's self-efficacy aspect relates to job engagement, job satisfaction, and motivation to leave a position as principal. On this basis, the purpose of this study is to explore how self-efficacy for the principal's instructional leadership relates to work engagement, job satisfaction, and motivation to leave a position as principal from the perspective of the Principal Strengthening Training (PST).

The educational context for instructional leadership in Indonesia: policy perspective
In Indonesia, instructional leadership models have been studied by previous researchers with contradictory results. Instructional leadership is considered one of the determinants of the successful implementation of the 2013 curriculum because the focus of this model is towards a curriculum where the principal coordinates the curriculum, supervises and evaluates the curriculum, and monitors student development (Author, 2019).
Although instructional leadership has been implemented, this model has not worked properly (Yunita, 2015). There are pros and cons to implementing instructional leadership in the Indonesian education system. Most principals consider that instructional leadership can be applied because it focuses on improving student achievement outcomes (Yunita, 2015). However, recent research shows that the leadership role of principals in Indonesia has not focused on managing the teaching and learning process to improve the school's ranking compared to Malaysia (Rizkita & Supriyanto, 2020). In fact, other research results revealed that the principal's instructional leadership practices in Indonesia are mostly conducted in the third dimension, namely maintaining high visibility compared to the other two dimensions, setting the vision and mission, and managing teaching programs (Dania & Andriani, 2020).
To solve this problem, the Directorate General of Teachers and Education Personnel of the Ministry of Education and Culture of the Republic of Indonesia made a policy to implement the Principal Strengthening Training (PST). This is one of the priority programs in 2020, which aims to improve the competence of the principal to be able to lead and manage the school well so that students get happiness (students wellbeing) during the education process in the school

Intervention: Principal strengthening training for instructional leadership
Principal Strengthening Training is conducted in 3 (three) learning stages, namely On the Job Training 1 (OJT 1), In-Service Training (IST), and On the Job Training 2 (OJT 2). Learning at the OJT 1 stage is face-to-face mentoring (offline mode) with a form of self-assignments using the syntax of Theory U (Scharmer, 2007). Assignment in the form of observation and reflection activities by studying learning materials, identifying learning problems faced in schools and unresolved using the instruments provided, and putting together good practices that have been done during the principal's tenure. IST learning is an activity using ILEAD (Introduction, Link, Enforcement, Awareness, and Development) approach with the learning steps written in this learning material. At this stage, participants (principals) share learning problems, find solutions to problems in discussion forums, and share good practices. After sharing, participants draw up a School Development Plan based on learning problems that are a priority to be completed and developed. OJT 2 learning using Theory U (Act, Evaluate, and Reflect) syntax is conducted face-to-face (offline mode). Participants (principals) actively communicate with teachers to compile school development reports and make video performances that will be sent through YouTube (Education Learning Materials and Principal Strengthening Training, 2020). The model of intervention in the implementation of the Principal Strengthening training can be seen in figure 1.
The time used to learn this learning material is 71 hours of lessons (JP) @ 45 minutes equivalent to 24 days (face-to-face takes as much as 21 days) as follows: 1). On-the-Job Training 1 (OJT 1), as much as 10 JP or equivalent to 5 working days (2 JP per day); 2). In-Service Training (IST), as much as 40 JP or equivalent to 8 days with virtual face-to-face details and chat in LMS as much as 5 JP/ day. As for face-to-face as much as 40 JP or equivalent to 5 days; and 3). On-the-Job Training 2 (OJT 2), as much as 21 JP or equivalent to 11 working days (2 JP per day) (Education Learning Materials and Principal Strengthening Training, 2020).
Training strengthening principals is developed by integrating the principle of freedom of learning (Merdeka Belajar) with the realization of students wellbeing through instructional leadership, which is leadership that focuses or emphasizes on learning which includes components of the curriculum, learning process, assessment (assessment of learning outcomes), assessment and development of teachers, excellent services in learning, and the development of learning communities inside and outside the school.

Motivation to leave
Several studies show that many work-related factors contribute to employees' motivation to quit their jobs or influence their turnover intentions (Chen & Scannapieco, 2010;Hayes et al., 2006;Hong, 2010). In a study of teachers in Hong Kong, a positive relationship was found between burnout and the intention to leave the profession (Leung &Lee, 2006). Contradictory results have been found regarding job satisfaction (Tzeng, 2002). Some studies have investigated the relationship between self-efficacy and motivation to quit. This research shows that self-efficacy can serve as a buffer against thoughts or intentions to leave work or turnover (Chen & Scannapieco, 2010;McNatt & Judge, 2008;Niu, 2010).
It is important to note that the motivation to leave or intention turnover is not the same as actual quitting behavior. According to LeCompte and Dworkin (1991), many people experience burnout and dissatisfaction but never leave their jobs. Previous studies have revealed a weak link between the desire to leave and actual quitting behavior. Belief in alternative roles is often a necessary precursor to actual quitting behavior because many people have invested heavily in their careers (Dworkin, 1987).

Self-efficacy for instructional leadership
Self-efficacy is a critical element in Bandura's social cognitive theory (Bandura, 1978(Bandura, , 1986(Bandura, , 1997. The social cognitive theory emphasizes human agency development and practice-the idea that people can exercise influence over what they do, and people are seen as self-organizing, proactive, self-reflective, self-management, and involved in their development. People can influence their actions and have the skills to control their mindsets and emotions. What they think, believe, and feel creates a code of conduct. Perceptions of reality and behavior are influenced by the controls and influences they experience during their lives. Individuals are considered products and manufacturers of their social environment (Bandura, 1986).
Link Enforcement Awareness Development

Figure 1. School principal strengthening training intervention model
The principal's self-efficacy can be defined as a type of leadership self-efficacy related to a certain level of confidence in one's knowledge, skills, and abilities in relation to leading others (Hannah et al., 2008). In this study, the principal's self-efficacy was operationally defined as the principal's assessment of the principal's ability to plan, organize, structure a particular action, and carry out tasks and handle their relationships with the people and schools they lead in their environment (Tschannen-Moran & Gareis, 2004). This definition covers all responsibilities of the principal. Thus, in accordance with the general conceptualization of the construct, the headmaster's self-efficacy is conceptualized as a multidimensional construct. According to Skaalvik (2020a), the concept of self-efficacy for instructional leadership as a multidimensional construction emphasizes teaching. It provides helpful feedback to teachers as an essential dimension of instructional leadership not captured by the instructional leadership subscale, the Principal Sense of Efficacy Scale (Tschannen-Moran & Gareis, 2004). Similarly, Principal Efficacy Beliefs for Instructional Leadership (PEBIL; Goddard et al., 2020). According to our analysis, the previous scale did not capture the goals and vision of the school and created a positive and safe learning environment which is an essential dimension in instructional leadership.
Based on previous research (Coldren & Spillane, 2007;Hallinger, 2010;V. Robinson, 2011;Ylimaki & Jacobsen, 2013;Osborne-Lampkin et al., 2015;Mitchell et al., 2015;Brandmo & Aas, 2017), this study distinguishes between five dimensions of self-efficacy of principals for instructional leadership: (a) developing educational goals and visions, (b) creating a collective culture among staff, (c) motivating teachers, (d) classroom observation and teacher guidance, and (e) creating a positive and safe learning environment for students (Skaalvik, 2020a). These five dimensions do not require the principal to perform all these tasks personally, but they can also complete them by delegating tasks.
The development of goals and visions is emphasized as an essential aspect of instructional leadership by some researchers (Hallinger, 2010;Hallinger & Murphy, 1985;V. Robinson, 2011). The goal and vision are to communicate the school's mission or purpose and clarify what should be the focus of attention (Hallinger & Murphy, 1985). Creating a collective culture also involves purpose and vision. Goals and visions need to be developed and communicated. It is also essential to work actively with teachers to create a collective culture where teachers internalize common goals and visions. The third dimension is to encourage, motivate, and engage teachers, take part in school development, and day jobs with students. In many schools, teachers claim that principals or school leaders are rarely present and that their efforts are not recognized or appreciated because they do not motivate teachers (Skaalvik & Skaalvik, 2015). The core dimension of instructional leadership that directly affects the learning environment is the observation and guidance of teachers. The fifth dimension, creating a positive and safe learning environment for students, is essential to ensure optimal learning and student wellbeing. The study did not focus on how principals practice five dimensions. Instead, the study focused on the principal's confidence to put the underlying principles of five dimensions of instructional leadership into practice after obtaining principal strengthening training.
Over the past two decades, many studies have shown a positive relationship between selfefficacy and principal instructional leadership (Brama, 2004;Domsch, 2009;Miller, (2015).High; Nir et al., 2006;Niyazi, 2013;Ross &Gray, 2006;Tschannen-Moran & Gareis, 2004;Hallinger et al., 2018). Research related to the principal's self-efficacy for instructional leadership is still very little, never even done in Indonesia. The most recent study was conducted in Norway (R.A. Federici & Skaalvik, 2011;Skaalvik, 2020a), Iran , and Midwestern (Goddard et al., 2020). In addition, many studies use PIMRS to test self-efficacy and instructional leadership (Atherley, 2013;Fancera & Bliss, 2011;Hallinger et al., 2018;Horton, (2013).The ;Miller, (2015). High), and a researcher who uses Norwegian self-efficacy for instructional leadership scale (NSEILS; Skaalvik, 2020a). In addition to the instructional leadership subscale of the Principal Sense of Efficacy Scale and Principal Efficacy Beliefs for Instructional Leadership (PEBIL) that do not capture the dimension of instructional leadership in its entirety. Therefore, this study uses the Norwegian Self-Efficacy for Instructional Leadership Scale (NSEILS). This instrument has never been used in samples in Indonesia.
A literature study confirms a positive relationship between self-efficacy and work engagement (Halbesleben, 2010;Prieto, 2009;Sweetman & Luthans, 2010). R.A. Federici and Skaalvik (2011) stated that the principal's self-efficacy was positively related to work engagement. Likewise, Skaalvik (2020a) observed a positive relationship between self-efficacy for instructional leadership and work engagement. In addition, the results of other studies state that the principal's self-efficacy is positively related to job satisfaction. (Evers et al., 2002;Friedman, 1995Friedman, , 2002R.A. Federici & Skaalvik, 2011;Skaalvik & Skaalvik, 2009. Skaalvik (2020a) states that self-efficacy for instructional leadership is negatively related to the motivation to leave a position as principal. Similar relationships were found in other studies (Chen & Scannapieco, 2010) and have shown that the headmaster's self-efficacy has a preventive influence on motivation to quit work. After discussing that, there was no research focusing on the relationship between self-efficacy for instructional leadership, work engagement, and job satisfaction in the same study, nor did any research investigate the relationship between these variables in the Indonesian context. With considering the literature outlined here, the following hypothesis is proposed: H 1 : There is a positive relationship between principal self-efficacy for instructional leadership and work engagement H 2 : There is a positive relationship between principal self-efficacy for instructional leadership and job satisfaction.
H 3 : There is a negative relationship between principal self-efficacy for instructional leadership and motivation to leave.

Work engagement
Work engagement is often defined as " . . . a positive, fulfilling work-related state of mind that is characterized by vigor, dedication, and absorption" (Bakker et al., 2011, p. 5). According to this definition, work engagement refers to feelings of fulfillment, persistent affective-cognitive conditions, and not focusing on particular objects, events, individuals, or behaviors W.B. Schaufeli & Bakker, 2010). Thus, work engagement is a motivational construction that positively relates to work efforts and performance (Bakker & Bal, 2010;Demerouti & Cropanzano, 2010). People who experience work engagement have an energetic and effective feeling of connecting with their work activities and see themselves able to face the demands of their work well (W. B. Schaufeli & Bakker, 2004).
In a study of teachers in Finland, Hakanen et al. (2006) found that work engagement positively relates to organizational commitments. As a result, it is reasonable to expect that principals who have work engagements are more motivated to continue working as school leaders. Demerouti et al. (2001) also found that work engagement was negatively associated with health problems such as psychosomatic symptoms. However, in general, research evidence fails to find a strong relationship between work engagement and health psychology indicators (Bakker et al., 2011). Previous research has also revealed a positive relationship between work engagement and self-efficacy (Halbesleben, 2010;Sweetman & Luthans, 2010). Skaalvik (2020a) shows that work engagement has a negative relationship with the motivation to leave the position as principal. Based on this discussion, the following is hypothesized: H 4 : There is a negative relationship between work engagement and the motivation to leave

Job satisfaction
Job satisfaction is generally defined as a pleasant or positive emotional state resulting from a person's job assessment or work experience (Weiss, 2002;Zembylas & Papanastasiou, 2006). Several similar definitions have been proposed by other researchers (Cranny et al., 1992;Schultz, 1982), who agreed that job satisfaction is considered an affective orientation to one's work (Newby, 1999).
Job satisfaction has also shown relationships with various organizational and individual outcomes. For example, Fried et al. (2008), in a meta-analysis of studies of various occupations, found that job satisfaction is negatively related to a person's intention to leave work. In line with that, Skaalvik (2020b) and R.A. Federici and Skaalvik (2011) stated that job satisfaction has a negative relationship with the motivation to leave the position as principal. However, there are no studies investigating self-efficacy for instructional leadership, job satisfaction, and motivation to leave a position as principal in a simultaneous study, and have not been found in Indonesian literature. In addition, there were no studies that tested the relationship of job satisfaction with the motivation to leave a position as a principal. Thus, the hypotheses that can be proposed are: H 5 : There is a negative relationship between job satisfaction and motivation to leave.
Based on theoretical frameworks and hypotheses, the following conceptual frameworks can be proposed to highlight the relationships between constructs in this study. More details, can be seen in Figure 2.

Method
This study used a cross-sectional survey design to examine the conceptual model proposed in figure 2. In this section, we present data collection and sample procedures, measurements, and data analysis methods.

Sample and data collection
The population of this study is junior high school (public) in Sikka District, Timor Tengah Utara (TTU), Alor, Timor Tengah Selatan (TTS), and Kupang City in East Nusa Tenggara Province, Indonesia. The selected school is a school where the principal has attended the training of the principal and has a Certificate of Completion of Education and Training (CCET). In Sikka district there are 21 principals who already have CCET and TTU 27 principals, Alor 23 principals, TTS 39 principals, and Kupang City 18 principals. Using a saturated sampling procedure, where all members of the population are used as samples. A total of 128 schools, questionnaires were distributed to 128 principals, 125 respondents participated in the study, and completed the filling of questionnaires, resulting in an effective response rate of 97.65%. A total of 125 responses were received after data collection in two months (May-April 2021). Of the 125 responses, 3 could not be used because of incomplete or outliers. The data for this study was collected using a questionnaire survey. Online questionnaires (google forms) are distributed via WhatsApp to all principals.
The sample consisted of 39.2% female principals and 60.8% male principals. Eight per cent of the participants were between 30 and 35 years of age, 18% were between 36 and 40, and 74% were between 41 and 56. The average teaching experience before becoming a principal was 10-26 years, the average number of years of experience as a principal was 1-5 years, and the average number of years of experience at the school was 21-40 years. Meanwhile, concerning education, 80.8% of principals are Bachelor. The sample consisted of principals from junior high schools.

Measurement
Self-efficacy for instructional leadership is measured using five items of Norwegian self-efficacy for instructional leadership scale (NSEILS) consisting of five sub-scales: (a) development of goals and visions for schools, (b) collective cultural development, (c) motivating teachers, (d) classroom observation and teacher guidance, and (e) creating a positive and safe learning environment for students. Responses are given on a 5-point scale from "Not sure at all" (1) to "Absolutely certain" (5). Cronbach's alpha for the total scale is .88. Alpha to sub-scale is for goal development (.90), collective culture (.84), motivating teachers (.72), observing and guiding learning (.76), and learning environment (.82;Skaalvik, 2020a). Meanwhile, engagement is measured by a five-item version of the Utrecht Work Engagement Scale . Responses are given on a 5-point scale from "Never" (1) to "Every day" (5). Cronbach's alpha for this scale is .89.
The principal's job satisfaction was measured by five scale items developed for this study (R. A. Federici & Skaalvik, 2012). This scale focuses on the principal's global feelings towards work. Principals are asked to rate statements that indicate their level of job satisfaction. Cronbach's alpha for this scale is .91. The items are: "I'm inspired by my work," "I really enjoy being a principal," "As a principal, I'm in my element," "I love being a principal," and "When I wake up in the morning I look forward to going to work." Responses are given on a 5-point scale ranging from "Not at all" (1) to "of course" (5).
The motivation to leave the position as principal was measured by three modified scale items, which have initially been developed for teacher research (Skaalvik & Skaalvik, 2011). Responses were given on this scale of 5 points from "Strongly disagree" (1) to "strongly agree" (5). Cronbach's alpha for this scale is .92.

Data analysis
The data was analyzed using PLS path modeling using SmartPLS 3.3.3. The primary considerations of choosing this technique are 1). SEM is a superior feature of regression in terms of simultaneous estimation of all parameters in the model (Iacobucci et al., 2007); 2). PLS-SEM is able to provide complete information on the extent to which the model is supported by data, such as measuring the goodness of fit and predictive relevance (Hair et al., 2017;Latan et al., 2018). This study uses a consistent estimator through the application of consistent PLS (PLSc) due to the confirmation nature of this study. Convergent validity, composite reliability, and discriminant validity are used to assess the validity of the construct. The hypothesis was tested with structural equation modeling using the consistent PLS bootstrapping inside SmartPLS 3.3.3.

Respondent profile
The study involved a total of 125 principals in the junior high school (public) of Sikka district, Timor Tengah Utara (TTU), Alor, Timor Tengah Selatan (TTS), and Kupang City in East Nusa Tenggara Province, Indonesia. Respondent profiles include age, gender, teaching experience, years of experience at the school, years of experience as a principal, and education. The respondent's demographic profile is shown in Table 1.

Convergent validity, composite reliability, and discriminant validity
Convergent validity is assessed by looking at outer loadings and average variance extracted (AVE). The outer loading for each item should be higher than 0.7, and the AVE of each construct should be above 0.5. However, outer loading 0.5 is still acceptable as long as the AVE for certain constructs meets the requirements of 0.7 (Hair et al., 2017). AVE less than 0.5 indicates that the item failed to describe most variants of the construct. In addition to convergent validity, composite reliability (CR) representing the consistency of internal indicators in measuring a construct, is also assessed. CR 0.7 shows sufficient internal consistency (Hair et al., 2017). The assessment of convergent validity and CR results is presented in Table 2.  Table 2 shows that all outer loading is greater than 0.6, while the average variance extracted (AVE) value is greater than 0.5, and the CR value is higher than 0.7. Therefore, convergent validity and construct CR are considered satisfactory.
Discriminant validity is assessed using Fornell and Larcker Criterion. Furthermore, to ensure the uniqueness of the construct, discriminant validity assessment is done by comparing the square root of the AVE value of the construct with the correlation coefficient of the construct with other constructs (Fornell & Larcker, 1981). Table 2 shows that all reflective measurement models have sufficient discriminant validity as the square root of all constructs is higher than the correlation with other constructs. In short, all constructs are different from each other can be seen in Table 3.
At this stage, a bias-corrected bootstrapping procedure is applied. According to Nawanir et al. (2020), the use of 5,000 bootstrap samples can ensure the stability of results. A one-way test with a significance level of 0.05 was set out in the study. figure 3 and Table 4 show the results of the structural model assessment.  The results showed that self-efficacy for instructional leadership was positively related to work engagement (β = 0.752, t = 17,679), job satisfaction (β = 0.734, t = 14,225), and according to the hypothesis, the relationship between self-efficacy for instructional leadership and motivation to leave a position as principal is not significant (β = 0.106, t = 1,146). So, H 1 , H 2 , and H 3 are strongly supported. Meanwhile, work engagement (β = 0.232, t = 2,042) and job satisfaction (β = 0.567, t = 5,090) were positively related to the motivation of leaving a position as principal and based on hypotheses found H 4 and H 5 were not supported. More details, can be seen in Table 4..

Discussion
The purpose of this study is to explore how self-efficacy for principal instructional leadership relates to work engagement, job satisfaction, and motivation to leave a position as principal. To achieve this goal, five hypotheses were tested. Three hypotheses are strongly supported, and two hypotheses are not supported.  The relationship between motivation leaving a position as to school and work engagement in this study was proven. The results of this study are in line with previous research (Halbesleben, 2010;Prieto, 2009;R.A. Federici & Skaalvik, 2011;Skaalvik, 2020a;Sweetman & Luthans, 2010). Principal self-efficacy for instructional leadership is predictive for high-level work engagement and predictive for motivation to leave at a low level. A possible interpretation is that principal selfefficacy for instructional leadership is high, impacting on high levels of work engagement, which in turn increases the motivation to remain in the position of principal. Instead, the principal selfefficacy for instructional leadership is low, resulting in a low level of work engagement, which increases the motivation to leave the position as principal. This interpretation highlights the importance of designing the principal's primary role in improving work engagement. An alternative understanding of Skaalvik (2020a) asserts that principals with low self-efficacy for instructional leadership ignore this dimension of school leadership. However, principals who avoid instructional leadership may also feel uncomfortable and unsuccessful, as they neglect the responsibilities for which they are expected to do so. This explanation highlights the need for principal education with a strong emphasis on instructional leadership values and the skills and competencies required by such leadership. This research proves that principals who have obtained the Certificate of Completion of Education and Training (based on instructional leadership competencies) give positive results to work engagement, job satisfaction and can prevent the principal from leaving his/her position as principal.
The findings of this study are different from previous research, which states that the leadership of principals in Indonesia has not focused on the management of the teaching and learning process (Rizkita & Supriyanto, 2020) and only focus on maintaining high visibility (Dania & Andriani, 2020). The study results showed that principal self-efficacy for instructional leadership could develop clear goals and expectations for learning, guide teachers about educational matters, ensure a learning environment where students feel safe, and motivate teachers for teaching and learning. In addition, these findings are following social cognitive theory. Social cognitive theory assumes that people are motivated to shy away from situations and activities where they have low expectations of mastery (Skaalvik, 2020a). Based on these assumptions, self-efficacy is essential for the involvement of principals in their work, especially instructional leadership compared to administrative management and the school environment (Federici &Skaalvik, 2011, p. 594-595). In addition, the efficacy of the principal is also related to the commitment of their followers to the duties of the organization. It has a positive influence on the work engagement of followers (Chemers et al., 2000).
The study explored the positive relationship between principal self-efficacy for instructional leadership and job satisfaction. These results are supported by previous research (Evers et al., 2002;Friedman, 1995Friedman, , 2002R.A. Federici & Skaalvik, 2011;Skaalvik & Skaalvik, 2009. While the relationship between job satisfaction and motivation to leave the position as principal is not supported because the results are positive. These results are also different from previous studies showing a negative relationship between job satisfaction and motivation to leave a position as principal (Fried et al., 2008;R.A. Federici & Skaalvik, 2011;Skaalvik, 2020b).
In addition, the results of this study imply that principal self-efficacy instructional leadership relates to the motivation to leave a position as principal in two ways. First, high self-efficacy can lead to higher job satisfaction, increasing motivation to continue working as a principal. At the same time, a high level of self-efficacy can reinforce the belief that a principal can succeed in another line of work and therefore increase the motivation to leave the position. Thus, this study demonstrates the importance of principal self-efficacy for instructional leadership for work engagement and job satisfaction and shows how these concepts relate to the principal's motivation to quit work. The principal's responsibility for the learning and welfare of students in the school is the principal need to develop instructional leadership competencies and self-efficacy.
This study shows that schools that have obtained headmaster training and have received the Certificate of Completion of Education and Training (CCET) have increased self-efficacy for instructional leadership, thus increasing the work engagement and job satisfaction of the principal. In the end, the principal had no intention of leaving their position as principal. In essence, the policy of the Indonesian government through the Directorate General of Teachers and Education Personnel of the Ministry of Education and Culture of the Republic of Indonesia related to the training of strengthening the principals is applicable and proven to improve the principal's instructional leadership.

Conclusion
This study showed a positive relationship between principal self-efficacy for instructional leadership, work engagement, and job satisfaction. In addition, the relationship between principal selfefficacy for instructional leadership and motivation to leave a position as principal is not significant, thus supporting this hypothesis. The relationship between work engagement and job satisfaction with the motivation to leave a position as principal was positively proven, so this hypothesis was not strongly supported in this study. This study showed that Indonesian government policies related to the training of strengthening principals succeed in improving self-efficacy for the principal's instructional leadership.
Based on the results of this study, the successful implementation of principal strengthening training in Indonesia can be adopted by policymakers from other countries because it can increase selfefficacy for instructional leadership, work engagement, job satisfaction, and prevent motivation to leave from the principal. In particular, this principal strengthening training program uses the ORPAER intervention model (observe, reflect, plan, act, evaluation, and reflect) combined from the syntax of Theory U and the ILEAD (Introduction, Link, Enforcement, Awareness, and Development) approach. This intervention model can improve the competence of principals, the principals' problem-solving skills, creative thinking, and emotional engagements. This principal strengthening training program has the power because it is mandatory to include all principals, in contrast to the Swedish national principal training programme (Jerdborg, 2021;Norberg, 2019), which is voluntary. In addition, this model considers social inclusion without distinguishing ethnicity, religion, race, class, gender, socioeconomic status, and special needs so that it can be applied in the context of other countries. This social inclusion is also applied to educators, education personnel and learners.