Does vocational passion matter? Minority case of Indonesian Christian educators

Abstract

This study investigated the role of vocational passion in the relationship between well-being, psychological distress, and job satisfaction among Indonesian Christian educators, who form part of a minority in a predominantly Muslim country. The research surveyed 2,421 active local Christian educators. We then conducted a confirmatory factor analysis and bootstrapping using SmartPLS. Indonesian Christian educators’ well-being and vocational passion have a direct positive effect on job satisfaction. Psychological distress has a direct negative effect on both job satisfaction and well-being. Notably, Indonesian Christian educators’ well-being partially mediates psychological distress and job satisfaction, indicating that educational institutions must address Indonesian Christian educators’ psychological distress rather than enhancing their well-being to improve their overall job satisfaction. Moreover, vocational passion negatively moderates the relationship between well-being and job satisfaction. This may be attributed to religious values and norms, and social expectations placed on Indonesian Christian educators. Last, vocational passion does not moderate the relationship between psychological distress and job satisfaction, indicating that, while vocational passion is essential, it may not serve as a buffer for the adverse effects of psychological distress on job satisfaction. Efforts should prioritize addressing well-being and psychological distress rather than relying solely on vocational passion to mitigate its impact.

Keywords:

• job satisfaction
• well-being
• psychological distress
• vocational passion
• education
• minority case

1. Introduction

Since March 2020, educational institutions have dealt with the impact of the government’s COVID-19 policy changes. Enforcing safety protocols, closing school buildings, and adopting distance learning were inevitable. As these changes had to be implemented in a short period, educators had insufficient time to receive proper training and adjust to the distance learning methodology or even for schools to plan and implement such changes properly (Chan et al., 2021). The situation was further exacerbated by the lack of readiness, expertise, and technology skills among educators, along with a lack of suitable facilities and infrastructure to support the implementation of distance learning. Consequently, educators have begun to question how long they can afford to continue, despite their passion for education and continued desire to support students. Chan et al. (2021) found that nearly half of the surveyed educators in the United States experienced high psychological stress during the first few months of teaching during school closures; this is supported by surveys conducted by Alberta Teachers’ Association (ATA), Ottawa-Carleton District School Board, and Canadian Teachers’ Federation (Canadian Teachers’ Federation, 2020; Janzen, 2020; Pringle, 2020). Unfortunately, even after almost two years into the COVID-19 pandemic, the situation has not changed in the region. ATA later conducted a study in January 2022 and reported that 85% of teachers still feel high levels of stress and anxiety coming into the second year of the pandemic (Edwardson, 2022). These data also apply to Indonesia; educators working in DKI Jakarta, West Java, and Banten experienced high stress and anxiety as the pandemic continued (Fauzan et al., 2022; Indra et al., 2021).

Indeed, research related to the above will contribute significantly to educational institutions in making strategic policies and providing support. However, previous research has a common denominator, namely the use of a popular sample target. People from minority ethnic groups are under-represented; without their participation, we would not uncover the existence of these disparities in the education literature or understand their underlying causes (Redwood & Gill, 2013). Therefore, this research aims to promote inclusivity and close the gap of the SAT literature by exploring Indonesian Christian educators, who are a minority in a predominantly Muslim country. Understanding the factors influencing Indonesian Christian educators’ job satisfaction in the minority group can inform the development of institutional strategic policies tailored to their needs. In addition, these findings can provide valuable knowledge for educational institutions globally, especially in multicultural environments, by highlighting its importance in shaping educator job satisfaction. In addition, Holmes et al. (2020), on behalf of The United Nations Educational, Scientific and Cultural Organization, also advocated studying the effects of the pandemic to apply this knowledge in the future. This research was conducted from December 2021 to January 2022, when there were several new changes and developments in the situation compared to previous studies. By being conducted at a time when there are new changes and developments in the pandemic, this study contributes to the existing literature by capturing the unique circumstances and challenges that educators face. This study uses educators’ job satisfaction (SAT) as the dependent variable, with two independent variables—educators’ well-being (WEB) and educators’ psychological distress (DIS)—and one moderating variable—educators’ vocational passion (PAS)—to explain SAT. Based on the background and problem identification, the following research questions were formed:

1. Does WEB or PAS among Indonesian Christian educators positively influence their SAT?

2. Does the presence of DIS negatively affect SAT or WEB in Indonesian Christian educators?

3. Does WEB mediate the impact of DIS on SAT of Indonesian Christian educators’?

4. Does PAS of Indonesian Christian educators moderate the relationship between their WEB and SAT, as well as the relationship between DIS and SAT?

This study aims to gain new knowledge about the relationship between the above variables, which will be considered by stakeholders of educational institutions when designing educator coaching strategies and developing the SAT models of Indonesian Christian educators, and to serve as reference material for further research.

2. Literature review

Figure 1 shows the proposed research model and hypotheses.

Figure 1. Proposed research model.

2.1. Educator’s job satisfaction

SAT is a central issue in human resource (HR) management within various organizations, including educational institutions. In educational institutions’ HR management strategies, assessing educator job satisfaction is crucial as a performance indicator. Different scales have been used to measure this construct. Chan et al. (2021) employed an adapted version of the Teacher Stress Inventory, Pepe et al. (2017) relied on the Teacher Job Satisfaction Scale (TJSS), and Margolis et al. (2019) developed the Riverside Life Satisfaction Scale, each focusing on different dimensions such as work environment, co-worker relations, and overall life satisfaction. The concept of job satisfaction has evolved over time. Hoppock’s seminal work in 1935 framed SAT as being influenced by psychological, physiological, and environmental circumstances (Aziri, 2011). Various theoretical frameworks have been used to investigate SAT. For example, Maslow’s hierarchy of needs has been applied to understand how job satisfaction is tied to fulfilling a range of needs, from basic to self-actualization (Noori, 2023). Herzberg’s two-factor theory categorizes job elements into “Hygiene Factors,” which can prevent dissatisfaction but do not effectively promote SAT, and “Motivators,” which have the potential to genuinely satisfy and engage employees in their work. The job characteristics model refines this further by focusing on specific job attributes like task variety and autonomy. However, these classical theories often fail to capture the multidimensional nature of SAT, as recently acknowledged in the literature. Ahrari et al. (2021) highlight that SAT is a complex construct with different meanings for different individuals. Moreover, Ntimba et al. (2021) and Adeka and Mede (2022) emphasize additional dimensions, such as compensation and interpersonal relations, which are increasingly important for educators’ SAT. Building on existing research, this study defines SAT as a subjective assessment of psychological satisfaction experienced by Indonesian Christian educators in fulfilling their teaching and learning responsibilities, including classes, classroom interactions, materialistic compensation, and relationships with colleagues, students, and educational institutions.

This study uses the Job Demands-Resources model as its theoretical framework. According to this model, DIS is seen as a job demand that can cause strain and decrease SAT. Conversely, WEB is viewed as a job resource that can protect against the negative effects of job demands and increase SAT. Additionally, PAS is considered a personal resource that may potentially moderate the relationships between these variables. This research aims to gain a broader understanding of the factors that impact SAT among educators, particularly among minority groups like Indonesian Christian educators.

2.2. Educator’s well-being

Understanding the importance of WEB in HR management is crucial, as it directly affects SAT (Adeka & Mede, 2022; Dreer, 2021; Yee et al., 2022). WEB is a complex concept with varying definitions in the literature. Huppert (2009) defines WEB as a combination of feeling good and functioning well, incorporating positive emotions, control over life, and meaningful relationships. Maddux (2018) describes it as a construct that considers material conditions, subjective experiences, and emotions. Ruggeri et al. (2020) and Kundi et al. (2021) describe WEB in terms of both hedonic experiences of pleasure and eudaimonic experiences of meaning and purpose. Adeka and Mede (2022) suggest that WEB arises from the harmony between environmental factors and personal needs and expectations. Previous empirical research has utilized various methods to operationalize WEB. Schat et al. (2005) employed a modified version of the Physical Health Questionnaire (PHQ), which includes questions such as “How often have you had difficulty falling asleep at night?” and “How often have you had to be careful about what you eat to avoid stomach upsets?” Kobau et al. (2010) took a different approach and adapted the Autonomy, Competence & Relatedness Scale. Their questions included “I get along well with the people I interact with.” In examining various definitions and operationalizations of WEB, some researchers, like Dreer (2021), have turned to Seligman’s model, which includes positive emotion, engagement, relationships, meaning, and achievement (PERMA) as domains of WEB. Dreer (2021) found that only positive emotions were significantly correlated with SAT among the PERMA components. Likewise, Yee et al. (2022) focused on investigating the psychological dimensions of WEB. Based on this related research, this study defines WEB as a measure of educators’ subjective assessment of their welfare in terms of health, economy, and social relations, as influenced by their work experience. Aligned with this definition, the PERMA model provides a comprehensive approach to understanding WEB, encompassing emotional health, engagement in tasks, relationship quality, the pursuit of meaning, and the sense of achievement. These aspects align with the broader evaluation of educators’ welfare in terms of health, economy, and social relations in this study, providing a holistic framework for the research. Due to the variability in how WEB is conceptualized and measured, this research employs a comprehensive and systematic approach. Specifically, this study uses the PERMA model, as well as other relevant frameworks, to measure WEB as comprehensively as possible.

Several previous studies have empirically supported a direct positive relationship between WEB and SAT (Adeka & Mede, 2022; Dreer, 2021; Yee et al., 2022). Adeka and Mede (2022) used a mixed method approach to determine that English-as-a-Foreign-Language teachers in Istanbul have a positive impact on SAT through WEB. Similarly, Dreer (2021) found that German school teachers’ WEB also positively affected SAT. Yee et al. (2022) reported a significant correlation between preschool teachers in Perak’s WEB have and SAT (r = 0.39, p < 0.01). However, in contrast, the meta-analytic study conducted by Bowling et al. (2010) found a two-way relationship between WEB and SAT. Ultimately, consistent with other studies, the effect of WEB on SAT is stronger than the reverse. Thus, the tested hypothesis is as follows:

H1a:

WEB has a positive effect on SAT.

2.3. Educator’s psychological distress

DIS is a major concern in HR management, especially in the field of education, where it impacts educators’ WEB (Allahyari et al., 2018; Tran et al., 2022) and SAT (Allahyari et al., 2018; Anastasiou & Papakonstantinou, 2014; Bonsaksen et al., 2021; Chan et al., 2021; Friganović et al., 2019; Viertiö et al., 2021; Xie et al., 2021; Zhang et al., 2021). Prior research, such as Dalrymple et al. (2013), used modified items from the Clinically Useful Social Anxiety Disorder Outcome Scale to measure DIS, such as “I was extremely afraid of social situations.” Similarly, Baker et al. (2019) used the Anxiety Symptoms Questionnaire (ASQ) with items measuring anxiety and nervousness, while Sarafis et al. (2016) used the Expanded Nursing Stress Scale to assess work stressors like conflicts with colleagues. Additionally, Chan et al. (2021) employed the Maslach Burnout Inventory Educators Survey (MBI) with statements like “I feel emotionally drained from my work.” DIS varies among individuals and is influenced by teaching-related factors such as frustration, emotional disturbance, burnout, and anxiety (Anastasiou & Papakonstantinou, 2014; Horwitz, 2007; Kokkinos, 2006; Maslach et al., 2001; Viertiö et al., 2021; Wheaton, 2007). According to Horwitz (2007), DIS is a result of exposure to stressful events that harm physical or mental health, leading to emotional turmoil due to ineffective coping. Marchand and Durand (2011) define DIS as a mental health outcome characterized by psychophysiological and behavioral symptoms. Goodwin et al. (2013) equate DIS with poor mental health, while Drapeau et al. (2012) consider it a normal emotional reaction to stress. Anastasiou and Papakonstantinou (2014) and Sarafis et al. (2016) define DIS in terms of negative emotions and work-related factors. Bonsaksen et al. (2021) highlight that DIS severity varies depending on the work context or type. To capture the complexity of DIS, this study adopts the transactional theory of stress and coping (TTSC) as a guiding framework. Building on previous literature, this study defines DIS as an assessment of the psychological stress experienced in work, including frustration, emotional disturbance, fatigue, and perceived anxiety.

Negative feelings arising from psychological stress, such as frustration, emotional disturbance, fatigue, and anxiety, will make it difficult for educators to achieve WEB and SAT. When educators consistently encounter challenges without practical solutions, it leads to frustration, an increase in DIS, and hinders their sense of accomplishment; consequently, this results in a decrease in WEB and SAT (Agyapong et al., 2022). This also applies to emotional disturbance from stressful situations, such as managing challenging students or classroom disruptions, which further add to the challenge of maintaining WEB and SAT (Agyapong et al., 2022). The demanding nature of the teaching profession, such as extended hours, heavy workloads, and multiple responsibilities contributes to fatigue, also making it difficult for educators to find joy and fulfillment in their work, ultimately diminishing their WEB and SAT (Agyapong et al., 2022). In line with previous descriptions, the pressure to meet educational standards, achieve positive student outcomes, and meet parental expectations can also induce significant DIS, eroding educators’ confidence and further impacting their WEB and SAT.

This is supported by Anastasiou and Papakonstantinou (2014)’s research where DIS was observed as a crucial determinant of WEB and SAT. Previous empirical studies also support a negative relationship between DIS and SAT (Allahyari et al., 2018; Anastasiou & Papakonstantinou, 2014; Bonsaksen et al., 2021; Chan et al., 2021; Friganović et al., 2019; Viertiö et al., 2021; Xie et al., 2021; Zhang et al., 2021) and DIS and WEB (Allahyari et al., 2018; Tran et al., 2022) directly. Anastasiou and Papakonstantinou (2014) found that DIS negatively affects SAT. In Epirus, younger secondary education teachers, as well as female secondary education teachers, reported higher DIS compared to their older and male counterparts, respectively (Anastasiou & Papakonstantinou, 2014). Friganović et al. (2019) literature review found that emotional disturbance has the most significant impact on SAT as a subscale of DIS. However, there was no indirect negative relationship between DIS and WEB-mediated SAT. Therefore, this study examines the negative indirect relationship of DIS on SAT mediated by WEB. Based on this description, the formulation of the tested hypotheses is as follows:

H2a:

DIS has a negative effect on SAT.

H2b:

DIS has a negative effect on WEB.

H3:

WEB mediates DIS and SAT.

2.4. Educators’ vocational passion

The PAS factor is originally not a part of any SAT model. However, this study examines PAS because of the effect it entails. PAS has mainly been studied concerning sports and leisure activities, but it also lends itself to work contexts like workaholism and organizational commitment (Spehar et al., 2016). This indicates that PAS is a valuable addition to the work engagement and organizational literature, as reflected by its ubiquity in popular and empirical discourse. Nevertheless, the latest studies need more scientific consensus on the definition of PAS relevant to workers across various vocations. In the early 2000s, researchers defined PAS as a purely affective experience (Chen et al., 2020). Vallerand and Houlfort (2003) offered the first definition of PAS that recognized components of this experience beyond mere positive affect: a strong inclination toward an activity that people like, that they find essential, and in which they invest time and energy. According to Chen et al. (2020), PAS means strongly identifying with a line of work that one feels motivated to engage in and derives a positive effect when performed (Chen et al., 2020). In operationalizing this construct, prior empirical research (Chen et al., 2020) developed the Work Passion Scale, which includes questions such as “How much do you love doing your work?” and “How often do you feel positive about your work?” This scale is designed to assess aspects of work passion such as emotional attachment, frequency of positive feelings, and the centrality of work to one’s identity. Based on the aforementioned description, this study defines PAS as a measure of the Indonesian Christian educator’s subjective assessment of how much they love the world of education, and whether they possess the desire to keep moving forward despite unfavorable circumstances.

Carbonneau et al. (2008) study among educators found that PAS is positively correlated with SAT. Spehar et al. (2016) also found that PAS has a direct positive effect on SAT. In line with previous research, Pathak and Srivastava (2020) also found that PAS positively affected SAT both directly and indirectly. However, there is currently a lack of existing literature exploring the moderating effects of the variable PAS on the relationships between WEB and SAT, as well as DIS and SAT. It is expected that PAS should positively moderate the relationship between both WEB and SAT, and DIS and SAT. PAS is often associated with increased motivation, which is a source of inspiration and drive (Chen et al., 2020). This enables educators to find WEB and SAT in their work, even in challenging circumstances. Furthermore, passion can also act as a buffer against negative effects such as DIS, enabling them to maintain SAT despite the presence of stressors. Therefore, this study examines the moderating effect of PAS on the relationship between WEB and SAT, and between DIS and SAT. Based on this description, the formulation of the tested hypotheses is as follows:

H1b:

PAS has a positive effect on SAT.

H4a:

PAS moderates the relationship between WEB and SAT.

H4b:

PAS moderates the relationship between DIS and SAT.

3. Research method

This study uses a descriptive quantitative approach with a cross-sectional method because data were collected once from each sample (Malhotra, 2019). The data collected were obtained directly from the primary source through a questionnaire, which was then processed using SmartPLS 4, and analyzed.

A purposive sampling technique was used, where the researcher selects a non-probability sample based on population characteristics and research objectives, relying on subjective judgment, a mixture of hunches, and prior knowledge to conduct the research survey (Sukumar et al., 2016). Although there is a standard rule for measuring minimum target samples, such as calculating 10 times the indicator (Hair et al., 2011; Kock & Hadaya, 2018), Hair et al. (2017), in another opus, suggested a different calculation method using Cohen’s power primer, which is used in this research. A link to the questionnaire was shared with 2,800 target samples, yielding 2,443 responses. Later, a careful data-cleaning process was conducted, identifying and excluding a small number of perfunctory responses by utilizing reversed questions to detect inconsistent responses. A final dataset of 2,421 respondents was obtained, reflecting an 86% response rate. Ethical review and approval were not required for the study on human participants in accordance with the local legislation and institutional requirements. Informed consent was obtained from all individual respondents involved in the research.

The questionnaire was adapted from several sources, translated backward using two independent translators (Tsang et al., 2017), and distributed to respondents. The questionnaire for measuring WEB was adapted from Schat et al. (2005) and Kobau et al. (2010) with sample items such as “I reach my daily nutritional intake goal regularly” and “I have enough time to spend with people I love.” The questionnaire for measuring DIS was adapted from Dalrymple et al. (2013) and Baker et al. (2019) with sample items such as “When I work, I always feel like I am at the end of a rope” and “I do not care what happens to my students.” The questionnaire for measuring PAS was adapted from Chen et al. (2020) with sample items such as “Becoming an educator is my passion” and “I will continue to be an educator even under unfavorable conditions.” Last, the questionnaire for measuring SAT was adapted from Pepe et al. (2017) and Margolis et al. (2019) with sample items such as “I am satisfied with all my responsibilities as an educator” and “I am happy with the teaching-learning interactions that I experience.” Table 1 presents the complete operational definitions for each variable.

Table 1. Operational definition

Indicator	Code	Source
I get enough quality sleep every night.	WEB-01	Adapted from Schat et al. (2005); Kobau et al. (2010)
I reach my daily nutritional intakes goal regularly.	WEB-02
I DO NOT have enough materialistic compensation for my job. (R)	WEB-03
I have enough time to spend with people I love.	WEB-04
Running an unpleasant task frustrates me.	DIS-01	Adapted from Dalrymple et al. (2013); Baker et al. (2019)
Adjusting to unfavorable working conditions frustrates me.	DIS-02
Dealing with student problems frustrates me.	DIS-03
I am NOT frustrated by having to endure unpleasant working conditions. (R)	DIS-04
Other people’s perception of my work frustrates me.	DIS-05
When I work, I always feel like I am at the end of a rope.	DIS-06
I feel HAPPY after working with my students. (R)	DIS-07
This job bothers me emotionally.	DIS-08
I feel like I have no energy for work when I wake up in the morning.	DIS-09
I do not care what happens to my students.	DIS-12
I want to remain an educator until the end of my life.	PAS-01	Adapted from Chen et al. (2020)
I did NOT enjoy my time as an educator. (R)	PAS-02
Becoming an educator is my passion.	PAS-03
I will continue to be an educator even under unfavorable conditions.	PAS-04
I believe that by becoming an educator, I will achieve my goals.	PAS-05
I am satisfied with all my responsibilities as an educator.	SAT-01	Adapted from Pepe et al. (2017); Margolis et al. (2019)
My co-workers are NOT very supportive of me. (R)	SAT-03
My boss really cares about my needs.	SAT-04
I am happy with the teaching-learning interactions that I experience.	SAT-06

Many similar studies conducted in Indonesia only used popular samples. However, this study examines active educators; That is, Christian Indonesian educators who have been actively teaching since 2019. This group represents a minority within the broader population and can be further classified based on their social determinants, such as gender, education level, work domicile, school type, age, work experience, categories of educator, employment status, marital status, number of household members, household status, and monthly expense. The data analysis technique used in this research was the structural equation model (SEM) with the partial least square (PLS) method. In addition to obtaining an evaluation of the measurement and structural models (Hair et al., 2017), PLS-SEM was used because the research model is complex, uses a pathway model, and has latent variables (Hair et al., 2021; Sarstedt et al., 2021).

Based on several previous studies, the five-point Likert scale is more widely understood by respondents (Bouranta et al., 2009) and enables them to express their views (Marton-Williams, 1986). It enhances response rates and quality while reducing respondents’ frustration levels (Buttle, 1996). With an odd and minimal number of points, this scale includes a neutral response option (Cox, 1980) and facilitates researchers to create a complete list of scale descriptors (Dawes, 2008). The five-point Likert scale also demonstrates higher reliability and validity (McKelvie, 1978). Moreover, its common usage enables comparison with other studies (Saleh & Ryan, 1991). Thus, we adapted the operational definition from previous research to fit the context and scale required for this study (Table 1). Based on this description, the scale used in this study was a five-point Likert scale with the following choices:

1. = Strongly Disagree (SD)

2. = Disagree (D)

3. = Neutral (N)

4. = Agree (A)

5. = Strongly Agree (SA)

4. Results

4.1. Descriptive statistics

In this study, the sample (Table 2) mostly comprised the following: female (75.92%), worked in West Java (24.25%), lived in private residences (46.30%), had households with three members (42.21%), held a bachelor’s degree (81.12%), belonged to the early working age group (35.60%), worked at Christian private schools (87.32%), had 10–14 years of experience (21.81%), were categorized as primary educators (36.84%), held permanent employment (61.17%), were married (63.65%), and had monthly expenses under 5 million IDR (61.30%). Table 3 presents the descriptive statistics of the variables under investigation. These variables exhibited varying means, medians, and standard deviations.

Table 2. Demographic characteristics of respondents

Characteristic (Class)	Frequency	%	Characteristic (Class)	Frequency	%
Gender			Education Level
Male	583	24.08	Trade School	46	1.90
Female	1,838	75.92	Community College	65	2.69
Work Domicile			Bachelor’s Degree	1,964	81.12
Aceh	7	0.29	Master’s Degree	339	14.00
Bali	52	2.15	Doctoral Degree	7	0.29
Bangka Belitung Islands	7	0.29	Age Group
Banten	104	4.30	15–24 (Young Age)	51	2.11
Bengkulu	2	0.08	25–34 (Early Working Age)	862	35.60
Central Java	113	4.67	35–44 (Middle Age)	812	33.54
Central Kalimantan	1	0.04	45–54 (Pre-Retirement Age)	526	21.73
Central Sulawesi	11	0.45	55–64 (Retirement Age)	166	6.86
DI Yogyakarta	29	1.20	≥65 (Post-retirement Age)	4	0.16
DKI Jakarta	249	10.29	School Type
East Java	338	13.96	Christian Private School	2,114	87.32
East Kalimantan	24	0.99	Non-Christian Private School	243	10.04
East Nusa Tenggara	26	1.07	Public School	64	2.64
Gorontalo	1	0.04	Work Experience
Jambi	36	1.49	2–4 Years	406	16.77
Lampung	26	1.07	5–6 Years	278	11.48
Maluku	21	0.87	7–9 Years	353	14.58
North Kalimantan	11	0.45	10–14 Years	528	21.81
North Maluku	16	0.66	15–19 Years	391	16.15
North Sulawesi	109	4.50	20–29 Years	313	12.93
North Sumatera	174	7.19	≥30 Years	152	6.28
Papua	94	3.88	Educator Categories
Riau	79	3.26	Nursery	354	14.62
Riau Islands	62	2.56	Primary	892	36.84
Southeast Sulawesi	14	0.58	Lower Secondary	407	16.81
South Kalimantan	6	0.25	Upper Secondary	403	16.65
South Sulawesi	52	2.15	Multi-Level	359	14.83
South Sumatera	18	0.74	Special Needs	6	0.25
West Java	587	24.25	Employment Status
West Kalimantan	108	4.46	Permanent	1,481	61.17
West Nusa Tenggara	5	0.21	Honorary	554	22.88
West Papua	5	0.21	Non-Permanent/Contract	49	2.03
West Sulawesi	1	0.04	Administrator	93	3.84
West Sumatera	33	1.36	Head of School	244	10.08
Household Status			Marital Status
Private Residence	1,121	46.30	Single	813	33.58
Rented Residence	666	27.51	Married	1,541	63.65
Official Residence	104	4.30	Widowed	45	1.86
Shared Living	530	21.89	Divorced	22	0.91
Household Member			Monthly Expense
1 Person/Living Alone	234	9.67	<5 Million IDR	1,484	61.30
2 People	318	13.13	5–10 Million IDR	801	33.09
3 People	1,022	42.21	10–15 Million IDR	94	3.88
4 People	641	26.48	>15 Million IDR	42	1.73
≥5 People	206	8.51

Table 3. Descriptive statistics

	Mean	Med	Min	Max	SD
(1) WEB	3.855	4.000	1.000	5.000	1.032
(2) DIS	2.092	2.000	1.000	5.000	1.047
(3) PAS	4.521	5.000	1.000	5.000	0.677
(4) SAT	4.119	4.000	1.000	5.000	0.865

4.2. Evaluation of reflective measurement model

The reflective outer model determines each indicator’s relationship with its latent variables. Convergent and discriminant validity testing is a measurement tool used to determine whether a data collection method is valid.

Convergent validity testing is conducted to determine whether the research concept is appropriate and is met if the values obtained from different instruments used to measure the same construct have a high correlation. When the indicator’s loading factor is lower than 0.7, Hair et al. (2019) suggest the removal of the indicator. If removing the indicator causes a decrease in the average variance extracted (AVE) value, then it is suggested to keep the indicators if the loading factor of the indicator remains above 0.6. Conversely, if the indicator’s loading factor is greater than 0.7 (Hair et al., 2017, 2019) or a minimum of 0.6 (Hair et al., 2010, 2019), and the AVE value of each construct is greater than 0.5 (Hair et al., 2017), then the convergent validity is met (Table 4).

Table 4. Results of convergent validity

		Loading Factor
	AVE	WEB	DIS	PAS	SAT
WEB	0.606
WEB-01		0.756
WEB-02		0.840
WEB-03		0.795
WEB-04		0.717
DIS	0.522
DIS-01			0.604
DIS-02			0.660
DIS-03			0.817
DIS-04			0.767
DIS-05			0.750
DIS-06			0.764
DIS-07			0.770
DIS-08			0.752
DIS-09			0.606
DIS-12			0.701
PAS	0.626
PAS-01				0.720
PAS-02				0.779
PAS-03				0.797
PAS-04				0.833
PAS-05				0.822
SAT	0.527
SAT-01					0.737
SAT-03					0.733
SAT-04					0.674
SAT-06					0.758

Discriminant validity testing is also conducted to determine whether the research concept is appropriate. Discriminant validity can be seen from the heterotrait-monotrait ratio of correlations (HTMT) and Fornell—Larcker values (Hair et al., 2017). When the HTMT value is smaller than 0.9, and the Fornell—Larcker diagonal value is greater than the correlation value of each construct, then the discriminant validity is also fulfilled (Table 5).

Table 5. Results of discriminant validity

	HTMT				Fornell—Larcker
	1	2	3	4	1	2	3	4
(1) WEB					0.778
(2) DIS	0.422				−0.361	0.722
(3) PAS	0.291	0.347			0.241	−0.310	0.791
(4) SAT	0.558	0.504	0.491		0.434	−0.423	0.394	0.726

Reliability analysis is a test used to determine how consistently an instrument can measure what is being measured (Hair et al., 2019). When the value of Cronbach’s alpha and composite reliability is greater than 0.7 (Hair et al., 2019), then the reliability is fulfilled (Table 6).

Table 6. Results of reliability

		Composite Reliability
	Cronbach’s Alpha	Rho_A	Rho_C
(1) WEB	0.782	0.787	0.860
(2) DIS	0.897	0.907	0.915
(3) PAS	0.850	0.858	0.893
(4) SAT	0.707	0.719	0.816

When the value of the inner variance inflation factor is smaller than 3.0 (Hair et al., 2019), it is concluded that there is no collinearity problem among the constructs (Table 7).

Table 7. Inner variance inflation factor

	Endogenous
	WEB	SAT
(1) WEB		1.180
(2) DIS	1.000	1.225
(3) PAS		1.221
(4) SAT
(5) PAS x WEB		1.264
(6) PAS x DIS		1.312

4.3. Evaluation of goodness-of-fit (GoF)

Model fit testing is conducted to determine how well the sample data fit the normal distribution of the population. The evaluation of the fit of the model can be seen from the results of the univariate normality test and the model fit test.

A univariate normality test can be conducted by looking at the value of skewness and kurtosis. When viewed from the skewness value, which is still between −2.0 and 2.0, and the kurtosis value, which is still between −7.0 and 7.0 (Byrne, 2010; Hair et al., 2010), it can be said that the residual univariate regression is normally distributed (Table 8). Meanwhile, the GoF test was conducted by looking at the standardized root mean square residual (SRMR), squared Euclidean distance (d_ULS), the geodesic distance (d_G), and chi-square. When the SRMR value is less than 0.080 (Hu & Bentler, 1998), the model can be considered to meet the fit criteria (Table 8). By looking at the value of d_ULS and d_G, which are greater than the significance level set in this research (Dijkstra & Henseler, 2015), the suitability of the model can be determined. As the chi-square value has a reasonably large nominal, it can be said that while this model is considerably complex, it meets all indications of GoF (Table 8).

Table 8. Goodness-of-fit parameters

	Normality			Model Fit
	Skew	Kurt		Saturated	Estimated
1. WEB	−0.774	0.011	SRMR	0.057	0.060
2. DIS	0.914	0.258	d_ULS	0.884	0.996
3. PAS	−1.537	3.034	d_G	0.215	0.217
4. SAT	−0.971	1.018	Chi-Square	3,115.763	3,120.710
			NFI	0.862	0.862

4.4. Evaluation of inner model

Inner model evaluation tests the relationship between variables (Hair et al., 2017) and can only be conducted if all variables and indicators contained in the research are valid and reliable. Evaluation of the structural model can be seen from the results of the direct effect significance test, indirect effect significance test, effect magnitude test, predictive relevance test, and coefficient of determination test.

When the value of the t-statistic is greater than the t-table (1.645) and the p-value smaller than 0.050 (Hair et al., 2019), it can be concluded that H1a—H4a are not rejected, but H4b is rejected (Table 9). Moreover, by looking at the effect size value, it can be concluded that H2b indicates a significant effect; H1a, H1b, and H2a indicate a moderate effect; and H4a and H4b indicate a small effect (Table 9).

Table 9. Direct effect significance test, indirect effect significance test, and effect size

Path	β	$\bar{x}$	s	t-statistics	p-value	f^2	Decision
H1a: WEB ➔ SAT	0.290	0.290	0.020	14.556	0.000	0.106	Not Rejected
H1b: PAS ➔ SAT	0.238	0.239	0.018	13.133	0.000	0.069	Not Rejected
H2a: DIS ➔ SAT	−0.240	−0.241	0.020	11.993	0.000	0.070	Not Rejected
H2b: DIS ➔ WEB	−0.361	−0.362	0.020	18.138	0.000	0.150	Not Rejected
H3: DIS ➔ WEB ➔ SAT	−0.105	−0.105	0.009	11.219	0.000	-	Not Rejected
H4a: PAS*WEB ➔ SAT	−0.058	−0.057	0.021	2.773	0.006	0.004	Not Rejected
H4b: PAS*DIS ➔ SAT	−0.000	0.000	0.023	0.017	0.986	0.000	Rejected

Based on the result of the structural measurement model in Figure 2, this study can be summarized using a multiple-step multiple mediator model approach (Hayes, 2009) with total effects into the following two equations:

(1) $WEB=\hspace{0.17em}-0.361\mathrm{\_}DIS\hspace{0.17em}+e_1$(1)

(2) $SAT\hspace{0.17em}=\hspace{0.17em}0.290\mathrm{\_}WEB\hspace{0.17em}-\hspace{0.17em}0.240\mathrm{\_}DIS\hspace{0.17em}+\hspace{0.17em}0.238\mathrm{\_}PAS\hspace{0.17em}-\hspace{0.17em}0.058\mathrm{\_}PAS\hspace{0.17em}\ast \hspace{0.17em}WEB\hspace{0.17em}+{\epsilon }_2$(2)

Figure 2. Structural measurement model.

Predictive relevance testing determines the ability to predict an indicator in a reflective model (Hair et al., 2019). If the value of Q2 ranges between 0.02 and 0.15, it can be concluded that the WEB and SAT indicators and variables show moderate predictive relevance (Table 10). By comparing the RMSE and MAE values of PLS-SEM in Table 10, which are all smaller than LM, it can be concluded that the model has high predictive power (Shmueli et al., 2019).

Table 10. Predictive relevance and coefficient of determination

		PLS-SEM		LM
	Q^2	RMSE	MAE	RMSE	MAE	R^2	Adjusted R^2
WEB	0.129	0.934	0.720			0.131	0.130
WEB-01	0.092	1.042	0.844	1.052	0.862
WEB-02	0.081	0.865	0.657	0.876	0.662
WEB-03	0.066	0.836	0.646	0.844	0.657
WEB-04	0.071	1.006	0.817	1.016	0.835
SAT	0.250	0.867	0.680			0.329	0.327
SAT-01	0.146	0.761	0.591	0.774	0.595
SAT-03	0.103	0.721	0.565	0.723	0.566
SAT-04	0.076	0.927	0.717	0.935	0.725
SAT-06	0.183	0.721	0.558	0.723	0.570

The coefficient of determination test measures the effect of the dependent variable on the independent variable. When the R2 value is between 0.10 and 0.50, then WEB can be explained by the study’s independent variable of 13.10%, while the SAT can be explained by the study’s independent variable of 32.90% (Table 10). According to Ozili (2022), a model with low R2 is generally accepted for studies in social science because human behavior cannot be accurately predicted. R2 between 0.10 and 0.50 is still good—if some or most of the explanatory variables are statistically significant.

5. Discussion

Based on the results of the tests, this study has succeeded in answering the questions that had been formulated, by reaching the following conclusions:

1. WEB and PAS among Indonesian Christian educators positively influence their SAT.

2. The presence of DIS negatively affects SAT and WEB of Indonesian Christian educators.

3. WEB partially mediates the impact of DIS on SAT of Indonesian Christian educators’.

4. PAS of Indonesian Christian educators negatively moderates the relationship between their WEB and SAT, but does not moderate the relationship between their DIS and SAT.

Our findings regarding the positive impact of WEB and PAS on SAT are particularly significant when considering Herzberg’s two-factor theory and the job characteristics model. These theories have long suggested that job characteristics and internal factors such as achievement and recognition play a vital role in enhancing SAT. This study provides empirical evidence to support the importance of WEB and PAS in SAT among Indonesian Christian educators. The results of the direct effect significance test for H1a indicate that WEB has a positive impact on SAT, consistent with previous studies (Adeka & Mede, 2022; Dreer, 2021; Yee et al., 2022). Similarly, as can be seen from the results of H1b, PAS also positively influences SAT, in line with the research of Carbonneau et al. (2008), Spehar et al. (2016), and Pathak and Srivastava (2020). Additionally, the holistic framework, which is based on the PERMA model and other relevant frameworks, validates the findings regarding the importance of WEB. These factors appear to be positively correlated with WEB, indirectly contributing to higher levels of SAT. Tailored WEB initiatives could be particularly useful for minority educators, who may face unique challenges such as discrimination or lack of representation. Moreover, professional development should include culturally sensitive techniques that can help foster PAS among these educators. In a multicultural school environment, it is essential to approach the enhancement of WEB with cultural sensitivity, taking into account the diverse backgrounds of the staff. Schools could provide various coping mechanisms and stress management techniques that align with the cultural norms and expectations of their staff.

As can be seen from the results for H2a and H2b, DIS negatively affects both WEB and SAT, as supported by Allahyari et al. (2018), Anastasiou and Papakonstantinou (2014), Bonsaksen et al. (2021), Chan et al. (2021), Friganović et al. (2019), Tran et al. (2022), Viertiö et al. (2021), Xie et al. (2021), and Zhang et al. (2021). In the broader context, the TTSC serves as a guiding lens to better understand the impact of DIS on SAT and WEB. This framework suggests that individual strategies for assessing and coping with stress directly contribute to the level of DIS experienced. Our study reveals that the path coefficient for DIS ranks as the second strongest, following WEB and preceding PAS, emphasizing its potent influence on SAT, as can be seen by comparing the results for H1a, H1b, and H2a. In essence, effective stress management interventions, particularly customized ones that address the unique stressors faced by minority educators, are crucial for improving SAT and WEB. These interventions can not only alleviate DIS but also substantially improve both WEB and SAT. Therefore, addressing these unique stressors is a crucial aspect of any intervention aimed at improving SAT and WEB among minority educators. In other words, implementing supportive measures, such as providing resources for stress management and coping strategies, can be beneficial for Indonesian Christian educators. By identifying the negative effect of DIS on WEB and SAT, this study highlights the importance of considering DIS as a crucial factor in the literature on SAT.

The study highlights the positive impact of WEB on SAT among Indonesian Christian educators. Looking back at the path coefficient value, WEB is considered the strongest determinant of SAT in this research. Educational institutions should prioritize the WEB of educators, as it has been shown to influence their SAT significantly. Strategies and interventions promoting WEB can increase SAT among Indonesian Christian educators. In a multicultural school environment, it is important to approach the enhancement of WEB with cultural sensitivity, taking into account students’ diverse backgrounds. Additionally, schools should provide a range of stress-reduction strategies and coping mechanisms that considers the various cultural expectations of staff members. These strategies benefit educators and contribute to creating a welcoming school climate. The findings reinforce the theoretical understanding of the significance of WEB in the context of SAT.

In addition, the analysis revealed that WEB partially mediates the relationship between DIS and SAT, as can be seen from the results for H3. Educational institutions should consider promoting WEB to mitigate the adverse effects of DIS on SAT. By addressing the underlying causes of DIS and promoting WEB, it is possible to mitigate its detrimental effects on SAT. In multicultural environments, analyzing current HR policies to ensure inclusivity and implementing specialized professional development programs that effectively manage classroom diversity will enhance educators’ WEB and SAT. Moreover, fostering community-building activities among educators in diverse settings can greatly improve SAT and lower DIS. Programs and initiatives focusing on enhancing WEB can help buffer the impact of DIS and contribute to greater SAT among Indonesian Christian educators. However, focusing on addressing DIS is still more crucial than enhancing WEB. The study provides empirical evidence supporting WEB as one of the critical determinants of SAT. It also establishes WEB’s mediating role, enhancing our understanding of the mechanisms through which DIS impacts SAT.

Furthermore, the study examined the positive impact of PAS on SAT. Educational institutions should foster an environment that nurtures and supports Indonesian Christian educators’ PAS. Recognizing and aligning job responsibilities with their PAS can enhance the overall SAT. While PAS was found to have a direct positive effect on SAT, as seen from the results for H1b, it did not significantly moderate the relationship between DIS and SAT, as suggested by the results for H4b. This is not in line with Benitez et al. (2023), who conducted their research on front-line workers from service organizations in southern Spain. Furthermore, they used the variable harmonious passion instead of PAS, which resulted in a substantial difference—while harmonious passion is driven by internalized desire, PAS is driven by a motivation to make a meaningful contribution in a specific field. In a multicultural school environment, considering PAS can be crucial. However, it is essential to remember that it may not serve as a buffer for the adverse effects of DIS on SAT. The essence of a multicultural setting requires recognizing that PAS can have different interpretations and impacts based on cultural norms. Therefore, relying solely on PAS to mitigate the effects of DIS may not be sufficient. Efforts should be directed toward addressing DIS rather than relying solely on PAS to mitigate its impact. One interesting finding is that PAS does moderate the relationship between WEB and SAT, but it does so negatively, as the results for H4a indicate. This is in line with Neubert and Halbesleben (2015), who suggested it is possible that what would usually be described as a positive factor, such as PAS, could have a negative impact in certain circumstances. This may be attributed to religious values and norms, where Indonesian Christian educators perceive their work as a spiritual calling. While PAS can be a powerful source of motivation, the high expectations and moral responsibilities attached to their role can also lead to pressure and stress. Social expectations from the environment may also contribute to the negative moderation effect of PAS on the relationship between WEB and SAT. Indonesian Christian educators are expected to exemplify moral integrity and dedication in their educational tasks, which can create pressure and affect their perceptions and experiences of work, despite their high PAS. In addition, other factors such as high workload, low-income level, or lack of resources could also possibly undermine the positive moderation effect of PAS between WEB and SAT. This highlights the importance of further exploration of alternative moderators and additional factors that could impact the relationship between DIS and SAT among Indonesian Christian educators, such as educator’s workload, income level, availability of material and financial resources, institutional support, professional development opportunities.

Another exciting finding could be found in the descriptive results of the DIS variable. Most of the respondents disagreed with any DIS indication. This is not in line with previous studies that were conducted in Indonesia, where Indra et al. (2021) and Fauzan et al. (2022) used the more popular sample. Therefore, these findings highlight the significance of understanding unpopular samples to fill the literature gap on the relevant model.

Despite the findings, this research also has certain limitations. The research framework can still be developed using other variables. Adding more variables can make the research results more useful and representative of the actual situation. The method used in this study is cross-sectional; to avoid common method bias and strengthen the causal relationships among the variables, future research can use the longitudinal method. Social determinants can also be used as variables that moderate the relationship of exogenous variables to the SAT. Future researchers are also encouraged to use other sampling techniques to avoid limited generalizability, potential bias, and subjectivity in respondent selection. Further, due to time constraints in implementing a larger research design, the respondents’ workplace background is overwhelmingly similar, with 87.32% placed in Christian private schools. Future research could consider a larger and more diverse sample to introduce greater variation into the study.

6. Conclusion

In conclusion, this study investigated the factors influencing the SAT of Indonesian Christian educators, who belong to a minority group in a predominantly Muslim country. The research findings have significant implications for educational policies, institutional support systems, and the broader theoretical landscape. These findings support and expand upon established theories like Herzberg’s two-factor theory, the job characteristics model, the PERMA framework, and the TTSC. Additionally, they offer novel insights into the application of these theories to Indonesian Christian educators in multicultural and diverse environments.

The study confirmed several significant relationships between the variables: WEB and PAS positively affect SAT, while DIS has a detrimental impact on both WEB and SAT. Additionally, WEB partially mediates the relationship between DIS and SAT, emphasizing the importance of reducing DIS to improve SAT. Educational institutions can manage DIS among Indonesian Christian educators by evaluating and modifying expectations, workload, and ensuring a work-life balance.

Notably, the path coefficient for DIS ranked as the second strongest, following WEB and preceding PAS, highlighting the critical need for effective stress management strategies in educational institutions. These findings contribute to the literature by emphasizing the critical role of DIS in SAT. They also highlight the positive impact of WEB on SAT and its mediatory role between DIS and SAT. Therefore, while enhancing educators’ WEB is a top priority, it is worth noting that managing DIS also plays a significant role in improving SAT.

Finally, although PAS had a positive influence on SAT, it did not significantly moderate the relationship between DIS and SAT. This suggests that interventions should prioritize addressing DIS rather than solely relying on PAS. Interestingly, PAS negatively moderated the relationship between WEB and SAT, which may be due to the unique religious values, norms, and social expectations imposed on Indonesian Christian educators. Considering these nuances, this study highlights the importance of adopting a multidimensional approach to enhance SAT among this particular demographic.

Geolocation information

Indonesia

Acknowledgments

We would like to thank Editage (www.editage.com) for editing and reviewing this manuscript for English language.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets analysed during the current study are available from the corresponding author on reasonable request.

Additional information

Funding

This research received no funding.