Employee Green Behaviour in the Vietnamese Healthcare Sector: Integrating Socially Responsible Human Resource Management and Environmental Transformational Leadership

Social Information Processing Theory (SIPT)

Salancik and Pfeffer’s (1978) SIPT proposes that the workplace social environment affects behaviour and performance. Social interactions help employees understand their workplace. The workplace influences employee attitudes, perceptions and motivation. According to Salancik and Pfeffer (1978), the social environment – in this case, the organisation – provides cues for constructing and interpreting events related to positive organisational attitudes and behaviours. Managers’ HRM use helps employees understand their workplace (Aryee et al., 2012; Cooper et al., 2019) in terms of important information, desired behaviour and results (Cooper et al., 2019). HRM policies should be firmly established within the organisation with effective communication and unwavering organisational backing to maximise the societal advantages of environmentally conscious employee actions (Dumont et al., 2017). According to SIPT, adaptive people change their mindsets in response to their social surroundings. Cognitive processes link social information to individual attitudes (Zhang et al., 2022).

Social cues and personal needs influence workplace behaviour (Salancik & Pfeffer, 1978). When making behavioural decisions, employees often consult their superiors (Boekhorst, 2015), who, in turn, shape culture and inform employees (Huhtala et al., 2013; Liu et al., 2021). SIPT posits that a nurse’s understanding of their work environments and interactions with co-workers influences their work attitudes and behaviours (Salancik & Pfeffer, 1978). Scholars have consistently shown the relevance of SIPT to organisational leadership studies (Wadei et al., 2021). According to the theory, employees’ impressions of the organisation are heavily influenced by its leaders. Leaders represent their organisations because of the authority and prestige they possess (Mohammed Sayed Mostafa & Shen, 2020). This leads employees to believe that the organisation is responsible for their actions (Gilbert et al., 2011). Leaders are responsible for understanding and implementing the organisation’s policies and practices. To make sense of these policies and procedures, followers consult the data supplied by their leaders (Mohammed Sayed Mostafa & Shen, 2020). Employees can learn the organisation’s rules, regulations and expectations through their leaders’ everyday interactions with them. Employees’ attitudes and behaviours on the job will be impacted by this (Jiang et al., 2015).

Socially Responsible Human Resource Management and Employee Green Behaviour

Green behaviour refers to personal actions and choices that positively influence the environment (Fawehinmi et al., 2020). These actions lessen the human impact on the environment (Stern, 2000) and promote long-term environmental sustainability (Ones & Dilchert, 2012). Organisations should establish formal procedures for promoting environmentally friendly behaviour (Fawehinmi et al., 2020). EGB highlights the actions and behaviours of employees in managing the environment, as aligned with the organisation’s expectations (Paillé & Boiral, 2013; Zhou & Zheng, 2024). By aligning employees with organisational CSR goals, SRHRM provides a new perspective on responsible management and provides additional micro-theory to explain the link between SRHRM and individual outcomes (Zhou & Zheng, 2024). Implementing a human resources practice that fosters CSR enhances employees’ sense of affiliation with the organisation (Lee et al., 2023; Shen & Benson, 2016). SIPT states that nurses’ work environment perceptions, attitudes and behaviours are socially constructed. Nurses proactively acquire information and comprehend their environment (Salancik & Pfeffer, 1978). Nurses may interpret SRHRM using social information processing, which includes both internal and external stakeholders, ultimately impacting their EGB.

SRHRM is ‘CSR policies and practices directed at employees. As such, SRHRM is not only an integral part of CSR initiatives but also an important tool for its successful implementation as it is through employees that the objectives of CSR are achieved’ (Shen & Benson, 2016, p. 1724). SRHRM can help nurses process social information and assess the organisation’s external image (Lee et al., 2023). Norton, Zacher, et al. (2017) found that EGB generates value and is prosocial, thereby helping an organisation achieve its CSR goals. When hospitals implement SRHRM strategies, nurses often believe that external stakeholders are being prioritised. These SRHRM practices help nurses recognise how their work benefits others and boost prosocial motivation (Wang et al., 2023). Hu and Liden (2014) defined prosocial employees as givers who place others before themselves and strive to improve their communities. Thus, prosocial employees are more inclined to engage in EGB (Wang et al., 2023). SIPT suggests prioritising sustainable values during recruitment, improving employees’ communication and engagement skills with external stakeholders related to promotion, rewards, and compensation to encourage EGB (Zhou & Zheng, 2024). These fundamentals influence nurses’ decisions as organisational cues. After reviewing the literature, we formulated the following hypothesis (H1):

The Mediating Role of Personal Moral Norms

Nurses are more inclined to view their employer as moral when they feel that they are internally accountable (Abdelmotaleb & Saha, 2020). Ellemers et al. (2011) found that socially responsible organisational practices and actions strongly influence employee morality. According to Abdelmotaleb and Saha (2020), nurses view SRHRM practices as a reflection of the moral and ethical character of their hospital. These practices include hiring and retaining socially responsible individuals, offering CSR training, and incorporating nurses’ social contributions into promotions, performance reviews and rewards. The SIPT states that observations, job behaviour, and others’ opinions are the main sources of information (Zhang et al., 2022). Managers portray the organisation as a social actor who upholds norms (Yam et al., 2018). Therefore, SRHRM practices communicate the organisation’s values and norms to nurses, which helps them understand how their organisations interact with stakeholders (Cooper et al., 2019; De Roeck & Farooq, 2018). Workplace practices and standards can strengthen or weaken morality (Elango et al., 2010). Victor and Cullen (1988) argued that individual traits cannot predict mortality. Organisational predictors significantly affect individual decisions. Organisational culture, rewards and beliefs strongly influence employee ethics (Elango et al., 2010). PMN are significant moral imperatives that compel employees to participate in pro-social behaviour (Schwartz, 1977), such as helping to protect the environment (Fawehinmi et al., 2020). Thøgersen (2006) defined PMN as an individual’s values-based inclination towards right and wrong. In this context, these feelings relate to environmental concerns (Fawehinmi et al., 2020). Accordingly, we argue that nurses’ SRHRM views can predict PMN.

Staddon et al. (2016) found that EGB has long-term economic and social benefits for the organisation; thus, EGB must be voluntarily adopted. Simply reminding workers of their environmental responsibilities can induce EGB (Fawehinmi et al., 2020). Bamberg and Möser (2007) stressed that EGB is ethically driven. Social discourse influences PMN, making it an individual duty. Schwartz and Howard (1981) defined taking responsibility as setting self-expectations for beneficial deeds and then feeling obligated to perform them. PMN is obligation-based intrinsic motivation. Energy conservation and recycling are part of this duty – employees may not enjoy it, but can derive satisfaction from it. While PMNs appear crucial to organisations, little research has linked them to EGB (Fawehinmi et al., 2020). Hence, we propose the following hypothesis:

The Mediating Role of Public Service Motivation

Some scholars have perceived PSM as a collection of beliefs and values (Ha & Doan, 2024). Vandenabeele (2007) characterised PSM as beliefs and attitudes transcending self and organisational interests, driving individuals towards actions benefiting broader entities. Given the SIPT framework, we can deduce that SRHRM initiatives stimulate PSM. The notion of working for a socially impactful organisation (SRHRM) can enhance nurses’ sense of purpose (Ha & Doan, 2024). For example, when nurses sense their organisation’s larger societal contribution, they experience enhanced job meaningfulness. By focusing on nurses’ well-being through SRHRM strategies, hospitals signal a loftier mission to their staff (Glavas & Kelley, 2014). SRHRM undertakings also convey pivotal messages about organisational morals and trustworthiness (Hansen et al., 2011). Building on this notion and the SIPT, this study posits a link between nurse perceptions of SRHRM and PSM.

The function of PSM as a mediator between HRM practices and employee outcomes has been the subject of several recent empirical investigations (Abdelmotaleb & Saha, 2019). The PSM studies discussed here have improved our comprehension of the idea. However, several academics have called for additional research to fully understand how organisational practices and policies affect the growth of PSM (Gould-Williams et al., 2014). PSM ‘may be formed by socio-historical factors before employees enter the organisation, but it will also be influenced by the work environment in which employees find themselves’ (Moynihan & Pandey, 2007, p. 42). This forms the basis of our hypothesis that nurses’ positive views of SRHRM are associated with PSM. Intrinsically motivated nurses have a higher chance of taking actions that benefit the workplace environment. EGBs taken by intrinsically motivated workers are likely to align with their values and priorities (Sabbir & Taufique, 2022). Put differently, green practices are more commonly employed by nurses when they have ‘a general altruistic motivation to serve the interests of a community of people’ (Piatak & Holt, 2020). Building upon the foundation provided by SIPT, we anticipate that the improved PSM, resulting from perceived SRHRM, will enhance nurses’ EGB.

Successive Mediation of PMN and PSM

As a set of principles for how one ought to act in one’s social environment, moral values encourage individuals to put their interests aside and work towards the greater good. Similarly, PSM encourages public servants to behave compassionately, put public values ahead of their interests, and act for the common good (Wang et al., 2020). The moral stance is intrinsic to and related to PSM-relevant beliefs and norms (Stazyk & Davis, 2015). Through socialisation, environmental reinforcement and value congruence, nurses with high levels of PSM can link institutional values to their morals (Stazyk & Davis, 2015; Wang et al., 2020). According to Piatak and Holt (2020), an example of a higher norm is the ‘commitment to the public interest or civic duty’. As a result, PSM’s results ought to align with the accepted social norms and values. Accordingly, it would make sense to assume that PMN can improve nurses’ workplace PSM.

Research has demonstrated that PSM extends beyond public institutions and encompasses a universal, selfless desire to serve the public (Wang et al., 2020). Regardless of whether one works in a public organisation, PSM is a combination of factors that motivate them to uphold social responsibility, stifle selfishness, and advance society – indeed, a range of prosocial behaviours have been linked to PSM (Piatak & Holt, 2020). We suggest a serial mediation model that connects SRHRM and EGB by taking into account the aforementioned arguments. Specifically, SRHRM, via cue signalling, will enhance nurses’ PMN, which, in turn, will benefit nurses’ PSM and promote EGB.

The Moderating Role of Environmental Transformational Leadership

According to SIPT, leaders influence their employees’ thoughts and actions (Boekhorst, 2015). Leaders’ positive information exchanges change employees’ attitudes, behaviours and perceptions, thus improving the relationship between SRHRM and EGB. Leaders send these signals to ensure employees understand the organisation’s environment and then use that knowledge to shape their thoughts (Gu et al., 2018). ETL, according to Chen and Chang (2013, p. 109), is ‘leaders’ behaviours that inspire followers to perform beyond expected levels of environmental performance and motivate followers to achieve environmental goals’. ETL can inspire followers and set a positive example for environmental values and behaviour (Asghar et al., 2022). ETL can also consider environmental remedies (Robertson & Barling, 2017) and inspire their subordinates to improve sustainability and engage in more long-term thinking (Asghar et al., 2022). SIPT states that employees’ perceived social information affects their behaviour. SRHRM and ELT practices affect employees’ attitudes and behaviours, and the expectations placed on them, thus making them important pieces of social information (Liu et al., 2023; Salancik & Pfeffer, 1978). As stated in the SIPT, nurses are better able to comprehend the hospital’s expectations, goals, and direction when they receive highly consistent cues from leadership and the organisational level. When nurses’ values match ETL and SRHRM practices, EGB rates rise (Liu et al., 2023). Leadership that prioritises environmental sustainability has the potential to increase intrinsic motivation among nurses (Graves & Sarkis, 2018). Nurses who are morally committed to sustainability and higher ideals are more likely to accept organisational goals (Graves & Sarkis, 2018). Moreover, ETLs have the potential to boost morale from within by giving nurses greater confidence in their ability to solve environmental problems (Asghar et al., 2022; Graves & Sarkis, 2018).

Employees perceive both formal HR practices and leadership behaviours as signals of their psychological contract with the organisation. These two elements of the employee management system may interact in ways that are either mutually reinforcing, conflicting, or functionally substitutive, depending on their alignment. When contract makers’ signals are inconsistent, there is a greater chance of unfavourable synergistic effects throughout the broader set of employee management systems (McClean & Collins, 2019). These effects can occur when ETL and SRHRM practices convey conflicting messages, creating a weak environment where employees are unclear about the firm’s values (Bowen & Ostroff, 2004). Additionally, an ETL can take the place of SRHRM procedures by giving more casual feedback, being considerate of members’ needs and skills, or rewarding staff in other, less conventional ways. By appreciating EGB in the lack of official SRHRM procedures, they might also inherently encourage independence and discretion (McClean & Collins, 2019). Consequently, the consistent informational cues shared by SRHRM and ETL can be seen as trustworthy ethical signals (Lv et al., 2024). When combined, they can impact nurses’ thinking and decision-making in a way that promotes ethics, helps them better understand moral issues, and ultimately improves their PMN and EGB. These earlier hypotheses point to a paradigm of moderated mediation. In particular, the various levels of ETL determine the indirect impact of SRHRM on EGB through PMN and PSM. Based on the above, we propose the following hypotheses:

The current research model is presented in Figure 1.

OPEN IN VIEWER

Figure 1.

Conceptual framework of the relationships among SRHRM, PMN, PSM, ETL, and EGB.

Data Collection and Sample

Data for the study were gathered from nurses working in Vietnamese public hospitals between October and December, 2023. Before the data collection, a pre-test was conducted in person with a small but representative sample of respondents (n = 20). The objective was to verify the questionnaire’s construct validity by determining whether any of the statements were difficult to understand (Colton & Covert, 2007; Ha et al., 2022; Shadish et al., 2002). Pre-testing serves as an initial step to gather empirical feedback on the appropriateness of the draft instrument. Participants are first asked to complete the questionnaire, after which they evaluate its format, content, clarity, terminology, usability and completion time. They are also invited to suggest specific additions or removals. Based on their feedback, the instrument should be reviewed and revised accordingly to improve its validity and usability. The questionnaire was pretested for participant comprehension (Lewis et al., 2005). Responses to the pre-test revealed that some SRHRM items were unclear. In every question, ‘company’ was changed to ‘hospital’. Moreover, the terms ‘social performance’ and ‘social responsibility’ were also ambiguous, so we provided examples to elucidate. For instance, a hospital’s social responsibility goes beyond providing quality medical examination and treatment services to protect and improve public health, ensure fair and humane working conditions for medical staff, and protect the environment through safe medical waste management. Besides, ‘social performance’ measures how well an organisation manages its social responsibilities and helps stakeholders and society flourish sustainably. Subsequently, we incorporated these clarifications into the questionnaire.

The survey sought to gather information from public nurses employed in 10 public hospitals in Vietnam. To sufficiently reflect the population, the sampling reflected the demographic features of nurses. Specifically, we employed the quota sampling technique, which aims to include a specific number of respondents from subgroups within the population. We then used a non-random selection mechanism to continue until each subgroup had the required number or quota of completed interviews (Acharyya & Bhattacharya, 2020). Quotas were established based on hospital affiliation, departmental designation (e.g., internal medicine, surgery, emergency, obstetrics and gynaecology), and gender distribution, mirroring essential demographic and structural characteristics of the nursing workforce. The distribution of quotas was informed by workforce data supplied by the hospitals’ administrative divisions and publicly accessible health sector reports. Samples can then be conveniently drawn during this second stage. Nevertheless, quota sampling is not a suitable substitute for probability sampling. Quota sampling is prone to favour those who are eager, readily available, and interested in the research issue. Therefore, representative samples that are comparable to those obtained using probability sampling cannot be obtained through quota sampling (Yang & Banamah, 2014). Ultimately, 334 nurses representing 10 public hospitals in Vietnam answered the questionnaires (66.8% of the 500 that were distributed). Demographic information of the survey participants is summarised in Table 1. According to Garver and Mentzer (1999), a ‘critical sample size’ of 200 yields reliable parameter estimations and has enough power to test a model. Although the response rate was 66.8%, we did not perform a formal non-response bias test, such as comparing early versus late respondents (Armstrong & Overton, 1977), due to limitations in time-stamped data collection or lack of access to full demographic information for all participants. We acknowledge this as a limitation of the study and recommend future research to include such an assessment in order to enhance the robustness of survey findings.

OPEN IN VIEWER

Table 1.

The Respondents’ Demographics.

Characteristics	Frequency	Percent (%)
Gender
Female	250	74.9
Male	84	25.1
Age (years)
Below 30	28	8.4
31–40	134	40.1
41–50	111	33.2
Above 50	61	18.3
Educational level
College	12	3.6
University	210	62.9
Others	112	33.5
Working experience (years)
Less than 3	43	12.9
3–5	68	20.3
6–10	113	33.8
11–15	65	19.5
Over 15	45	13.5
N	334	100%

We thoroughly briefed the participants on the study’s objectives, methodologies, potential hazards and advantages. All participants submitted written consent before engaging in the study. The privacy and confidentiality of participant data were fully ensured following applicable regulations.

Measurement of Constructs

An evaluation was conducted for each question using a five-point Likert scale, where strong disagreement (1) represented the lowest score and strong agreement (5) the highest. We used a distinct scale to evaluate each construct. Orlitzky and Swanson (2006) developed the six-item unidimensional scale, which Shen and Benson (2016) further enhanced, to assess SRHRM. PMN was assessed using four items adapted from Fawehinmi et al. (2020) and Steg and de Groot (2010). ETL was assessed using five items derived from Graves et al. (2013) and Asghar et al. (2022). The five-item PSM scale was measured according to that used by Wright et al. (2013). Wesselink et al. (2017) developed the five-item scale for EGB assessment, which they adapted from Fawehinmi et al. (2020).

Data Analysis

Using structural equation modelling (SEM), this study analyses survey data from 334 nurses. While multiple regression and SEM are fairly similar, the latter is far more robust and offers more analytical flexibility. Error terms, correlations, interactions, and other independent and dependent variables can all be modelled using SEM (Collier, 2020). Through model validation and visualisation, SEM is typically used to explain multiple statistical correlations at once. This method makes it easy to explain complex models. It is an expansion of conventional linear modelling methods, such as multiple regression analysis, which are necessary to master SEM. Simply put, it is the simultaneous combination of multiple regression analysis and factor analysis (Dash & Paul, 2021).

One of the most significant limitations of SEM is the required sample size, which is widely considered extensive (Collier, 2020). Indeed, even the most basic models require extensive samples, let alone intricate structures (Hair et al., 2019). Study reliability, power and generalisability are compromised by inadequate sample sizes (Martínez-López et al., 2013). The sample quantity is a critical factor in the use of SEM. Despite the widespread consensus regarding the necessity of a large sample size for SEM, there is no definitive guideline for the selection of an appropriate sample size. Additionally, normality is a concern in SEM analysis. General least squares and maximum likelihood estimation are the most widely used methods for SEM estimation, which presume that the data follow a normal distribution. Goodness-of-fit measures may be impacted and standard error terms may be undervalued if this restriction is violated (MacCallum et al., 1992). Thus, researchers must confirm multivariate normality and rectify non-normality, and calculate the skewness and kurtosis of the data to evaluate their normality (Kline, 2016).

The Sample’s Demographics

Evaluation of Measurement Model

We eliminated Items 5 and 6 from the SRHRM scale, Item 5 from the ETL scale, and Item 4 from the PSM scale due to cross-loading issues, as well as Items 4 and 5 from the EGB scale due to factor loading under .5 before evaluating the measurement model (Hair et al., 2019). That the data follow a normal distribution is a crucial assumption to make at the outset (Collier, 2020; Ha et al., 2022). Skewness and kurtosis tests are used to determine the normality of the data. A kurtosis of less than 7.0 and a skewness of less than 3.0 are indicators of data normality (Collier, 2020). The normality test confirmed that the data follow a normal distribution, which showed that kurtosis ranged from −.101 (EGB3) to 3.805 (ETL2) and skewness between −1.559 (EGB1) and −.578 (PSM2). Following this, we examined common method variance (CMV) through Harman’s single-factor test because the same instrument was used to collect all indicators (Jordan & Troth, 2019; Podsakoff & Organ, 1986). The results showed that one factor explained 46.753% of the variation in the 19 variables, well below the 50% threshold. Consequently, the dataset seemed unaffected by CMV (Podsakoff & Organ, 1986). It is important to note that although Harman’s single-factor test was conducted to evaluate common technique bias, relying entirely on this approach has several drawbacks. CMV is not adequately evaluated by this analysis, and it frequently allows researchers to assert support for data that is, in fact, unduly skewed by CMV (Fuller et al., 2016). Thus, more robust approaches, such as the marker variable test (Lindell & Whitney, 2001) or a full collinearity test (Kock, 2015), should be used in future research, as they may boost the reliability of the results.

To ensure the validity and accuracy of the measurement tools, Hair et al. (2019) suggested examining composite reliability (CR), average variance extracted (AVE), and standardised regression weights, among other factors. According to Table 2, every AVE value exceeded the recommended threshold of .50, with ETL having the lowest value at .560 and PMN having the highest at .809. Furthermore, the CR values for all measures were higher than the recommended threshold of .70, with ETL having the lowest value at .835 and PMN having the highest at .944. All items also had standardised regression weights that were higher than the .5 threshold (Hair et al., 2019).

OPEN IN VIEWER

Table 2.

The Constructs and their Corresponding Measures.

Construct	Code	Item	Loading	AVE	CR
Socially responsible HRM (SRHRM)	SRHRM1	‘My hospital considers individual identity-social responsibility identity fit in recruitment and selection’	.814***	.745	.921
	SRHRM2	‘My hospital provides adequate training to promote social responsibility as a core hospital value’	.867***
	SRHRM3	‘My hospital provides social responsibility training to develop employees’ skills in receptive stakeholder engagement and communication’	.878***
	SRHRM4	‘My hospital considers employee social performance in promotions’	.890***
	SRHRM5	‘My hospital considers employee social performance in performance appraisals’ (d)	–
	SRHRM6	‘My hospital relates employee social performance to rewards and compensation’ (d)	–
Personal moral norms (PMN)	PMN1	‘I feel guilty if I do not act pro-environmentally at work’	.834***	.809	.944
	PMN2	‘I feel morally obliged to act pro-environmentally at work’	.912***
	PMN3	‘I feel proud when I act pro-environmentally at work’	.948***
	PMN4	‘I would violate my principles if I did not act pro-environmentally at work’	.902***
Public service motivation (PSM)	PSM1	‘Meaningful public service is very important to me’	.824***	.604	.858
	PSM2	‘I am often reminded by daily events about how dependent we are on one another’	.875***
	PSM3	‘Making a difference in society means more to me than personal achievements’	.761***
	PSM4	‘I am prepared to make enormous sacrifices for the good of society’	–
	PSM5	‘I am not afraid to go to bat for the rights of others even if it means I will be ridiculed’	.627***
Environmental transformational leadership (ETL)	ETL1	‘My manager displays confidence about environmental issues’	.753***	.560	.835
	ETL2	‘My manager talks about the importance of protecting nature’	.812***
	ETL3	‘My manager talks enthusiastically about what we need to do to protect nature’	.787***
	ETL4	‘My manager gets me to look at environmental problems in new ways’	.630***
	ETL5	‘My manager provides teaching and coaching on environmental issues’ (d)	–
Employee green behaviour (EGB)	EGB1	‘I make sure that air-conditioning is switched off when not in the office’	.879***	.783	.915
	EGB2	‘I print and photocopy double-sided’	.940***
	EGB3	‘I switch off my computer/notebook when I leave my office for a considerable period’	.833***
	EGB4	‘I switch off the lights when I leave my office for a considerable period of time, and there is no one else’ (d)	–
	EGB5	‘I recycle plastics’ (d)	–

Note. d = indicator failed. EGB = employee green behaviour; ETL = environmental transformational leadership; PMN = personal moral norms; PSM = public service motivation; SRHRM = socially responsible human resource management.

***

Significant at p < .001.

Henseler et al. (2015) suggested a method for evaluating a scale’s discriminant validity that involves assessing the Heterotrait–Monotrait (HTMT) ratio of correlations, which measures the degree of association between related concepts. If the ratio exceeds .85, the discriminant validity is compromised (Kline, 2016). The constructs’ satisfactory discriminant validity was confirmed by our study’s results (HTMT values < .85). The particular HTMT values derived from our study are displayed in Table 3.

OPEN IN VIEWER

Table 3.

Heterotrait–Monotrait Ratio Results.

Construct	PMN	SRHRM	PSM	ETL	EGB
PMN
SRHRM	.647
PSM	.579	.514
ETL	.554	.594	.468
EGB	.604	.549	.608	.468

Note. HTMT discriminate at HTMT < .85. EGB = employee green behaviour; ETL = environmental transformational leadership; PMN = personal moral norms; PSM = public service motivation; SRHRM = socially responsible human resource management.

With χ² = 367.351 (df = 142, p = .000), χ²/df = 2.587, standardised root mean squared residual (SRMR) = .0565, comparative fit index (CFI) = .953, Tucker–Lewis Index (TLI) = .944, and root mean square error of approximation (RMSEA) = .069, the model fit statistics showed an excellent fit. This allowed us to confirm that the constructs were valid and reliable.

Evaluation of the Structural Model

A maximum likelihood estimation was then performed on the structural model to validate the hypotheses (Collier, 2020). The metrics used to evaluate the model’s fit are as follows: SRMR = .0701; CFI = .950; TLI = .940; RMSEA = .069; χ² = 408.069 (df = 158; p = .000). The fact that the χ²/df value of 2.583 fell between the 2 and 5 thresholds further demonstrates the model’s acceptability (Hu & Bentler, 1999). Therefore, the model’s goodness of fit was in the range of acceptable to excellent, as indicated by all of the fit indices (Ha et al., 2022). Different levels of statistical significance were demonstrated for each of the hypotheses. Table 4 shows the outcomes of the structural mode hypothesis testing.

OPEN IN VIEWER

Table 4.

Hypothesis Testing Using the Structural Model.

Relationship hypotheses	Expected effects	Standardised regression weights	Results
H1: SRHRM → EGB	Positive	.210**	Supported
H2: SRHRM → PMN → EGB	Positive	.142***	Supported
H3: SRHRM → PSM → EGB	Positive	.068**	Supported
H4: SRHRM → PMN → PSM → EGB	Positive	.187***	Supported
H5: SRHRM × ETL → EGB	Positive	.186***	Supported

Note. EGB = employee green behaviour; ETL = environmental transformational leadership; PMN = personal moral norms; PSM = public service motivation; SRHRM = socially responsible human resource management.

***

p < .001. **p < .010.

Mediation Analysis

A bootstrapping mediation analysis revealed the relationships between SRHRM, PMN and EGB. The permutation methods used approximately 1,999 permuted samples, while the bootstrap methods used 2,000 bootstrap samples per sample (Taylor & MacKinnon, 2012). The data showed that PMN significantly affects EGB and that SRHRM correlates with it. It appears that PSM mediates the SRHRM–EGB relationship. The bootstrap test also demonstrated that SRHRM indirectly influenced EGB consecutively via PMN and PSM. Table 5 summarises the mediation test results.

OPEN IN VIEWER

Table 5.

Results from a Mediation Evaluation.

Relationship	Mediating route	Estimate	Probability	Conclusion
H3: SRHRM – PMN – EGB		.142	<.001	Partial mediation
H4: SRHRM – PSM – EGB		.068	<.010	Partial mediation
H5: SRHRM – PMN – PSM – EGB		.187	<.001	Partial mediation

Note. EGB = employee green behaviour; ETL = environmental transformational leadership; PMN = personal moral norms; PSM = public service motivation; SRHRM = socially responsible human resource management.

***

p < .001. **p < .010.

Moderation Analysis

The models used contained all of the relevant variables, including those acting as moderators, dependents, and independents, as well as an interaction term (Collier, 2020). A strong positive effect of SRHRMxETL on EGB was found through these models (β = .187, p < .001), thus supporting H5. To distinguish between the high and low values of SRHRM and to identify the two degrees of this variable, the moderating effects of ETL were mapped out (Dawson, 2014). As shown in Figure 2, the relationship between SRHRM and EGB was positively moderated by ETL, especially at higher ETL levels. Furthermore, H6 was not supported by the insignificant effect of SRHRM × ETL on PMN.

OPEN IN VIEWER

Figure 2.

The moderating effect of ETL on the link between SRHRM and EGB.

Moderated Mediation Analysis

Using SPSS PROCESS Macro 4.2 Model 86, the moderated mediation model (H6) was analysed and tested with 5,000 bootstrapping re-samples and 95% confidence intervals (CI). Table 6 presents the results of the moderating effect of ETL on the indirect effects of SRHRM on EGB via the sequential path PMN and PSM. For the sequential indirect path (SRHRM → PMN → PSM → EGB), the index of moderated mediation was −.0111, SE = .0072, with a 95% confidence interval (BootLLCI = −.0274, BootULCI = .0006) that narrowly included zero, indicating that the moderated mediation effect was not statistically significant. H6 was therefore not substantiated.

OPEN IN VIEWER

Table 6.

The Moderated Mediation Effect of ETL.

Moderator (ETL)	Effect	BootSE	BootLLCI	BootULCI
Indirect effect: SRHRM → PMN → PSM→ EGB
−1 SD (−.6632)	.0613	.0173	.0287	.0964
M	.0539	.0154	.0250	.0851
+1 SD (.6632)	.0465	.0147	.0195	.0773
Index of moderated mediation
	Index	BootSE	BootLLCI	BootULCI
	−.0111	.0072	−.0274	.0006

Note. LLCI = lower confidence interval; ULCI = upper confidence interval.

Conclusion

This study sought to use the SIPT to investigate the relationship between SRHRM practices and EGB in Vietnam’s public healthcare sector. We found strong evidence for the positive relationship between SRHRM practices and nurses’ EGB. Additionally, our findings add credence to the idea that nurses’ PMN and PSM play mediating roles in the positive correlation between SRHRM and EGB. Finally, ETL was shown to play a moderating role between SRHRM practices and nurses’ EGB. However, the moderating of ETL in this relationship through PMN and PSM was not statistically significant. Our research shows that SRHRM practices and ETL style are useful tools for healthcare organisations to encourage environmentally conscious actions among their staff. This is especially true for organisations with environmental sustainability goals and objectives. Furthermore, the behaviours assessed by EGB are relatively simple and may exert little influence. Subsequent research should replicate the model to explain more complex behaviours.

Theoretical and Practical Implications

Limitations and Implications for Further Research

Naturally, this study carries certain limitations. First, we gathered our datasets from a cross-section of Ho Chi Minh City nurses working in the public healthcare system. Consequently, future research should include samples from different locations to assess the totality of the experimental findings. Second, this study merely examined PMN’s function as a go-between for SRHRM and EGB. No other variables were considered for their possible effects. Subsequent studies can undertake further investigations into the internal workings of the SRHRM–EGB relationship. Third, because intrinsic motivation is considered a strong predictor of job performance, motivational theories, such as self-determination theory (Deci & Ryan, 1985), can be applied (Norton, Parker, et al., 2015). Future research could address different levels of employee engagement with EGB by exploring whether different types of formal motivation can do so. Finally, our use of a cross-sectional design may have presented endogeneity as a potential weakness. Consequently, future studies may wish to use longitudinal or experimental methodologies to establish causality more definitively.