Crafting for sustainability: a daily diary study and self-training intervention on proactive employee engagement in sustainability

ABSTRACT

This research aims to stimulate employee engagement in sustainability by investigating whether self-initiated and trained job crafting functions as a behavioural tool to engage more in sustainability in daily work. Using a daily diary design in Study 1, we found that employees’ general attitude and perceptions of control towards sustainability predicted daily sustainability intentions which related to daily sustainability behaviour. Importantly, optimizing demands mediated the intention-behaviour gap. In Study 2, we tested a self-training intervention in which participants were encouraged to work on daily self-set crafting for sustainability goals. Results showed that the intervention group had increased sustainability intentions after the training. Furthermore, participants who actively followed the training increased more in optimizing demands throughout the training than control participants. The self-training did not affect overall proactive sustainability behaviour but resulted in increased specific sustainability activities through optimizing demands. Stimulating job crafting seems promising for implementing sustainability into organizational life.

KEYWORDS:

• Intervention
• job crafting
• proactive work behaviour
• sustainability
• theory of planned behaviour

Introduction

Sustainability has become an important movement in many organizations, often translated into corporate sustainability initiatives or human resource practices. Although these organizational initiatives may function as critical benchmarks to meet governmental regulations and communicate corporate consciousness to customers and stakeholders (Aguinis & Glavas, 2012), they may not necessarily lead to employee engagement in sustainability (Davis & Challenger, 2015). This is problematic because, in the end, employees need to act more sustainably to make an organization more sustainable. Important reasons include the absence of rewards for sustainable work behaviour (Daily & Huang, 2001) and a lack of individual sustainability motivation (Unsworth & McNeill, 2017). These reasons can be better understood when realizing that engaging in sustainability necessitates employees to change or adapt their work routines (Young et al., 2015). That is, working more sustainably usually requires extra effort or time in already demanding jobs (e.g., travelling by public transport instead of the car, adhering to strict safety procedures). Therefore, employees not only need reasons to change their behaviour (i.e., rewards, motivation), but they also need the skills and resources to rearrange their work environment accordingly (Stern, 2000).

The current research examines how sustainability motivation can be translated into sustainability behaviour at work. Active engagement in sustainability is usually not part of employees’ formal job requirements and is often even incompatible with expected profits or deadlines (Merriman et al., 2016), meaning that the work environment is usually not designed to facilitate engagement in this extra-role behaviour (Rupp et al., 2018). Building on the Job Demands-Resources (JD-R) theory (Demerouti et al., 2001), we specify the context that may stimulate engagement in sustainability by means of job demands and resources and propose that job crafting, or employees’ self-initiated changes in their job demands and job resources (Tims & Bakker, 2010), may be an effective strategy that enables employees to engage more in sustainability. A first reason is that by optimizing their job demands (e.g., working more efficiently) and seeking the necessary job resources (e.g., decision support, expertise) and challenges (e.g., searching for sustainable alternatives to commonly used products at work), employees can proactively create a challenging but resourceful work environment that facilitates contextual performance such as contributing to sustainability in their work (cf. Demerouti et al., 2015). Secondly, through crafting employees can establish a better fit between their job and their personal goals or values (Kooij et al., 2017). And third, crafting behaviour may complement (top-down) corporate sustainability initiatives by producing creative and hands-on sustainability solutions that can uniquely be developed from a bottom-up perspective (Le Blanc et al., 2017).

By suggesting that job crafting can be a helpful strategy to engage more in sustainability during daily work, we complement previous research on employee sustainability motivation. That is, employees’ sustainability behaviour is driven by their attitudes towards sustainability, goals and values, self-efficacy, and (perceived) social norms (Stern, 2000; Swaim et al., 2016; Unsworth & McNeill, 2017). These insights have been useful to understand why employees would be motivated to engage in sustainability, but they cannot explain how employees can do this. By integrating the job crafting literature with the Theory of Planned Behaviour (TPB) and the JD-R theory, the current research aims to examine whether employees can (be trained to) craft a resourceful but challenging work environment (i.e., “an active job”; Petrou et al., 2012) that is required to implement their sustainability intentions. We investigate which stable motivational sustainability beliefs result in daily sustainability intentions, and whether employees can bridge the daily sustainability intention-behaviour gap by mobilizing the necessary job resources, setting new job challenges to themselves, and optimizing their job demands that frustrate the implementation of sustainability in their work. We conduct a daily diary study to examine our hypothesized model among employees who report on their work experiences as they occur, and we conduct a self-training intervention study to examine whether we can boost this hypothesized process by training employees in crafting for sustainability.

In doing so, we contribute to the sustainability literature by uncovering how employees themselves can proactively contribute to sustainability (Aguinis & Glavas, 2012; Stern, 2000). We contribute to the behavioural change literature by adopting a JD-R perspective to explain how the TPB’s psychological drivers relate to actual work behaviour. This adds to the current understanding of the TPB because we draw attention to so far over-looked obstacles in the work environment (and potential ways to cope with them) that impact the behavioural mechanism connecting intentions and behaviours. Hence, our research may help to identify why individuals who intend to act in a certain way sometimes fail to do so: They may lack the skills, means, and resources to put their intentions into action (cf. Ajzen, 2015). This is especially relevant for sustainability behaviour given the typical costs (i.e., time, know-how, flexibility) that are associated with it. However, we aim to reveal that employees can mobilize these means themselves through crafting. These insights may help to activate the responsibility of employees who may feel insignificant or useless to the complex puzzle of sustainable development (Lamm et al., 2015). Finally, we add to job crafting theory because our intervention may empirically demonstrate the value of job crafting beyond individual work outcomes like employee well-being and performance to more communal outcomes (i.e., responsible crafting; Bizzi, 2020). In addition, we introduce one of the first job crafting interventions explicitly designed according to the TPB, thereby strengthening the theoretical underpinnings of job crafting interventions (see also Costantini et al., 2022).

Implementing sustainability at work

Sustainability is about the conservation of resources. This broad definition origins from the Brundtland Report (World Commission on Environment and Development WCED, 1987) which defined sustainable development as “development that meets the needs of the present without compromising the ability of future generations to meet their own needs” (pp. 16). A widely adopted framework in the work domain is the triple bottom line of corporate sustainability consisting of economic prosperity, environmental quality, and social equity (Elkington, 1997), also named “the triple P framework”: People, Planet, and Prosperity. We draw from this framework in the current research because it signals that sustainability is concerned with more than just the environment. Many organizations have adopted sustainability strategies that are usually concerned with the raw materials they use and activities in their supply chain (i.e., logistics, labour standards; Carter & Rogers, 2008). Although these organizational initiatives may contribute to more conscious industries, they do not necessarily influence employees’ daily work behaviour. This is problematic because all the decisions that employees make during their work form a great deal of an organization’s actual impact (Lamm et al., 2015). The current research zooms in on employees’ daily proactive sustainability behaviour using a diary study and a self-training intervention. An important reason for this daily focus is that daily behaviours can be learned or changed (Ohly et al., 2010), and that daily behavioural processes may uncover the real-time drivers and barriers that employees face when trying to engage in sustainability. We focus on employees’ proactive sustainability behaviour, which represents the general initiatives employees take to engage in the people, planet, and prosperity domains of sustainability beyond their required work tasks (Bissing-Olson et al., 2013).

The theory of planned sustainability behaviour

An important particularity of proactive sustainability behaviour at work is that it is usually considered extra-role behaviour and not part of one’s key performance indicators (Swaim et al., 2016). Secondly, compared to other work behaviours or sustainability behaviour at home, there are no direct (financial) benefits for the individual employee (Inauen et al., 2016). Third, engaging in sustainability at work occurs within the context of organizational or task-related boundaries meaning that employees may need to seek or create new solutions to integrate sustainability with their existing work duties (Lülfs & Hahn, 2014). These particularities emphasize the importance of employees’ motivation to behave more sustainably at work. A framework that explains the individual motivation for behavioural change is the TPB. The TPB predicts that the primary driver of behaviour is individuals’ behavioural intention. This behavioural intention is preceded by individuals’ attitudes towards the outcomes of the behaviour, their perceptions of the social evaluation of performing that behaviour, and their (self-) efficacy beliefs regarding their ability to engage in the behaviour (Ajzen, 1991). Behavioural intentions consist of a relatively stable, trait-like component on which people differ from each other, as well as a state-like component that fluctuates over time (Inauen et al., 2016). Notwithstanding, the TPB has predominantly focused on this stable component and applied it to describe between-person differences in behavioural change to find that the intention-behaviour gap might be blurred by daily or momentary factors hindering the implementation of intended behaviours (Ajzen, 2015). The current study, therefore, zooms in on daily (within-person) processes impacting the intention-behaviour processes of proactive sustainability behaviour.

From the TPB it is expected that employees who positively value sustainability (i.e., attitude), who have the perception that their social network values sustainability (i.e., subjective norm), and who feel able to engage in sustainability during their work (i.e., perceptions of control) are most willing to put the effort in implementing sustainability (i.e., behavioural intention), and, consequently, will act accordingly (i.e., behaviour). In this light, Swaim et al.’s (2016) vignette study showed that personal values and control beliefs exert a significant influence on how supply managers make decisions to pursue environmental sustainability for the organization. A comparable field study showed that senior marketing managers holding positive attitudes and who are exposed to a favourable social norm towards sustainability seem to be more motivated to engage in sustainable marketing practices (Ferdous, 2010). The relevance of personal attitudes for behavioural sustainability outcomes is further nuanced by Unsworth and McNeill (2017) who showed that as long as employees have self-concordant goals that are related to sustainability (i.e., saving money or getting more exercise by commuting with public transport instead of the car) they are willing to change their behaviour. Stern (2000) summarized causal factors for environmentally significant behaviours and concluded that attitudinal factors seem a driving force but that they may interact (especially when the behaviour is costly or difficult) with personal capabilities, contextual factors, and habits and routines. Based on this literature, we hypothesize:

Hypothesis 1:

Employees with more favourable attitudes, subjective norms, and perceptions of control towards sustainability will have more daily behavioural intentions towards sustainability, and, in turn, engage more in daily proactive sustainability behaviour.

Job crafting as a mechanism

Using the TPB we have theorized that employees’ daily sustainability intentions and behaviours are driven by their attitudes, subjective norm, and perceptions of behavioural control (Swaim et al., 2016). However, research has shown that there are ample situations in which people have positive intentions but fail to act on them which has been related to (i) a lack of means, skills, and resources to perform the behaviour or (ii) potential barriers in the (work) environment (Ajzen, 2015). This seems especially true for proactive sustainability behaviour at work which requires specific knowledge, flexibility, and additional resources to adapt existing work procedures (Bhattacharya et al., 2023). Moreover, even when employees have the intention to embrace sustainability in their daily work, obstacles like busy work schedules, (financial) pressures to perform, a lack of (supervisor) support, and standard work procedures may complicate the enactment of their sustainability intentions (Merriman et al., 2016; Rupp et al., 2018). At this point, we integrate JD-R theory with the TPB and we argue that job crafting for sustainability is a hands-on and accessible tool to proactively make the work environment more resourceful, challenging, and not overly demanding so that it facilitates employees to work more sustainably. Job crafting can be defined as the self-initiated changes employees make to balance their job demands and resources with their individual needs, goals, and competencies (Tims & Bakker, 2010). Job demands are those aspects of work that cost effort or time. Job resources are those aspects of work that help employees cope with their job demands and contribute to goal achievement and development (Demerouti et al., 2001). Jobs with manageable demands and sufficiently high resources are suggested to stimulate high task and adaptive performance (Demerouti, 2020).

Drawing from JD-R theory (Bakker et al., 2023), we theorize that job crafting facilitates proactive engagement in sustainability for the following reasons. First, through crafting employees arrange their work environment such that they are not hindered by job demands and that they have sufficient job resources to invest in contextual performance domains like the implementation of sustainability (cf. Demerouti et al., 2015). This is important because working more sustainably comes with specific demands (e.g., ethical dilemmas, complexity, longer-term perspective) and requires specific resources (e.g., independence, flexibility, time) that need to be balanced to make it work (Bhattacharya et al., 2023). Second, job crafting increases the fit between the job and what a person desires (Kooij et al., 2017). As sustainability behaviour in the work context has shown to be strongly driven by personal values (Swaim et al., 2016), crafting may be a very useful vehicle to act upon one’s own sustainability intentions. And finally, job crafting may lead to innovative and novel sustainability solutions that can only be developed from a bottom-up perspective (cf. Le Blanc et al., 2017). People that are closest to the daily sustainability decisions in a firm may proactively spot opportunities to change aspects of the work environment that corporate strategies may overlook when dealing with the complexity of sustainable development (Lamm et al., 2015).

Employees can craft their job using different job crafting strategies, namely seeking resources, seeking challenges, and optimizing demands (Demerouti & Peeters, 2018; Petrou et al., 2012). We propose that on days when employees engage in each of these strategies their proactive sustainability behaviour will thrive. Specifically, seeking resources involves asking colleagues or supervisors for advice or support and seeking learning opportunities (Petrou et al., 2012). On days that employees target their resource-seeking behaviour to the sustainability domain, they require the support or know-how that they need to find sustainable solutions to the everyday demands or dilemmas that they face (Aguinis & Glavas, 2012; Swaim et al., 2016) which will enhance their sustainability behaviour. Seeking challenges involves searching for new tasks when work is finished or taking on extra responsibilities (Petrou et al., 2012). On days that employees seek challenges in the sustainability domain (i.e., which is typically seen as contextual performance; Bissing-Olson et al., 2013) they may boost their work engagement and experience of meaningful work, which may, in turn, fuel their contextual sustainability behaviour (cf. Demerouti et al., 2015). The final job crafting strategy, optimizing demands, involves finding ways to work more efficiently and simplifying processes or procedures (Demerouti & Peeters, 2018). On days that employees optimize their demands such that they save time and energy to engage in sustainability, their daily sustainability behaviour will be enhanced (Merriman et al., 2016; Rupp et al., 2018). We intentionally focus on optimizing demands instead of reducing, minimizing, or decreasing hindering demands (Demerouti et al., 2021) because employees who want to implement sustainability in their work need to perform this behaviour without neglecting their work duties. Hence, finding ways to make work more efficient and to eliminate obstacles seems a more effective strategy to engage in sustainability than avoiding existing strenuous aspects of their job (Demerouti & Peeters, 2018). Another reason to examine the role of optimizing demands instead of minimizing demands is that minimizing demands does not seem to be an effective strategy. Minimizing demands has been found to be either unrelated or negatively related to performance outcomes (e.g., Demerouti et al., 2017) which potentially signals that it represents attempts to diminish stress or health problems. Recently, Demerouti et al., (2021) included optimizing demands in their training which was shown to increase safety behaviour.

Hypothesis 2:

The positive link between daily behavioural intention towards sustainability and daily proactive sustainability behaviour is mediated by daily seeking resources (2a), daily seeking challenges (2b), and daily optimizing demands (2c).

Study 1

Materials and methods

The current research includes a daily diary study that we tested in Study 1 and a self-training intervention that we tested in Study 2. For pragmatic reasons, we used the control group of the self-training intervention to examine the daily hypotheses. This means that the sample of Study 1 is also part of the sample of Study 2 and that the studies were conducted in parallel. The present research adheres to the APA ethics code and Dutch ethics guidelines. We received institutional ethics approval for both studies. The data of our studies is available upon request by the first author.

Procedure and participants

Participants were recruited through the networks of a group of master students at a Dutch university. Each student recruited four participants (two participated in the intervention and 2 were part of the control group), and the inclusion criteria were (i) that participants were 18 years or older and (ii) that they were minimal 50% employed. Participants were provided with paper diary booklets that included the person-level measures and all daily diaries. Participants were asked to complete the person-level measures, including demographics, before the daily diaries. Furthermore, they were instructed to complete the daily diaries on five consecutive working days: The morning diary at the start of their working day and the afternoon diary at the end. In doing so, we aligned the measurements’ timing with our theoretical model, which focuses on behavioural intentions (assessed in the morning) that may translate into work behaviour (assessed in the afternoon). To ensure data quality, participants were asked to indicate the date and specific time of filling out the diaries. Furthermore, students provided the completed diary booklets and contact details of their participants to the researchers to verify the data’s trustworthiness.

Sixty-eight employees, including 26 women, participated in the diary study. Their mean age was 34.5 years (SD = 14.0). Participants primarily worked in the industrial or construction sector (29.8%), the healthcare or educational industry (20.9%), or financial or business services (14.9%). The mean working hours per week were 37.5 (SD = 6.6). Participants had worked on average for 6.4 years (SD = 10.3) for their current employer, and 20.9% fulfilled a supervisory position. We based the required sample size for sufficient power of our study on Ohly et al. (2010) recommendations, which state that multilevel studies aimed at explaining daily within-person fluctuations should aim for at least 5 measurement days per person and include a minimum of 30 participants.

Measures

Person-level-measures

Theory of planned sustainability behaviour.

Attitude, subjective norm, and perceived behavioural control towards sustainability were measured with three items each from Swaim et al. (2016) that were adapted to capture sustainability in accordance with the triple P framework (Elkington, 1997). The attitude items are: “It is important for me to act sustainable”, “It is important for me to conserve natural and human resources”, “I am concerned about the long-term condition of the environment”. The subjective norm items are: “People important to me think I should buy sustainable products”, “People important to me think I should support the environment”, and “People important to me think I should conserve natural and human resources”. The items for perceived behavioural control are: “I can perform sustainability activities”, “I have control over my actions to support the environment”, “I have the ability to carry out sustainability activities”. Items were scored on a 7-point scale, ranging from (1) “strongly disagree” to (7) “strongly agree”.

Daily diaries

The diary booklet consisted of five identical morning surveys and five identical afternoon surveys. The included scales were short versions of existing scales to limit completion time (Ohly et al., 2010). Furthermore, the items were adapted to the day level so that respondents could indicate how representative each statement was for the respective day. Unless stated otherwise, items were scored on a 7-point scale, ranging from (1) “does not apply to me” to (7) “totally applies to me”.

Daily behavioural intention towards sustainability

Daily behavioural intention towards sustainability was measured in the morning with the following three items from Swaim et al. (2016) that were adapted to capture sustainability in a broader sense: “Today, I intend to seek out ways to be more sustainable”, “Today, I plan to support sustainability activities”, “Today, I plan to play a role in reducing harm to the environment”.

Daily job crafting

Daily job crafting was assessed in the afternoon with six items developed by Petrou et al. (2012) measuring seeking resources and seeking challenges, and three items developed by Demerouti and Peeters (2018) measuring optimizing demands. Example items are “Today, I asked colleagues for advice” (seeking resources), “Today, I asked for more responsibilities” (seeking challenges), and “Today, I looked for ways to do my work more efficiently” (optimizing demands).

Daily proactive sustainability behaviour

Daily proactive sustainability behaviour was measured in the afternoon with the adapted 3-item daily proactive pro-environmental behaviour scale by Bissing-Olson et al. (2013). The adapted items are: “Today, I took initiative to act in sustainable ways at work”, “Today, I took a chance to get actively involved in sustainability activities at work”, Today, I did more for the environment at work than I am expected to”.

Analytical approach

The current dataset has a multilevel structure with 340 daily measurements at the lower level nested within 68 persons at the higher level. To account for this multilevel structure, we analysed our data using multilevel structural equation modelling in Mplus (Muthén & Muthén, 1998–2017). To justify this approach, we first examined our daily constructs’ between-person and within-person variance components by calculating their intraclass correlation coefficients (ICCs). The ICCs showed that variance in the daily constructs explained by the higher level, thus attributed to between-person variations, ranged between 55% and 69% (Table 1). This means that there is enough variance to be explained at both the between-level and within-level of the daily constructs.

Table 1. Means, standard deviations, Cronbach’s alphas, intraclass correlations, and correlations among the study variables in the diary study.

	M	SD	α^a	ICC	1	2	3	4	5	6	7	8
General level
1.Attitude towards sustainability	5.50	0.87	0.65
2.Subjective norm towards sustainability	4.29	1.24	0.87		.24*
3.Perceived behavioural control towards sustainability	5.15	0.99	0.71		.22	.18
Day level
4.Behavioural intention towards sustainability	4.53	1.14	0.87	0.66	.35**	.22	.32**		.06	.01	.15**	.15**
5.Seeking resources	4.47	1.25	0.78	0.58	.21	.08	.27*	.51***		.30***	.29***	.17**
6.Seeking challenges	2.89	1.45	0.91	0.69	−.14	−.05	.05	.25*	.51***		.20***	.28***
7.Optimizing demands	4.50	1.24	0.90	0.58	.14	−.02	−.06	.31*	.52***	.47***		.29***
8.Proactive sustainability behaviour	3.85	1.28	0.86	0.62	.18	.17	.26*	.67***	.51***	.39**	.54***

Notes. Correlations at the within-person level are displayed above the diagonal (N = 340; centered to the person mean), whereas correlations at the between-person level are displayed below the diagonal (N = 68; averaged across the five days).

^aCronbach’s alphas for the daily variables were averaged across the five days.

*p < .05; **p < .01; ***p < .001.

Figure 1 shows our hypothesized model. We treated attitude, subjective norm, and perceived behavioural control towards sustainability as between-level variables because they are relatively stable (Ajzen, 1991), and the three daily job crafting strategies as within-level variables because they may fluctuate across days. We did not specify the other variables’ measurement levels because they are used on both levels of our model (for a similar approach, see Petrou et al., 2012). Explanatory variables at the between-level were centered to the grand mean, and explanatory variables at the within-level were centered to the group mean. The maximum likelihood estimator was used in all analyses.

Figure 1. Hypothesized model of the diary study.

Results

Descriptive statistics

Means, standard deviations, Cronbach’s alphas, intra-class correlations, and correlations among the study variables are displayed in Table 1.

Hypotheses testing

Our multilevel structural equation model showed a good fit to the data χ² = 5.43, df = 4, p = .246, χ² /df = 1.36, CFI = 0.99, TLI = 0.98, RMSEA = 0.03, SRMR = 0.01 (within-person level) and 0.08 (between-person level). Figure 2 displays the standardized estimates and 95% CI’s of the hypothesized paths, and Table 2 the standardized estimates and 95% CI’s of the indirect effects.

Figure 2. Standardized estimates and 95% CI’'s (in brackets) for multilevel structural equation modelling of the diary study.

Note. Only the significant paths are shown.

Table 2. Standardized maximum likelihood estimates, standard errors, and 95% CI’s for the indirect effects in the diary study.

	Indirect effect	Est.	SE	95% CI	p
Between-person level
1.	Attitude → Behavioural intention → Proactive sustainability behaviour	0.18	0.08	0.01, 0.34	0.030
2.	Subjective norm → Behavioural intention → Proactive sustainability behaviour	0.09	0.08	−0.07, 0.25	0.256
3.	Perceived behavioural control → Behavioural intention → Proactive sustainability behaviour	0.18	0.08	0.02, 0,34	0.033
Within-person level
4.	Behavioural intention → Seeking resources → Proactive sustainability behaviour	0.01	0.02	−0.02, 0.05	0.467
5.	Behavioural intention → Seeking challenges → Proactive sustainability behaviour	0.05	0.03	−0.01, 0.11	0.072
6.	Behavioural intention → Optimizing demands → Proactive sustainability behaviour	0.09	0.03	0.03, 0.16	0.004

Notes. Attitude, subjective norm, and perceived behavioural control towards sustainability are measured at the general level. Behavioural intention towards sustainability, job crafting (seeking resources, seeking challenges, and optimizing demands), and proactive sustainability behaviour are measured at the day level.

In line with the TPB (Hypothesis 1), we found that employees with a more favourable attitude and higher perceptions of behavioural control towards sustainability engaged more in daily proactive sustainability behaviour through increased daily behavioural intentions towards sustainability (Table 2, indirect effects 1 and 3). The subjective norm did not affect daily sustainability intentions or behaviour and we also did not find a direct relationship between employees’ perceived behavioural control towards sustainability and their daily proactive sustainability behaviour.

Hypothesis 2 predicted that the positive link between daily behavioural intention towards sustainability and daily proactive sustainability behaviour was mediated by daily seeking resources (2a), daily seeking challenges (2b), and daily optimizing demands (2c). Our results showed that the indirect effect through optimizing demands was significant (Table 2, indirect effect 6). There was no indirect effect through seeking challenges (albeit this was marginally significant) or seeking resources. These results confirm Hypothesis 2c and reject Hypotheses 2a and 2b.

Brief discussion of study 1

Our daily diary study largely confirmed that the TPB antecedents predict daily intentions to engage in sustainability and, in turn, daily proactive sustainability behaviour. This finding extends previous research showing that the TPB explains general employee engagement in sustainability (Ferdous, 2010; Swaim et al., 2016). The TPB antecedent that had no effect on daily sustainability intentions and behaviour was the subjective norm. This finding is not unique; meta-analytic findings on sustainability behaviour explained by the TPB show that norm-based influences on intentions and behaviour tend to be weaker than attitude- or control-based influences (Klöckner, 2013).

These daily analyses revealed that the job crafting strategy optimizing demands functioned as a behavioural mechanism to translate daily sustainability intentions into proactive sustainability behaviour. Building on the notion that engaging in sustainability is effortful and time-consuming for employees (Merriman et al., 2016; Rupp et al., 2018), we show that on days when employees proactively optimize the demands in their work environment, they transform their sustainability intentions into behaviour. Against our expectations, we found no such effect for the other job crafting strategies. A post-hoc explanation could be that employees did not seek resources or challenges related to the implementation of sustainability but that they asked for feedback or extra responsibilities in a different performance domain. Yet, to examine whether crafting resources and demands/challenges may lead to sustainable work behaviour when employees are explicitly trained to use these strategies to facilitate sustainability engagement, we now introduce our self-training intervention on crafting for sustainability.

Job crafting interventions

Although job crafting is self-initiated behaviour, it can also be trained using a job crafting intervention. The meta-analysis of Oprea et al. (2019) showed that job crafting interventions effectively stimulate job crafting behaviour and, to some extent, performance and well-being. We employ the job crafting intervention to motivate participants to craft their job in order to engage in sustainability. This means that we introduce job crafting as a tool that participants can target towards the sustainability domain. To increase the scope of our intervention, we use a self-training format which seems to be effective to train job crafting (Demerouti, 2023). Similar to classic job crafting interventions, our sustainability self-training targets a change in attitudes, cognitions, and behaviour (Demerouti et al., 2019; Oprea et al., 2019). Moreover, we have embedded the self-training in the TPB, which has recently been postulated as a useful framework for designing job crafting interventions (Costantini et al., 2019, 2022). More specifically, our self-training intervention first aims to stimulate participants’ internal motivation to craft for sustainability by making the attitudes and subjective norms towards crafting for sustainability more salient and by increasing participants’ sense of control towards it. In addition, the intervention focuses on the implementation of crafting for sustainability goals by encouraging participants to translate these goals into real behaviours (Ajzen, 2015; Costantini et al., 2022). To accomplish this, we use several behavioural change techniques (cf. Michie et al., 2013) such as focusing on past successes, listing the pros and cons of crafting for sustainability, goal-setting, action planning, and self-monitoring. Given the complexity of stimulating proactive sustainability behaviour, we consider this mixture of behavioural change techniques a feasible method to achieve behavioural change in crafting for sustainability. That is, our intervention goes far beyond information-based training interventions proven to be inadequate to achieve behavioural change in sustainability by providing intellectual information on sustainability problems alone (cf. Maiteny, 2002). Although prior research has shown that positive sustainability attitudes may be absent in many people (Unsworth & McNeill, 2017), we expect that by heightening awareness on the positive outcomes that can be derived from crafting for sustainability (i.e., listing pros and cons, describing past successes) and by strengthening positive control beliefs concerning it (i.e., making an action plan), participants in the intervention will boost their beliefs towards sustainability in general (Costantini et al., 2022).

Hypothesis 3:

Employees participating in the self-training intervention will have a more positive attitude (3a), subjective norm (3b), perceived behavioural control (3c), and behavioural intention (3d) towards sustainability after the intervention compared to the employees in the control group.

The central skill of our self-training intervention is job crafting. Participants are familiarized with the different job crafting strategies and encouraged to practice them during the training period as a means to engage more in sustainability (Table 3). Specifically, each morning, participants (i) set a personal SMART-formulated crafting for sustainability goal and they indicate (ii) what they are going to change in their work environment to enable sustainability engagement, and (iii) when they are going to do that (see Table 4 for the specific instructions). They are encouraged to work on their crafting for sustainability goals during the day and reflect on their behaviour in the afternoon. This method is grounded in the job crafting intervention literature and has been shown to be effective in promoting job crafting behaviour (e.g., Dubbelt et al., 2019; Gordon et al., 2018). We expect that participants will develop their job crafting behaviour during the training period through these exercises.

Table 3. Overview of the crafting for sustainability self-training intervention.

Timing	Theoretical constructs	Behaviour change techniques in the intervention exercises	Prior job crafting interventions applying similar exercises
Pre-measures	Attitude	Reading information about potential positive consequences	Demerouti (2020); Dubbelt et al. (2019)
Morning Day 1	Attitude	Focusing on past successes Listing pros and cons	Costantini et al. (2022); Demerouti (2020); Dubbelt et al. (2019)
Morning Day 2	Subjective norm and social support	Reflection on others’ perceived approval Reflection on expected practical and emotional support	Costantini et al. (2022)
Morning Day 3–5^a	Perceived behavioural control and implementation intentions	Goal setting and action planning Coping planning Behavioural rehearsal	Costantini et al. (2022); Demerouti (2020); Dubbelt et al. (2019); Gordon et al. (2018); Kooij et al. (2017); van den Heuvel et al. (2015); van Wingerden et al. (2017); van Wingerden et al. (2017)
Afternoon Day 3–5^a	Self-monitoring	Self-monitoring Imagining (and listing) future improvements	Demerouti (2020); van Wingerden et al. (2017); van Wingerden et al. (2017)

Note. Behaviour change techniques are described using the taxonomy of Michie et al. (2013).

^aDay 3, 4, and 5 are similar in content.

Table 4. Instructions and examples of the daily crafting for sustainability exercises.

Day	Instructions	Examples
1	Have you already used crafting techniques (i.e., proactively making small adjustments to your work environment) to make your work more sustainable?	Bringing reusable bottles and mugs for coffee and water Planning meetings via Skype vs. in real life
1	Think about the consequences of crafting your current work environment to make it more sustainable. What benefits would this have?	Benefits in terms of energy costs for the organization It gives me a good feeling and I work on my fitness I contribute to a better world
1	And what are the downsides of such behaviours?	People might think that I am a freak It could be annoying for my colleagues It is time-consuming
2	Describe what people important to you would think about crafting for sustainability and how they could support you (distinguish between family, friends, colleagues, managers, and customers/clients). What would they think about crafting for sustainability?	My friends care about this a lot and would consider this a point of improvement for the work that I do (aviation) My colleagues would be in favour because we work in the same environment
2	How could they support me?	My manager could function as a role model Customers could pay a better price and agree on longer shipping times
3–5	Please think of a concrete work situation that you will encounter today in which you can make a (little) change that will motivate sustainability behaviour at work. What I can craft today	This morning during work I take the bicycle instead of the scooter Combining the deliveries for next week Healthy working by using a desk with a bicycle seat (desk bike)
3–5	When I will do this	On the way to work and when visiting clients at home During the planning meeting Three times a day for 30 minutes
3–5	Possible barriers and how I can cope with them	Longer travel time, plan more time between clients The buying firm might not agree because they need certain supplies as soon as possible It could be that there are no desk bikes available so I am going to reserve my desk in time
3–5	Provide a detailed description of a situation at work today in which you made a (small) adjustment that stimulates sustainability behaviour (crafting for sustainability). What did you change in your work environment to make your work more sustainable today?	I printed the screening forms double-sided I took rest between appointments
3–5	How effective was it?	For me it was ok but my colleague did not like it. It reduced my stress.
3–5	How could you improve your crafting for sustainability action in the future?	Only print my own form double-sided. I find it important to go along with my colleagues. Next time I might do a short meditation exercise.

Hypothesis 4:

Daily seeking resources (4a), daily seeking challenges (4b), and daily optimizing demands (4c) will increase over time for employees participating in the self-training intervention while remaining stable for employees in the control group.

Our intervention’s primary aim is to stimulate proactive sustainability behaviour through job crafting. According to JD-R theory (Bakker et al., 2023), job crafting is a way to optimize employees’ pool of resources and demands which allows them to implement changes, such as increasing the level of sustainability, in one’s job (Young et al., 2015). To accomplish this, we introduce job crafting as a means to implement sustainability in daily work. In doing so, we build on the intervention design developed by Demerouti, Soyer et al., (2021) who also effectively framed the job crafting intervention as a means to engage in behaviour with uncertain and long-term outcomes (i.e., safety behaviour). In doing so, we instructed our participants to use the resources and optimization processes they gain through crafting to proactively engage more in sustainability (i.e., proactive sustainability behaviour). Next to monitoring more general proactive sustainability behaviour, we are also interested in the specific sustainable work activities in which this intervention may result (cf. Greaves et al., 2013). We consider this insight useful because (1) sustainability is known to be quite an abstract and complex concept that may mean different things to people (Kuhlman & Farrington, 2010) and because (2) the current job crafting intervention is first in targeting communal (versus individual) outcomes such as reduced pollution and energy usage (Oprea et al., 2019) and (3) implementation intentions create an advance link between a specific behaviour and a specific situation (Clement et al., 2014). Understanding which specific sustainable work activities are triggered may help to picture the potential and the boundaries of the job crafting intervention to stimulate modern complex outcomes of work (Tims & Parker, 2020).

Hypothesis 5:

The self-training intervention will lead to an increase in daily proactive sustainability behaviour (5a) and daily perceived sustainable work activities (5b) over time through an increase in daily seeking resources (5c), daily seeking challenges (5d), and daily optimizing demands (5e) over time.

Study 2

Materials and methods

Our self-training intervention design combines a 5-day diary study with a field experiment that contains an intervention group and a control group. This design allows capturing within-person fluctuations in job crafting and sustainability behaviours over days while at the same time manipulating the effect of the self-training at the between-person level (Hülsheger et al., 2015).

Participants and procedure

The procedure is similar to the diary study that we performed in Study 1. The students randomly assigned two of their participants to the control group and two to the intervention group. The diary booklets of the intervention group also contained the intervention exercises.

In total, 138 valid diary booklets were returned (n = 70 in the intervention group and n = 68 in the control group). The 138 employees that participated had a mean age of 34.5 years (SD = 14.0). The sample contained 60 women, 77 men, and 1 participant who preferred not to indicate its gender identity. Most participants were employed in the industrial or construction sector (27.7%), the healthcare or educational sector (22.6%), or in financial or business services (16.0%). Their average working hours per week were 37.2 (SD = 6.4), their average tenure was 6.7 (SD = 10.3) years, and 29.2% of the participants had a supervisory role. A sensitivity power analysis has shown that the sample size of 138 would allow us to detect a small-sized difference between the conditions over time (f = .12, two-tailed test, alpha .05) with 80% power.

The self-training intervention

The self-training consists of three key elements. The aim of the first element is to activate employees’ positive attitudes towards crafting for sustainability and to explore the subjective norm and potential social support of significant others’ regarding this behaviour. The behavioural change techniques (cf. Michie et al., 2013) used to accomplish this are (i) demonstrating the value of job crafting to engage in sustainable work behaviour to make the positive consequences more salient, (ii) encouraging participants to list pros and cons of crafting for sustainability and by having them reflect on past success experiences, and (iii) identifying significant others’ perceptions, expectations and potential support towards this behaviour. This element completely focuses on stimulating the internal motivation towards crafting for sustainability. The second element aims to enhance participants’ sense of control (i.e., perceived behavioural control) over their crafting for sustainability behaviour and to support the implementation of sustainability intentions. Hence, it helps to transfer internal motivation into real behaviours (Stern, 2000). The behavioural change techniques that are used here are setting daily crafting for sustainability goals and creating a detailed action plan for this (what they plan to craft and when they plan to do this). To further support the implementation of these crafting for sustainability goals, participants are asked to list possible barriers and to describe how they could cope with these barriers in a coping plan (cf. Michie et al., 2013). The third element is a self-monitoring exercise. Participants are asked to describe what they changed in their work situation to engage more in sustainability that day. They are asked to reflect on their crafting action’s effectiveness and describe how it could be improved in the future. This element intends to boost the learning and internalization process as it emphasizes participants’ ability to self-regulate their newly learned behaviour (Bandura, 1989) and it may help to cross the TPB intention-behaviour bridge (Ajzen, 2015). Hence, the general set-up of the self-training is that the participants read their daily exercise in the morning (i.e., reflection exercises and formulating SMART goals as of day 3), and they will be asked whether they reached their goal in the evening. Tables 3 and 4 provide a detailed overview of the content and structure of the self-training including the targeted theoretical constructs, the behavioural change methods used in the specific exercises, prior job crafting interventions applying similar exercises, and the literal instructions of the daily crafting for sustainability exercise.

Measures

Pre- and Post-Measures

The pre-measures (T0) were filled out before the daily diaries, and the post-measures (T1) were part of the final afternoon diary that participants completed on day 5 of the data collection period. The pre- and post-measures included all TPB constructs assessed at the general level (Swaim et al., 2016).

Daily diaries

The format and content of the measures included in the daily diaries (including seeking resources, seeking challenges, optimizing demands, and proactive sustainability behaviour) were similar to Study 1. To evaluate the effect of the intervention in more detail, we also asked participants to indicate how sustainable they would rate their work activities in specific sustainability domains (i.e., perceived sustainable work activities).

Daily perceived sustainable work activities

Daily perceived sustainable work activities were measured in the afternoon with a rating scale consisting of activities in four specific sustainability domains, namely (i) pollution, (ii) energy usage, (iii) transport, and (iv) health, safety, and well-being, that was developed for the current study. Participants were provided with the labels of these domains and asked to rate their associated activities on a 7-point scale ranging from (1) “not at all sustainable” to (7) “very sustainable”. Each domain label was accompanied by examples such as two-sided printing (pollution), switching off devices (energy usage), bicycle usage (transport), and taking micro-breaks (health, safety, and well-being). To validate this measure, we conducted multilevel exploratory factor analyses in Mplus (Muthén & Muthén, 1998–2017). These analyses showed that a model with one latent within-level factor and one latent between-level factor on which all four sustainability domains load showed a good fit to the data, χ² = 4.50, df = 4, p = .343, χ² /df = 1.13, CFI = 0.99, TLI = 0.99, RMSEA = 0.02, SRMR = 0.02 (within-person level) and 0.04 (between-person level). Moreover, all individual items loaded significantly on the within-level and between-level factors.

Analytical approach

We analysed our data based on the intention-to-treat (ITT) and the per-protocol (PP) principle (Ranganathan et al., 2016). The ITT analyses compare participants in the originally randomly assigned conditions, irrespective of non-compliance or dropout. This analytical approach has two main advantages; (i) it safeguards the principle of randomization, and (ii) it provides a realistic estimate of the effectiveness of the intervention in practice since non-compliance and dropout are also common there. The PP analyses exclude participants who deviated from the protocol, meaning that only participants who showed high compliance are analysed. Therefore, PP provides an estimate of the effectiveness of the intervention under ideal circumstances (in theory).

To examine the self-training intervention’s effect on the pre- and post-measures (hypothesis 3), we performed 2 × 2 repeated measures ANOVAs. Partial eta-squared (η²) values were used to estimate the effect sizes. We classified values between 0.01–0.06 as small effects, 0.06–0.14 as medium effects, and ≥ 0.14 as large effects. Using Cohen’s d, within-group effects were estimated with values between 0.2–0.5 representing small effects, 0.5–0.8 representing medium effects, and ≥ 0.8 representing large effects (Cohen, 1988).

We used structural equation-based latent growth curve analyses in AMOS to examine the change in our daily dependent variables over time and how these change trajectories were a function of condition (coded as 0 = control group and 1 = intervention group; Curran et al., 2004). Following Raudenbush’s (2001) recommendations, we modelled the change in our dependent variables over time according to our self-training intervention design. Specifically, the self-training intervention structure was such that participants were explicitly instructed to start changing their behaviours as of day three when they were encouraged to achieve their (self-set) crafting goals on sustainability (see Table 3). Therefore, we fixed the factor loadings of the latent slope factor to 0 0 1 2 and 3 for the five consecutive days, respectively. We also fitted linear growth curves to our data for statistical completeness reasons, which resulted in comparable model fits (results of these linear growth curve analyses are available upon request). We first fitted an unconditional latent growth curve model to the data for every outcome. To test hypotheses 4 and 5a-b, we consequently fitted a conditional latent growth curve model to the data in which we regressed the latent intercept and latent slope factor on condition. These conditional models’ results can be interpreted as follows: A significant regression of the latent slope factor on condition indicates that the change trajectory of a variable varies as a function of condition (intervention versus control group), which would support the hypotheses. A significant regression of the latent intercept factor on condition indicates that the intervention and control group have different T1 (day 1) scores on the respective variable, suggesting that randomization failed.

To test for mediation (hypothesis 5c-e), we specified parallel process latent growth models for mediation (Cheong et al., 2003) in which the slope of the job crafting strategies functioned as a mediator between the intervention and the slope of the sustainability behaviour outcomes (i.e., a longitudinal slope-only mediation model). Consequently, we used 95% bias-corrected bootstraps (based on 5000 samples) to estimate the CI’s of the indirect effects (von Soest & Hagtvet, 2011).

Results

Sample characteristics

Multivariate analyses and Chi-square tests revealed that participants in the two conditions were demographically equal. There were also no significant differences between conditions on the pre-intervention measures or the daily outcome variables at baseline (day 1).

Compliance

To inspect whether participants in the intervention group had complied with instructions to do the daily intervention exercises, we looked at their qualitative answers (see Table 4 for examples of the daily crafting sustainability goals that participants had set). On average, participants completed 4.57 (SD = 1.04) of 5 daily intervention exercises. There were, however, 13 participants who did not complete any of the intervention exercises or who only completed 1 (n = 1) or 2 (n = 3) exercises. We identified these careless responses by scanning for blank answer boxes or answers like “I did not engage in today’s exercise”.

The 13 participants who did not complete or discontinued the intervention were similar to completers concerning demographics and daily outcome variables at baseline (day 1). Still, they scored lower on the pre-intervention perceived behavioural control towards sustainability measure. These 13 participants are included in the ITT analyses (N = 138) and excluded in the PP analyses (N = 125).

Descriptive statistics

Means, standard deviations, Cronbach’s alphas, intra-class correlations, and correlations among the study variables are displayed in Table 5.

Table 5. Means, standard deviations, Cronbach’s alphas, intraclass correlations, and correlations among the study variables in the self-training intervention.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
Pre-measures
(1) Attitude
(2)Subjective norm	.49***
(3)Perceived behavioural control	.31***	.24**
(4)Behavioural intention	.67***	.47***	.50***
Post-measures
(5) Attitude	.60***	.36***	.36***	.48***
(6)Subjective norm	.48***	.67***	.15	.36***	.49***
(7)Perceived behavioural control	.21*	.15	.54***	.30***	.45***	.27**
(8)Behavioural intention	.49***	.31***	.25**	.57***	.66***	.50***	.54***
Day level
(9)Seeking resources	.20*	.16	.15	.30***	.31***	.31***	.26**	.42***		.28***	.17***	.14***	−.00	−.01	.02	.01	−.01
(10)Seeking challenges	−.10	−.01	.01	.05	−.02	.08	−.02	.12	.54***		.18***	.23***	.10*	.00	.08*	.06	.12**
(11)Optimizing demands	−.05	−.04	−.03	.01	−.08	.07	.15	.18*	.54***	.51***		.32***	.11**	.02	.04	.10*	.15***
(12)Proactive sustainability behaviours	.26**	.27**	.27**	.39***	.34***	.40***	.49***	.59***	.49***	.43***	.52***		.32***	.20***	.23***	.18***	.23***
(13)Perceived sustainable work activities	.33***	.36***	.26**	.37***	.33***	.36***	.30***	.36***	.34***	.20*	.10	.35***		.72***	.73***	.65***	.56***
(14) Pollution	.26**	.33***	.29**	.35***	.25**	.35***	.29**	.34***	.27**	.10	.06	.32***	.86***		.42***	.27***	.17***
(15)Energy usage	.19*	.29**	.23**	.36***	.25**	.23**	.27**	.29**	.32***	.20*	.15	.32***	.76***	.65***		.28***	.27***
(16) Transport	.23**	.14	.04	.20*	.24**	.18*	.11	.23**	.20*	.13	−.01	.12	.71***	.42***	.32***		.11**
(17)Health, safety, and well-being	.32***	.37***	.27**	.25**	.24**	.35***	.28**	.23**	.26**	.20*	.14	.37***	.71***	.59***	.40***	.25**
M	5.51	4.38	5.18	4.89	5.45	4.62	5.25	5.15	4.52	2.85	4.46	3.86	4.88	5.10	4.70	4.81	4.90
SD	1.00	1.20	1.02	1.15	0.98	1.38	1.17	1.19	1.26	1.47	1.26	1.30	0.94	1.17	1.15	1.50	1.14
α ^a	0.77	0.89	0.74	0.84	0.74	0.93	0.91	0.86	0.77	0.91	0.90	0.86
ICC									0.58	0.68	0.59	0.62	0.58	0.44	0.49	0.58	0.50

Notes. Correlations at the within-person level are displayed above the diagonal (N = 690; centred to the person mean), whereas correlations at the between-person level are displayed below the diagonal (N = 138; averaged across the five days).

^aCronbach’s alphas for the daily variables were averaged across the five days.

*p < .05; **p < .01; ***p < .001.

Hypotheses testing

Table 6 shows the results of the 2 × 2 repeated measures ANOVAs. Post-intervention, the intervention group reported an increased behavioural intention towards sustainability in general compared to the control group (small between-group effect, medium within-group effect). There were no significant differences between conditions for the other pre-post comparisons (i.e., attitude, subjective norms, and perceived behavioural control towards sustainability in general), which means that hypotheses 3a-c are rejected and hypothesis 3d is confirmed. Findings were similar for the ITT and the PP analyses.

Table 6. Means, standard deviations, and 95% CI’s of the means for participants in the intervention and control group of the theory of planned sustainability behaviour constructs pre- (T0) and post- (T1) intervention.

Outcome	Intention-to-treat (N = 138)							Per-protocol (N = 125)
	T0		T1		Intervention effect^a			T0		T1		Intervention effect^a
	M	SD	M	SD	d ^b	F	η^2	M	SD	M	SD	d ^b	F	η^2
Attitude towards sustainability
Intervention group	5.51	1.10	5.50	0.99	−0.01	0.41	0.00	5.60	1.13	5.66	0.92	0.08	0.41	0.01
	[5.25, 5.77]		[5.26,5.73]					[5.30,5.90]		[5.42,5.91]
Control group	5.51	0.88	5.39	0.96	−0.12			5.51	0.88	5.39	0.96	−0.12
	[5.29, 5.72]		[5.16,5.63]					[5.29,5.72]		[5.16,5.63]
Subjective norm towards sustainability
Intervention group	4.47	1.16	4.80	1.36	0.33	0.86	0.01	4.56	1.21	4.98	1.30	0.48	2.06	0.02
	[4.20, 4.75]		[4.48,5.13]					[4.23,4.88]		[4.64,5.33]
Control group	4.26	1.25	4.42	1.39	0.14			4.26	1.25	4.42	1.39	0.14
	[3.96, 4.57]		[4.08,4.76]					[3.96,4.57]		[4.08,4.76]
Perceived behavioural control towards sustainability
Intervention group	5.20	1.04	5.34	1.15	0.13	0.80	0.01	5.33	0.94	5.60	0.96	0.24	2.17	0.02
	[4.95, 5.45]		[5.07,5.62]					[5.08,5.58]		[5.34,5.85]
Control group	5.16	1.00	5.14	1.20	−0.02			5.16	1.00	5.14	1.20	−0.02
	[4.91, 5.41]		[4.85,5.44]					[4.92,5.41]		[4.85,5.44]
Behavioural intention towards sustainability
Intervention group	4.81	1.18	5.29	1.13	0.54	5.49*	0.04	4.90	1.24	5.52	1.05	0.70	8.27**	0.06
	[4.53, 5.09]		[5.02,5.56]					[4.58,5.23]		[5.24,5.80]
Control group	4.94	1.11	4.99	1.24	0.04			4.94	1.11	4.99	1.24	0.04
	[4.67, 5.22]		[4.69,5.30]					[4.67,5.22]		[4.69,5.30]

Notes. 95% CI’s of the means are displayed between brackets.

^aRepeated measures ANOVAs on the pre- and post-measures with condition as between-subjects factor.

^bWithin group Cohen’s d. *p < .05; **p < .01.

The latent growth curve analyses are summarized in Table 7 (model fit indices) and Table 8 (parameter estimates). Hypothesis 4 predicted that daily job crafting would increase over time for employees participating in the self-training intervention while remaining stable for the control group. The ITT analyses revealed no significant regression of the latent slopes on condition for seeking resources (4a), seeking challenges (4b), and optimizing demands (4c). This means that the self-training intervention did not affect daily job crafting behaviour over time. The unconditional models for the three job crafting dimensions in the PP analyses showed a moderate to satisfactory fit to the data, which was improved in the conditional models (i.e., the self-training explained additional variance). Supporting hypothesis 4c, we found a significant regression of the latent slope on condition for optimizing demands, indicating that participants who followed the self-training intervention showed a steeper increase in daily optimizing demands than participants in the control group (Figure 3). There was no such pattern for daily seeking resources (4a) or daily seeking challenges (4b).

Figure 3. Changes in daily optimizing demands over time as a function of condition in the self-training intervention.

Note. Results from the PP analyses (N = 125).

Table 7. Model fit indices for the unconditional and conditional latent growth curve models predicting the daily outcomes in the self-training intervention.

	Intention-to-treat (N = 138)					Per-protocol (N = 125)
	χ^2	df	CFI	TLI	RMSEA	χ^2	df	CFI	TLI	RMSEA
Unconditional model
Seeking resources	19.47	10	0.97	0.96	0.08	17.82	10	0.97	0.95	0.08
Seeking challenges	27.44	10	0.97	0.95	0.11	26.59	10	0.96	0.94	0.12
Optimizing demands	13.73	10	0.99	0.98	0.05	20.63	10	0.97	0.95	0.09
Proactive sustainability behaviour	22.17	10	0.97	0.96	0.09	24.13	10	0.96	0.94	0.11
Perceived sustainable work activities	15.08	10	0.99	0.98	0.06	19.39	10	0.97	0.95	0.09
Pollution	7.60	10	1.00	1.00	0.00	12.95	10	0.98	0.97	0.05
Energy usage	19.81	10	0.96	0.94	0.09	25.28	10	0.93	0.90	0.11
Transport	12.30	10	0.99	0.99	0.06	12.27	10	0.99	0.99	0.04
Health, safety, and well-being	8.07	10	1.00	1.00	0.00	8.25	10	1.00	1.00	0.00
Conditional model
Seeking resources	19.49	13	0.98	0.97	0.06	17.86	13	0.98	0.97	0.06
Seeking challenges	28.94	13	0.97	0.95	0.10	27.92	13	0.97	0.95	0.10
Optimizing demands	15.38	13	0.99	0.99	0.04	23.13	13	0.97	0.95	0.08
Proactive sustainability behaviour	27.64	13	0.97	0.94	0.09	27.71	13	0.96	0.93	0.10
Perceived sustainable work activities	18.97	13	0.98	0.98	0.06	20.49	13	0.97	0.96	0.07
Pollution	9.02	13	1.00	1.00	0.00	13.84	13	1.00	0.99	0.02
Energy usage	20.68	13	0.97	0.95	0.07	27.37	13	0.93	0.89	0.09
Transport	12.74	13	1.00	1.00	0.00	13.00	13	1.00	1.00	0.00
Health, safety, and well-being	13.07	13	1.00	1.00	0.01	11.15	13	1.00	1.00	0.00

Notes. The unconditional model describes the development of the daily outcomes over time with a latent intercept and a latent slope factor. The conditional model includes condition (intervention or control group) as predictor of the latent intercept and latent slope factor of the daily outcomes.

Table 8. Parameter estimates and standard errors from the intervention predicting the intercept and slope factors of the daily outcomes.

		Intention-to-treat (N = 138)				Per-protocol (N = 125)
		Intercept factor		Slope factor		Intercept factor		Slope factor
Outcome variable	Predictor variable	Parameter	SE	Parameter	SE	Parameter	SE	Parameter	SE
Seeking resources	Condition	0.17	0.22	−0.06	0.07	0.29	0.21	−0.05	0.07
	Intercept^a	4.46***	0.15	0.01	0.05	4.46***	0.14	0.01	0.05
Seeking challenges	Condition	−0.11	0.26	0.03	0.06	−0.05	0.28	0.00	0.07
	Intercept^a	2.88***	0.19	0.03	0.05	2.88***	0.19	0.02	0.05
Optimizing demands	Condition	−0.16	0.23	0.08	0.06	−0.13	0.24	0.13*	0.06
	Intercept^a	4.44***	0.16	0.05	0.04	4.44***	0.16	0.05	0.04
Proactive sustainability behaviour	Condition	−0.10	0.23	0.08	0.07	0.08	0.24	0.09	0.07
	Intercept^a	3.75***	0.16	0.11*	0.05	3.76***	0.16	0.11*	0.05
Perceived sustainable work activities	Condition	0.03	0.17	0.07	0.05	0.12	0.17	0.09	0.05
	Intercept^a	4.71***	0.12	0.10**	0.04	4.70***	0.12	0.10**	0.04
Pollution	Condition	0.18	0.22	0.16*	0.08	0.28	0.23	0.18*	0.08
	Intercept^a	4.90***	0.16	0.01	0.06	4.90***	0.16	0.01	0.06
Energy usage	Condition	−0.12	0.21	0.10	0.07	−0.07	0.22	0.12	0.08
	Intercept^a	4.56***	0.15	0.12*	0.05	4.56***	0.15	0.12*	0.05
Transport	Condition	0.19	0.29	−0.13	0.09	0.26	0.30	−0.13	0.09
	Intercept^a	4.61***	0.20	0.15*	0.06	4.61***	0.20	0.15*	0.06
Health, safety, and well-being	Condition	−0.23	0.22	0.20**	0.07	−0.09	0.23	0.21**	0.07
	Intercept^a	4.79***	0.16	0.08	0.05	4.80***	0.16	0.08	0.05

Notes. Condition was coded as 0 = control group (N = 68) and 1 = intervention group (N = 70).

^aIntercept term for the prediction equation.

*p < .05; **p < .01; ***p < .001.

According to hypothesis 5, daily proactive sustainability behaviour (5a) and perceived sustainable work activities (5b) will increase over time for employees in the intervention group while remaining stable for the control group. The ITT analyses showed that the unconditional and conditional models revealed no significant regression of the latent slope on condition for proactive sustainability behaviour or perceived sustainable work activities (i.e., no effect of the self-training on sustainability behaviour over time). However, when we looked at the perceived sustainability of work activities in the specific sustainability domains, we found that condition could predict the growth trajectories for perceived sustainability activities in the pollution domain and the health, safety, and well-being domain (Table 8 and Figure 4 and 5). These results were mirrored in the PP analyses.

Figure 4. Changes in daily perceived sustainability regarding work activities in the pollution domain over time as a function of condition in the self-training intervention.

Note. Results from the ITT analyses (N = 138).

Figure 5. Changes in daily perceived sustainability regarding work activities in the health, safety, and well-being domain over time as a function of condition in the self-training intervention.

Note. Results from the ITT analyses (N = 138).

Hypothesis 5c-e predicted that the effect of the self-training intervention on the increase in daily sustainability outcomes over time could be explained by an increase in daily seeking resources (5c), daily seeking challenges (5d), and daily optimizing demands (5e) over time. To test this, we first inspected whether the intervention related significantly to an increase in the mediators (i.e., job crafting behaviours) over time, which was only true for optimizing demands in the PP sample. Based on these results, the hypothesized indirect effect could only be tested for optimizing demands in the PP sample. Mediation analyses showed no indirect effect of the intervention via the growth of optimizing demands on the growth of proactive sustainability behaviour (γ = 0.07, SE = 0.07, 95% CI [−0.00, 0.14]) or perceived sustainable work activities (γ = 0.10, SE = 0.06, 95% CI [−0.02, 0.21]) over time. However, we found that the increase in optimizing demands over time (5e) mediated the effect of the self-training intervention on the increase in perceived sustainable work activities in the two specific sustainability domains of pollution (γ = 0.20, SE = 0.08, 95% CI [0.03, 0.34]) and health, safety, and well-being over time (γ = 0.21, SE = 0.08, 95% CI [0.03, 0.36]).

Brief discussion of study 2

In our intervention study, we extended the findings from the diary study by testing a self-training intervention based on the TPB in which employees were trained in crafting as a means to engage more in sustainability. One important result is that the intervention group had an increased behavioural intention towards sustainability in general after the training compared to the control group. Providing evidence of causality, this means that the self-training intervention boosted employees’ sustainability intentions and suggests that its crafting exercises influenced the enactment of the TPB (cf. Costantini et al., 2019, 2022).

The ITT and PP analyses revealed a notable difference regarding daily job crafting: Compliant participants started to optimize their job demands more throughout the training than participants in the control group; however, this effect diminished when we also considered non-compliant participants. Although PP analyses typically show more substantial results than ITT analyses (Ranganathan et al., 2016), the difference between receiving and following our job crafting intervention may have been inflated by the self-training format that we used. Another important finding is that the self-training did not affect participants’ daily use of seeking resources or seeking challenges. This is in accordance with meta-analytic evidence showing that job crafting interventions have a more pronounced effect on reducing demands than on the other job crafting strategies (Oprea et al., 2019). A post-hoc explanation is that the daily exercises’ instructions did not specify which crafting strategy participants should practice, leaving room to pick the strategy that seems most practical and appealing to engage more in sustainability in daily work (Rupp et al., 2018).

We also examined whether the self-training stimulated daily sustainability behaviour and found no effect between conditions on proactive sustainability behaviour or the overall measure of perceived sustainable work activities. Yet, by looking more closely, we found that participants who received the self-training increased in their perceived level of sustainability for work activities in the pollution and health, safety, and well-being domain over time. Importantly, mediation analyses showed that this effect was mediated by the growth of daily optimizing demands over time. Hence, optimizing demands “activated” employee engagement in several specific sustainability activities rather than in sustainability in general.

General discussion

Based on the central role that employee motivation plays in the sustainable development of organizations (Swaim et al., 2016; Unsworth & McNeill, 2017), this article combines JD-R theory with the TPB to introduce job crafting as a behavioural tool to engage more in sustainability in daily work (“crafting for sustainability”). This perspective does justice to realistic behavioural challenges and obstacles (i.e., conflicting demands, costs, time; Merriman et al., 2016; Rupp et al., 2018) that employees who want to implement sustainability face in their work environment. Our diary study suggests that optimizing demands enables employees to translate their daily sustainability intentions into daily proactive sustainability behaviour thereby bridging the intention-behaviour gap of the TPB (Ajzen, 2015). Our intervention study replicated this pattern to a large extent by showing that participants who actively followed our self-training intervention increased in their daily level of optimizing demands and, in turn, in several perceived sustainable work activities. Furthermore, they had a stronger behavioural intention to engage in sustainability after the training than participants who did not receive the training. Together, our study findings make three important contributions. First, the findings contribute to the sustainability literature as they show that implementing sustainability in organizations does not only have to originate from corporate business strategies or top management decisions (Baumgartner & Ebner, 2010) but can also start with proactive efforts (i.e., optimizing demands) that employees make to balance their job demands and resources. Importantly, we show that employees’ favourable attitudes and perceived behavioural control towards sustainability play a role in this, and these could, potentially, be boosted by the organizational context. Second, in combining JD-R theory with the TPB we find that job crafting can function as a means to facilitate the daily intention-behaviour transfer and to cope with daily barriers (cf. Costantini et al., 2019, 2022). Herewith, we empirically show that “for a sustainability intention to lead to sustainability behaviour, employees must ensure to have the requisite resources and potential barriers to behavioural performance must be removed” (Ajzen, 2015, p. 134).

Third, the present findings also contribute to job crafting theory by showing that optimizing demands can have pro-environmental and pro-social consequences. Where former job crafting research predominantly focused on individual or organizational work outcomes (Oprea et al., 2019), we provide the first empirical evidence for the impact that employee job crafting has on communal outcomes such as implementing sustainability. This finding resonates with the recently introduced concept of responsible job crafting in which managers proactively engage in responsible activities, such as sustainability, to boost their self-image (Bizzi, 2020). Furthermore, linking the intervention exercises explicitly to the TPB components could disentangle which exercise (and TPB component) was most effective in stimulating crafting for sustainability. Specifically, the observed growth in the targeted behaviours as of day 3 (see Figures 3–5) suggests that the personal crafting plan and the daily self-monitoring exercise are the most substantial aspects of the training. Note that both exercises were meant to boost participants’ control beliefs, which is assumed to be critical in a context when most people already value the intended change (cf. Ajzen, 2015).

Both our studies exclusively reveal optimizing demands as the most profitable job crafting strategy to engage more in sustainability. The main hindrance that employees face when they are willing to engage in sustainability in daily work is that it is effortful and time-consuming (Aguinis & Glavas, 2012; Merriman et al., 2016). Hence, making sure that current demands take less time and become less straining may be the first necessary step towards working more sustainably. Through optimization, employees work more efficiently, which means that the free time and attention to invest in sustainability would otherwise have been used for cumbersome procedures or in-efficient administrative tasks.

Limitations and future research

In contrast to previous empirical research, we conceptualized sustainability using the People Planet and Prosperity framework (Elkington, 1997). However, this endeavour complicated the measurement of sustainable work behaviour because existing measures predominantly capture pre-environmental attitudes and behaviour (e.g., Bissing-Olson et al., 2013; Swaim et al., 2016). Although we carefully adapted items that solely referred to the environment using broader and more general terms (e.g., sustainability, natural and human resources), the qualitative answers to the intervention exercises clarified that most participants still conceptualized sustainability as referring to the Planet only. This may signal that the public understanding of sustainability is not in line with the triple P framework (cf. Kuhlman & Farrington, 2010), implying that our measures (and intervention exercises) may not have captured the intended wide-ranging nature of sustainability. Future research may select measures that more explicitly refer to sustainability’s social and financial aspects. An alternative, even superior, option could be to collect objective behavioural measures in these domains such as recovery time after work (People), water usage per day (Planet), or savings (Prosperity). Doing so may solve potential common-method bias in the data as well. Another issue related to measuring sustainable work behaviour is that our daily diary design may have excluded best practice situations that occur less frequently (e.g., speaking up in a monthly meeting, initiating a new project). Complementing daily measures with weekly or monthly assessments could, in that sense, provide a better proxy of the real impact that employees make.

In the present studies, we focused on employee engagement in sustainability. Notably, the sustainability literature emphasizes other factors that determine the level of sustainability in organizations. To illustrate, stakeholder pressures, regulation, organizational strategies, and top-management commitment have been listed as critical catalysts of corporate sustainability initiatives (Aguinis & Glavas, 2012). Former studies and interventions that did focus on micro-level processes zoomed in on the motivation of employees to engage in sustainability (e.g., Maiteny, 2002; Swaim et al., 2016; Unsworth & McNeill, 2017). Therefore, an important contribution of our study to the sustainability literature is that we show that employee engagement in sustainability did not increase when we stimulated participants’ internal motivation to craft for sustainability but when we encouraged them to engage in goal-setting and action planning and experiment with the new behaviour. Given that, we wish to draw attention to other processes that may specifically impact employee engagement in sustainability – namely, the role of leadership. In times of organizational change, such as implementing sustainability, leaders act as change agents. Effective leadership styles that facilitate and support organizational change processes among employees are transactional and empowering leadership (Amundsen & Martinsen, 2014; Holten & Brenner, 2015). Especially in the context of sustainability, which often implies more effortful and costlier work procedures, employees need this kind of support, resources, and vision from their managers to integrate it proactively into their daily work. Future research may extend our scope and investigate these and other potential organizational or institutional top-down processes that may influence the (bottom-up) engagement in sustainability among employees.

To our knowledge, our paper contains the first experimental field study that investigated the effectiveness of a self-training intervention in crafting for sustainability. The first test of a new intervention allows learning from its growing pains. One critical drawback is that we used a general job crafting measure (without a purpose in the items) while we trained participants to craft for sustainability. Hence, we cannot be sure that participants targeted their crafting behaviour towards the sustainability domain. This might be a reason for the weak relation between (some of the) job crafting strategies and the outcomes in Study 2. Although we chose to use the validated scale of job crafting instead of an adjusted, unvalidated version, we think that this general scale may have led to less significant relationships and thus to a more conservative test of the effectiveness of our intervention. Future research may consider creating an adapted version of the job crafting scale to improve monitoring of the trained crafting for sustainability behaviours. Related to this, we also realized that our participants’ intervention exercises and feedback revealed that the provided written information is insufficient to stimulate the wide-ranging and multifaceted nature of job crafting and sustainability outcomes. This may have caused the limited variety in crafting strategies (i.e., predominantly optimizing demands) and a predominant focus on the Planet component of sustainability (see Table 4 for some examples). This aspect may be improved by integrating illustrative examples, explicit instructions, or short videos in the training, making the “message” more engaging.

A potential post-hoc explanation for the limited effects of our self-training could be that it targeted too many aspects in a short period of time. That is, our intervention first tried to activate sustainability attitudes and the perceived subjective norm thereof and consequently prompted goal-setting and action planning. Although this is in line with establishing behavioural change according to the TPB (Ajzen, 2015) and has proven to be successful in a recent job crafting intervention (Costantini et al., 2022), stimulating both participants’ internal motivation and behavioural implementation at the same time may be too ambitious for such a complex outcome as sustainability behaviour. We learned from linking the job crafting exercises with the TPB components that especially the second phase of the training in which participants had to make a personal crafting for sustainability plan and complete daily reflections on successes and learning points (i.e., targeting perceived behavioral control and implementation intentions) stimulated a growth in the trained strategies and some outcomes. To improve the training’s effectiveness, we suggest to focus on these aspects from the start.

Another crucial difference between our intervention and previous job crafting interventions is that participants were encouraged to set a crafting for sustainability goal instead of a crafting goal that would benefit their own well-being or performance. We did this because the TPB postulates that people need to first set an intention before they can and will engage in new behaviour (Ajzen, 2015). Although we carefully tried to boost their motivation to engage more in sustainability in the first days of the training, some participants may potentially still have experienced the crafting for sustainability goals as being more extrinsically motivated. This issue relates to the expectancy-value component of their attitude by placing the expectation of an additional demand (Ajzen, 1991). Possible ways to improve this aspect could be to link the crafting for sustainability goals more to participants’ pre-existing (sustainability) intentions (cf. Unsworth & McNeill, 2017), to frame the crafting for sustainability goals more as challenges (vs. demands), and to increase the salience of potential positive outcomes for them. A final difficulty is that we had limited control of our participants’ compliance, possibly resulting in substantial differences between our ITT and PP analyses. We could also not provide participants with feedback on their personal crafting for sustainability goals, which may have contributed to the (aforementioned) limited variety in activities. This is a severe limitation of our self-training intervention because such interactive elements enhance the effectiveness of job crafting training (Demerouti et al., 2019). Future research may seek ways to establish more interaction with or between participants, such as using an app or building an online community. A final important question that future research should address is how long-lasting the effects of the intervention are. For example, to support the long-term learning of the training to daily work routines, it may be worthwhile to invest in (self-initiated) reminders.

Practical implications

The present studies hold some important practical implications for individual employees and organizations. The findings are relevant for employees who experience pressures to work more sustainably. For many, the additional demanding, complex, and at the same time important aspects that this mission adds to their work agenda seem hard to combine with their limited time and resources (Davis & Challenger, 2015). Instead of being overwhelmed by this, employees could learn that they can already engage more in sustainability by making small daily changes in their work environment. They can proactively create opportunities to make more impact by using practical strategies such as simplifying work procedures, delegating work tasks, or letting go of perfectionistic standards. This research also responds to jobseekers’ emerging drive to seek jobs with impact (Cone Communications, 2016). Next to the responsibility of employers to contribute to a better world, employees could realize that these opportunities exist in any job as long as they initiate action to facilitate this themselves.

Our findings suggest that next to top-down initiatives aiming to influence employees’ attitudes, subjective norms, and control beliefs concerning sustainability, organizations may benefit from and invest in the bottom-up initiatives employees take to implement sustainability (cf. Davis & Challenger, 2015). Employees’ crafting behaviour can help to embed sustainability across all layers and functions of an organization, which, in turn, may boost the overall sustainability level of the organization. To stimulate proactive sustainability behaviour among employees, organizations may consider training employees in crafting for sustainability. The self-training intervention that we developed provides a first hands-on and cost-effective option to enhance crafting for sustainability skills by encouraging employees to set SMART daily crafting for sustainability goals, work on these goals, and monitor their progress. Yet, this method’s effectiveness may increase when organizations ensure that the conditions under which employees are trained in crafting for sustainability are supportive (cf. Petrou et al., 2012). For example, managers could encourage employees to experiment with crafting for sustainability, focus on the learning process, and assist them in combining daily crafting for sustainability goals with other organizational objectives (Demerouti et al., 2019).

Conclusions

Engaging in sustainability is a complex task that must be shouldered throughout an organization. Therefore, examining ways that help employees deal with this issue is valuable. The current research suggests that job crafting supports employees in translating their sustainability intentions into daily work behaviour. By optimizing their job demands, employees create time and energy to invest in sustainability. Furthermore, we provide initial evidence that crafting for sustainability can be influenced by employing a relatively short self-training intervention. These insights deepen and combine actual knowledge of sustainability management and job crafting.

Disclosure statement

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

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Additional information

Funding

The work was supported by TKI DINALOG .