English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Introduction
In the last decades, studies on passion have been extended to different areas, especially research on passion in work environment (Forest et al., 2012). According to the Passion for Work Assessment Model (MAPT. Modelo de Avaliação da Paixão pelo Trabalho) (Zigarmi et al., 2012), the referred construct consists of a manifestation of well-being at work, which mediates the relationship of affections and work cognitions (individual antecedents) with work intentions (individual consequences). The model also predicts that the organizational and work characteristics constitute distal contextual antecedents of such a process.
MAPT-based studies are still scarce (Egan et al., 2019: Peyton & Zigarmi, 2021: Shuck et al., 2015: Zigarmi. et al., 2018): they are generally cross-sectional studies (Roberts & Zigarmi. 2014: Shuck, et al., 2015), and they have focused primarily on the individual antecedents of passion for work (such as locus of control and cynicism). However, a few have focused the analysis of passion contextual background, such as. for example, self-concern and guidance from the leader (Zigarmi & Roberts. 2012). However, no studies were found concerning the adoption of servant leadership as a contextual distal antecedent of passion for work.
However, the servant leadership has received increasing attention from scholars, possibly in response to the finding that the behavior of certain leaders has been the main source of ethically questionable decisions, unsustainable business practices and abusive behavior towards followers (Liden, 2010). In this connection, servant leadership can contribute to the explanation of the positive effects of leadership (van Dierendonck & Xuijten.2011), since servant leaders are committed to the growth of their employees, while maintaining the focus on the development of the organization and responsibility towards the community (Reinke, 2004).
In addition, servant leaders often support, empower and encourage their immediate followers, thus helping them to succeed in their work and solving organizational problems (Liden et al., 2008). This is why, in the present investigation, servant leadership was adopted as an antecedent of work passion.
Also according to the Passion for Work Assessment Model (Zigarmi et al., 2012), the employees' positive or negative intentions about their own work (Zigarmi et al., 2015) are the consequence of passion. However, studies about this relationship are still scarce (Zigarmi et al., 2018), despite the fact that work intentions are characterized as an important concept in the attitude-intention-behavior chain (Webb & Sheeran. 2006). Additionally, employees who report more favorable affective states are more likely to striive and stay in then organization (Roberts & Zigarmi. 2014).
Although there are already studies focusing on the relationship between servant leadership and work intentions (Kg et al., 2016: Zigarmi et al., 2015), investigations on the processes underlying these relationships have not been found. In view of these considerations, the present investigation had the general objective of evaluating in a longitudinal study, the relationship between servant leadership and work intentions, mediated by work passion.
This investigation's relevance is justified by applying, for the first time in Brazilian samples, a recent and expanding model. In this way, it can contribute to a better understanding of organizational dynamics, with regard to processes related to passion for work, as well as assisting human resources professionals, improving the employees' quality of work life (Zigarmi et al., 2015). In addition, it may prove useful for carrying out interventions designed to increase employees' passion and bring long-term benefits to organizations (Panaccio et al., 2015).
Servant leadership and work intentions
Servant leadership concerns the motivation of the leader guided by a natural feeling of serving his own team (Greenleaf, 1977). Thus, servant leaders strive for then work goals, but they also turn to their followers (McNeff & Irving. 2017). In this way, their concerns go beyond the organization itself, insofar as they seek to foster the well-being of their employees, customers, other interested parties and society in general (Barbuto & Wheeler, 2006).
According to van Dierendonck and Nuijten (2011), the servant leader has eight characteristics: empowerment (Promoting a proactive and self-confident attitude among followers): humility (Putting the interest of followers first): authenticity (expression of the "true self", consistent with internal thoughts and feelings); interpersonal acceptance or forgiveness (ability to understand and experience the feelings of others): accountability (ability to hold people accountable for performance): courage (ability to dare and take risks): support (ability of the leader to put the interest of others first): servitude (willingness to take responsibility for the institution and act as a model).
The present work evaluated the impact of the servant leader's perception on the work intentions, which are mental images of the behavior that the employee plans to manifest in his work environment (Bagozzi. 1992). They unfold hi five dimensions: intention to use discretionary effort (intention to make efforts in favor of the organization, which go beyond the formal requirements of the position): intention to perform (intention to do one's work at an above-average level): intention to endorse (intention to endorse the organization to others as a good place to work): intention to stay (intention to stay in the organization): intention to be an organizational citizen (intention to behave in a respectful, attentive and sensitive manner with others) (Nimon & Zigarmi, 2011).
The intentions result from the work assessment process (Zigarmi et al., 2012), which encompasses the selection that the individual makes between alternative results and the possibilities of action that will eventually shape his/ her future (Lazarus, 1984). They derive, therefore, from the need to solve problems and from positive or negative assessments of well-being at work (Nimon & Zigarmi. 2015).
The Social Cognitive theory (Bandura. 1986) and MAPT (Zigarmi et al., 2012) suggest that individuals have the ability to reflect, through self-awareness about the social environment that sur rounds them. This allows them to review their experiences and think about their own thought processes. By reflecting on their different previous experiences and evaluating them, they can derive generic knowledge about themselves and their work, which allows them to monitor then future ideas and predict the occurrence of events in the workplace.
In other words, such an assessment process allows individuals to decide on their intentions. Applying this line of thought to the servant leadership, it is expected that employees, when evaluating the behavior of their leader in a positive way, will also manifest intentions of positive behaviors towards their organization.
In a congruent way with such theoretical foundations, it has been verified that the perception of servant leadership usually leads to the increase of several positive behaviors and the reduction of several negative behaviors (de Waal & Sivro, 2012). In this connection, empirical studies have shown that the perception of servant leadership tends to increase levels of organizational performance (Panaccio et al., 2015), job satisfaction and intentions to stay in the organization (Hajjaj, 2014), as well as reducing turnover intentions (Mitterer, 2017: Turgut et al, 2017).
In summary, servant leaders usually promote fairer and more humane workplaces, in addition to being more sensitive to social expectations (Gotsis & Grimari. 2016). Therefore, they end up acting as facilitators of work intentions, since they also consist of positive behaviors intended to benefit the work framework and the organization. Therefore, based on the Passion for Work Assessment Model, the Servant Leadership Model and the empirical studies mentioned, the following hypothesis was formulated:
H1: Servant leadership relates positively and significantly with work intentions, over six months.
The mediating role of harmonious passion for work
The passion for work consists of an inclination towards own work, which leads the employee to invest time and energy in doing it (Vallerand & Houlfort, 2019). Work passion stems from cognitive and affective assessments of then own work experiences, which can be positive or negative (Zigarmi et al., 2012). Such a dualistic perspective thus indicates the presence of two distinct types of passion: harmonious passion and obsessive passion.
Harmonious passion is revealed in the situations in which the work and its results are congruent with the pre-existing values and objectives of the self (Vallerand & Houlfort. 2019), which is why activities can be chosen freely and without contingencies. Employees do not feel obliged to do then job, but, on the contrary, work awakens in them a sense of identity and pleasure. Obsessive passion, on the other hand, emerges from the non-integration of work and its results into the employee's signature, which thus conflicts with its pre-existing goals and values (Vallerand & Houlfort, 2019). In the present study, we chose to consider only harmonious passion, since only this type of passion is a positive indicator of job well-being, which affects the performance of work activities and provides individuals with feelings of pleasure, satisfaction and vitality (Curran et al., 2015).
According to the Passion for Work Assessment Model (MAPT) (Zigarmi et al., 2012), personal and organizational phenomena, such as the case of servant leadership, expressed in experiences and in perceptions that the worker has of his work environment, they constitute antecedents of passion for work (Zigarmi et al., 2015). The model also predicts that the harmonious passion for work has several consequences, such as, for example, positive intentions related to the work context.
In this way, workers with higher levels of harmonious passion for work tend to manifest more often their intentions to stay in the organization, to speak well of their work to other people and to comply with organizational rules (Zigarmi et al., 2012: Zigarmi et al., 2015). This is because when individuals have more energy and dedication in then tasks, they consequently have more positive attitudes and behaviors in favor of their organization.
MAPT proposes, finally, that the passion for work mediates the relationship between its antecedents and its consequences (Nimon & Zigarmi. 2015: Zigarmi et al., 2012). In this connection, contextual resources help the individual to develop the perception of passion for work, which, in turn, facilitates work results. Therefore, passion for work is a link between the context of work and its results. Therefore, it is considered that the employees' perceptions about the environmental and contextual factors that surround them trigger affective and cognitive assessments that materialize through well-being indicators, such as the harmonious passion for work. Such passion for work will then positively influence employees' intentions (Zigarmi & Roberts, 2012). Therefore, it is possible to assume that servant leadership, as a factor in the individuals' work environment, is assessed by them affectively and cognitively as positive, which would awake their harmonious passion for work and. consequently, impact their work intentions.
In line with these assumptions, recent studies have pointed out that harmonious passion mediates the relationship of the leader's self-concern and guidance (factor of the work framework) with work intentions (Zigarmi et al., 2018). Similarly, Robertson and Barling (2013) identified indirect effects on the relationship between transformational leadership and organizational behaviors, through harmonious passion for work. Therefore, based on the Passion Assessment Model, as well as on the empirical studies cited, the following hypothesis was formulated:
H2: Harmonious passion (T2) acts as a mediator of the positive relationship between servant leadership (T1) and work intentions (T3).
Method
Longitudinal Study
Longitudinal studies allow a temporal analysis of subjective factors in an investigation overtime. Longitudinal analysis therefore provides answers to the gaps in most cross-sectional studies, enabling a broader understanding of the antecedents and consequences of a construct. In addition, the longitudinal study provides a complex view of the study models, indicating patterns of stability and change over the lifetime of individuals (Liu, 2016).
Participants
The total sample consisted of 479 workers of different companies and from all Brazilian regions, coming mostly from the State of Rio de Janeiro (59.7%). It is noteworthy, most workers are from two large companies in the educational sector. Participants were of both genders (70.4% women), with a minimum age of 18 years and a maximum age of 85 years (M = 37.4: SD = 11.5) and the majority was married (50.1 %). Regarding education. 58% had completed higher education, and out of these. 50.3% had attended a postgraduate course. With regard to organizational variables. 46.4% made between 1 and 3 minimum wages, most working in the private sector (56.9%): they had been doing the current job for an average of 7.83 years (SD = 12.6) and had been working for a total of 16.3 years (SD = 26.4). As a criterion for sample selection, the worker should be participating in a team of at least three members, who had a direct leader. In the present sample, such teams were composed of 3 to 10 members (39.7%), 11 to 50 members (41.3%) and more than 50 members (19%).
Data collection and analysis procedures
The data collection was carried out on-line, upon prior agreement with those willing to participate in the investigation, who filled out the Free and Informed Consent Form (ICF). As it is a longitudinal study, the collection was performed in three waves (T1, T2 T3), with 3 months lag between them, following Taris and Kompier's (2014) guidelines. The authors indicated that when there is no robust explanation for the choice of time, then a smaller interval between the waves should be used. Furthermore, it was the same interval used by Lavigne et al. (2014).
Participants were contacted via email or WhatsApp, and the sample was not-probabilistic. Participants hi T1 were 479 workers, in T2. 47.39% of the total sample (N = 227) adhered, while T3 counted on 27.34% of the total sample (N= 131). The rationale pointed out by the participants for withdrawing from the investigation was leaving the company (41.2%). dismissal (42.8%). changing job (10.7%). transferring (3%), maternity leave (1.3%) and retirement (1%).
The data were organized in a database in wide format, with the data of each participant arranged in a single line and with all the variables grouped in columns. Initially, the assumptions for confirmatory factor analysis were verified, such as the variables are categorical and non-normal distribution. Then the Confirmatory Factor Analysis of each instrument was performed, based on the total sample collected in wave one, through Structural Equation Modeling, using the software Mplus v.8, with the parameters being estimated by the WLSMV estimator (Weighted Least Squares Mean and Variance Ajusted), in order to verify the structure of each scale for the sample collected. The missing values were processed with the general procedure of the missing type (i.e. the models were estimated using all available data) (Kinnunen et al., 2019). Correlations were estimated between latent factors, inside the structural equation model.
Then, the measurement model was verified using Confirmatory Factor Analysis, initially testing the model of three distinct and correlated factors, in order to ensure that the three latent variables were represented by their relevant items. After CFA, factor scores were saved and used to estimate the longitudinal models. The internal consistency of the instruments was analyzed using the Composite Reliability Coefficient (Omega), calculated using the Composite Reliability Calculator (Colwell. 2016).
To investigate the invariance of the measures over the three waves, we tested three models, which progressively fixed the number of items and factors (configural invariance - M0), the factor loads (metric invariance - M1) and the intercepts (scalar invariance - M3). Hie tested model should not present differences hi CFI (ΔCFI)> 0.01, Gamma-Hat (ΔGamma-hat)> 0.001 and McDonald's NCI (ΔMcDonald’s NCI)> 0.02, when compared to previous models (Damasio. 2013). The demonstration of the structural invariance of latent variables over time allowed us to separate the effect of the growth of eventual differences in parameterization of the items. Then, the structural equation modeling with cross-lagged panel design (ACLD) was adopted, to assess the temporal association between the study variables, in other words, check how the variables interact with each other over time (three waves).
In the mediation analysis, the relationship between the independent variable (servant leadership) and the dependent variable (work intentions) was initially verified, which had to be significant for the confirmation of the model. In the second step, the mediating variable (harmonious passion for work) was inserted into the model. The model is considered mediational if there is a reduction in the strength of the relationship between the predictor variable and the dependent variable, with the entry of the mediator variable, and if the indirect effect of the independent variable on the dependent variable is significant.
To assess the fit of the models to the data, the following indicators were considered: Comparative Fit Index (CFI). Tucker-Lewis Index (TLI), Root-Mean-Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR) (Brown. 2015). Values below 0.06 for RMSEA and SRMR, and values greater than 0.95 for CFI and TLI were adopted as good adjustment indexes (Brown, 2015).
Ethical Procedures
The research was submitted to the Research Ethics Committee of the authors ' institution and approved under CAAE No. 17070819.7.0000.5289. To complete the questionnaire, prior agreement on participation in the research was required by completing the Informed Consent Form (ICF).
Instruments
Servant leadership was assessed using the van Dierendonck and Nuijten's (2011) Server Leadership Scale (SLS), validated for the Brazilian context by Pereira and Ferreira (2019). It consists of 28 items, distributed in eight dimensions, to be answered on a six-point Likert scale, ranging from 1 (Never) to 6 (Always). Example of one item: "My immediate boss learns from criticism''. In the Server Leadership Scale, the model's adjustment indexes were X 2(gl)=600.10(397): CFI=0.94: TLI=0.93; RMSEA=0.05: in the Brazilian version they were X 2(-gl)=247.266 (322): CFI=0.95: TLI=0.94: RMSEA=0.05, and in this study they were X 2(gl)=125.70 (30); CFI=0.97: TLI=0.96: RMSEA=0.08. The Composite Reliability Coefficient was excellent (0.92).
Passion for work was assessed by the Vallerand and Houlfort Scale of Passion for Work (2003), adapted to Portuguese by Pereira et al. (2018). It consists of 14 items, distributed in two dimensions: harmonious passion and obsessive passion. In the present study, however, only the dimension of harmonious passion was used. It presents seven items (example of one item "My work is hi harmony with the other activities of my life"), to be answered on seven-point Likert scales, ranging from totally disagree (1) to totally agree (7). In the Scale of Passion for Work, the model's adjustment indexes were X 2(gl) = 171.70(76): NNFI=0.91: CFI=0.93: RMSEA=0.73; in the Brazilian version they were X 2(gl)=254.03 (64);TLI=0.94;CFI=0.96: RMSEA=0.08, and m this study they were X 2(gl)=66.13 (II): CFI=0.99:TLI=0.99:RMSEA=0.06. The Composite Reliability Coefficient was excellent (0.90).
Work intentions were assessed using the reduced version of the Work Intent Inventory (Nimon & Zigarmi, 2015), validated for the Brazilian context by Pereira and Ferreira (2020). This version consists of 15 items, distributed in five dimensions, to be answered on a 6-point Likert scale, ranging from 1 = no extension to 6 = throughout the extension. Example of one item: "I intend to speak positively about leadership hi my organization". In the Work Intent Inventory, the model's adjustment indexes were X 2(gl)=165,73(80): CFI=0.98: TLI=0.97: RMSEA=0.06: in the Brazilian version they were X 2(gl)=233.64 (80); TL=0.98: CFI=0.98; RMSEA=0.05. The Composite Reliability Coefficient was good (0.75).
A sociodemographic questionnaire was also applied to describe die sample. The characteristics collected were: gender, age, salary range, marital status. State of residence, total working time, working time in the current job. nature of the organization in which you work and size of the team under his/her leadership.
Results
Initially, the Confirmatory Factor Analysis was performed for all three instruments used in the investigation, in the sample of the first wave (T1). All scales were modeled with only dimension. The results showed that all models showed good adjustment rates. Then, with view at verifying whether the variables were discriminating against each other, the general measurement model was calculated, with the inclusion of all the variables proposed in the study, taking into account also the data of wave one. The model's adjustment indexes were good (X 2[gl]=1136.68 [567]; CFI=0.98; TLI=0.98: RMSEA=0.05). In addition, averages, standard deviations and correlations between scales were calculated for wave one sample data (Table 1).
Table 1 Latent correlations and reliability between study variables for the complete sample - TI (N= 479).
| M | SD | 1 | 2 | 3 | ||
|---|---|---|---|---|---|---|
| 1. | Servant Leadership | 3.83 | 1.24 | 0.92 | 0.47 | 0.46 |
| 2. | Harmonious Passion | 5.01 | 1.44 | 0.90 | 0.81 | |
| 3. | Work Intentions | 4.64 | 0.97 | 0.75 |
Note: The main diagonal shows the values of the composed reliability indexes of the instruments: all relationships were significant at the p<.001 level. Correlations were estimated between latent factors, inside the structural equation model.
In the analysis of the instrument's invariance, the results presented in Table 2 show that the decrease in the model's fit is tolerable, when the number of items and factors (configural invariance), factor loads (metric invariance) and intercepts are progressively fixed (scalar invariance). In this connection, hi the three invariance modalities, no greater than expected increase differences were found (ΔCFI <0.01; ΔGamma-Hat <0.001 and McDonald's NCI <0.02), both for the Servant Leadership Scale, as well as for the Harmonious Passion for Work Scale. With regard to the Work Intent Inventory, the Gamma-Hat and XCI differences were slightly above the cutoff point, for the scalar invariance model. However, the differences in CFI and RMSEA were very small, and thus this instrument was also considered to be invariant.
Table 2 Analysis of instrument's longitudinal invariance.
| Model | X 2 (gl) | RMSEA (CI 90%) | CFI | Gamma-Hat | McDonald's NCI |
|---|---|---|---|---|---|
| Servant Leadership | |||||
| 1. Configural | 513.92 (225) | 0.05 (0.05-0.06) 0.05 | 0.945 | 0.870 | 0.739 |
| 2. Metric | 522.61 (241) | (0.04-0.05) 0.05 | 0.946 | 0.869 | 0.744 |
| 3. Scalar | 534.07 (253) | (0.04-0.05) | 0.946 | 0.869 | 0.734 |
| Harmonious Passion | |||||
| 1. Configural | 352.74 (156) | 0.05 (0.04-0.06) | 0.955 | 0.895 | 0.813 |
| 2. Metric | 370.46 (170) | 0.05 (0.04-0.06) | 0.955 | 0.894 | 0.810 |
| 3. Scalar | 398.12 (182) | 0.05 (0.04-0.06) | 0.952 | 0.893 | 0.797 |
| Work Intentions | |||||
| 1. Configural | 94.04 (77) | 0.03 (0.00-0.04) | 0.984 | 0.982 | 0.977 |
| 2. Metric | 106.62 (82) | 0.03 (0.00-0.04) | 0.982 | 0.981 | 0.975 |
| 3. Scalar | 118.48 (88) | 0.03 (0.01-0.04) | 0.978 | 0.975 | 0.969 |
In the hypothesis test, all variables hi the model were considered to be latent and configured based on the factor loads obtained in the final measurement model. To this end, two steps were taken. In the first one, the direct effects between the variables, through the cross-lagged panel analysis, were verified. In this connection, there was, initially, a direct effect between servant leadership and work intentions over time, without the harmonious passion for work (mediating variable). The results indicated that the servant leadership hi T1 predicted positively and significantly the work intentions in T2 (β = 0.06: p < .01), that the servant leadership in T2 predicted positively and significantly the work intentions in T3 (β = 0.11: p <.01), as well as the servant leadership in T1 positively and significantly predicted work intentions in T3 (β = 0.13: p <.01) (Figure 1). Thus these findings support hypothesis 1, although it should be noted that the values of the observed relationships were weak.
Note. **p <.001, *p <.05.
Figure 1 Graphical representation of the final model, with non-standard results.
Then, the effects of servant leadership on the harmonious passion for work were verified. The indices demonstrated that the servant leadership hi T1 positively and significantly predicted the harmonious passion for work inT2 (β = 0.05: p < .05), just as the servant leadership hi T2 positively and significantly predicted the harmonious passion for work in T3 (β = 0.04; p<.05), and the servant leadership in T1 predicted positively and significantly the harmonious passion for work in T3 (β = 0.08; p < .05).
Subsequently, the effects of harmonious passion for work on work intentions were tested. The results indicated that the harmonious passion for work hi T1 predicted positively and significantly the intentions of work in T2 (β = 0.21: p <.001), and that the harmonious passion for work in T2 predicted positively and significantly the intentions of work inT3 (β = 0.16.p < .001), as well as the harmonious passion for work hi T1 positively and significantly predicted work intentions in T3 (β = 0.08: p < .05).
In the second step, the harmonious passion for work was included in the model as a mediating variable in the relationship between servant leadership and work intentions, and the direct and indirect effects over time were analyzed. The results obtained indicated that the relationship between the servant leadership in T1 and the work intentions hi T3 was significant and positive (β = 0.03: p <.05). Previous research may have indicated greater effects between these variables. But hi our study, after controlling for the effect of time, the relationship is not that great (although it remains statistically significant).
However, contrary to expectations, a negative and significant indirect effect was observed in such a relationship (β = -0.16: p <.001). indicating that there was a partial mediation of the harmonious passion for work in T2 in this relationship, however, hi a negative sense, due to the fact that the effect of harmonious passion for work in T2 on work intentions in T3 was significant, but negative (β = -0.12: p < .001). These results thus prevented the empirical support of hypothesis 2 (Figure 1). The final model presented the following adjustment indices: X 2 = 23.51 (2), CFI = 0.99. TLI = 0.96. RMSEA = 0.09/0.08-0.09.
Discussion
Although the Passion for Work Assessment Model (MAPT) involves the explanation of the passion process, its temporal nature has been little investigated. In order to contribute to further elucidate tins issue, the present study aimed to investigate the mediating role of harmonious passion for work in the positive relationship between servant leadership and work intentions, in a longitudinal study, over six months.
Regarding the testing of hypotheses, there was support for hypothesis one, as it was found that the servant leadership (T1) positively and significantly predicted work intentions (T3), over the six months. Such findings are similar to the relationships observed between servant leadership and positive work results, such as intentions to stay, individual initiative and innovative behaviors (Hajjaj. 2014: Panaccio et al., 2015). Servant leaders often turn to then followers. With this, their concerns go beyond the organization itself, since they seek to foster the well-being of all the organizational characters involved, whether they are customers or employees, which causes their subordinates to develop a strong desire to engage in behaviors that benefit the organization (Barbuto & Wheeler 2006: Liden et al., 2008).
With respect to hypothesis two, the results obtained did not allow it to be sustained, as harmonious passion mediated the relationship between servant leadership and work intentions: however, such indirect effect was negative.
Thus, the relationship between the servant leadership (T1) and the mediator variable, harmonious passion for work (T2) was significant and positive, but very low. According to Ho and Astakhova's Passion Transfer Model (2020), passionate leaders transmit their passion for work to then team members through a process of contagion that leads their subordinates to also develop greater passion for their work. It is thus possible that the positive influence of leadership styles on work passion is conditioned by the passion for work of the leader, winch implies the need to perform future investigations to clarify tins issue.
The relationship between the mediating variable (harmonious passion for work) and the dependent variable (work intentions) was significant, although negative, thus causing a negative mediation of the harmonious passion for work hi the relationship between the servant leadership and the intentions of work. Results consistent with these findings were also observed in a study hi which a negative relationship was found between passion for work and other types of work intentions (Tarkar et al., 2019).
According to the Social Cognitive theory (Bandura. 1986) and MAPT (Zigarmi et al., 2012), employees assess the social/organizational environment that surrounds them, which will affect their perception of that environment and. consequently, the development of then harmonious passion, which, hi turn, can have positive or negative repercussions on the results of then work. Therefore, it is likely that, although the servant leadership positively impacts the harmonious passion for work, other characteristics present hi the work context and that were not investigated in the present study, such as the organizational culture, were contributing to negatively impact employees' assessments about such framework. This impact was reflected, therefore, hi the decrease of their passion for work and. consequently, hi the reduction of their work intentions. This argument is further supported by the leadership substitute model (Kerr & Jennier, 1978), according to winch the combination of individual, contextual and organizational characteristics usually explains not only the reasons why some leaders' behaviors trigger positive results hi their subordinates, but also the reasons why other leaders ' behaviors have no effect on then subordinates, or lead to dysfunctional effects.
It should also be noted that the final model may have suffered from the lack of statistical power. (Aguinis, 1995: Bidder et al, 2022), due to the sample drop that occurred between the different waves. In other words, the statistical power for detecting significant relationships has been substantially reduced due to sample loss, which may have been reflected hi the results now obtained.
This investigation also presents important practical contributions. Firstly, the importance of the leadership style for maintaining a harmonious passion for work is highlighted. Thus, it is important for the Human Resources function to foster the development of leaders with the characteristics of a servant leader, who are thus able to cultivate, develop and or maintain the harmonious passion for their subordinates ' work for a longer time. To this end, leaders should be guided to focus on strategies such as the recognition of the work performed by their subordinates and hi providing guidelines and appropriate work demands. Another practical implication concerns professional retention, since positive and ethical leaders, who value their subordinates ' involvement in their work, tend to have subordinates who manifest more positive intentions to stay in the organization.
In summary, the findings of the present study contributed to the broader understanding of the harmonious passion for work construct, by testing a procedural model on the impact of servant leadership on work intentions, with the mediation of harmonious passion, through a longitudinal study. Thus, they corroborate for an advance hi the set of empirical evidence that has been accumulated, in recent years, on the Passion for Work Assessment Model.
Limitations and future studies
The present study showed that the effects investigated here persisted over six months, despite the reduction hi the prediction strength, an indication that the time interval between measurements is important to verify the maintenance and continuity of these predictions. In this connection, there was a positive influence of the perception of the servant leader on the harmonious passion for work, which tends, however, to dwindle over time, as well as a negative influence of the passion on the work intentions, over time, which leads passion to play a negative mediational role in the relationship of servant leadership with work intentions. Such findings, therefore, indicate the need for future studies capable of deepening the understanding of the different ways in winch leaders transfer their harmonious passion to their subordinates, as well as the different situations in which this passion turns into positive outcome at work.
It is also worth noting some limitations of the present study. The main issue refers to the fact that the data were collected, in their entirety, on-line, without the presence of the investigators, which makes it difficult to control the differences that the type of work and the workplace exert, for example, in the perception of the leader. Another limitation concerns the collection was only self-report In addition, the online application may result in a lack of involvement and less reliable responses. Future investigations could therefore be carried out in person.
A second limitation occurs in the fact that although longitudinal investigations provide data which are more consistent with a causal model (Orth et al., 2009), they still do not allow the establishment of causal relationships. In this connection, experimental investigations with the manipulation of harmonious passion for work, for example, could be an alternative for investigating causality further. Another limitation observed is related to sample loss in the last wave, which can jeopardize the generalization of results.
With regard to future investigations, it is suggested that the mediating role of passion for work hi the context of the Passion for Work Assessment Model be deepened, by investigating other antecedent variables (such as. for example, work cognitions, other leadership styles and the passion of leaders for then work) and consequent variables (such as job crafting behaviors, job performance and health problems) of such a construct, hi addition, the investigation of more complex models, with the simultaneous inclusion of mediation and moderation relations, may contribute to a deeper understanding of the Passion for Work Assessment Model hi Brazilian organizations, as well as to the implementation of measures capable of fostering the development and maintenance of the harmonious passion for work hi such a context.
