Super (1980) was one of the forerunners of enrichment theory. When studying career development and organizational commitment, the author brought to the discussion the idea that multiple roles can be extensive, that is, they can support each other. According to Greenhaus & Powell (2006, p. 73), work-family enrichment refers to “the extent to which experiences in one role improve the quality of life in the other role”. For these authors, “quality of life” means high performance and positive affect. Furthermore, it is understood that work-family enrichment can occur in both directions, in a wide range of family contexts and with different professionals (Kim, et al., 2021). In other words, regardless of the individual's profession and family configuration, this phenomenon can occur bidirectionally.

Shein and Chen (2010) reported that the theoretical model proposed by Greenhaus & Powell (2006) extends the literature by incorporating a greater amount of resources generated in one role that can be applied in another role and by proposing two ways in which these resources can be applied: the instrumental path and affective path. Through the instrumental path, resources generated in one domain act directly on the other domain, increasing performance in the receiving domain. On the other hand, on the affective path, there is an indirect influence of the experiences of one domain over the other. That is, resources generated in one domain can promote positive effects within that role, which in turn produces high performance and positive effects in the other, in a bidirectional sense (Greenhaus & Powell, 2006). Furthermore, the authors incorporated moderating variables into the model, which determine the conditions under which resources in one role are more likely to enrich the quality of life in the other role. New moderating variables are welcome with regard to testing new ETF relationships and behaviors.

In this way, current studies have sought to identify new variables and deepen the knowledge of their effects when relating them to WFE. As an example, there are issues related to the work performance of individuals (Allen, French, Dumani & Shockley, 2020; Solat, Abrar, Shabbir, Bashir, Saleem & Saqib, 2020), perception of marital support (Silveira-Rodrigues, Chambel & Carvalho, 2021), well-being and personality traits (Kim, et al., 2021), subjective success and engagement at work (Awan, et al., 2021) and gender influences (Mishra & Bhatnagar, 2019; Bansal & Agarwal, 2019; Kim, et al., 2019).

The results of these studies demonstrate direct relationships between different realities and professions with work-family enrichment. Solat et al. (2020), when studying nurses in Pakistan, found that social interactions at work have a positive effect on work performance and consequently on WFE. In Brazil, it was found that military personnel on a peace mission in Haiti, when they perceive support from their family and especially from their spouse, feel satisfied, and this positively influences the relationship with WFE and FEW (Silveira-Rodrigues, Chambel & Carvalho, 2021).

In a longitudinal study with employed adults, to gauge the influence of well-being and individual psychological traits on WFE, the main finding of Kim et al. (2021) was the discovery that interventions to improve, over the years, individual psychological well-being also increase work-family enrichment. Individual personality traits, on the other hand, are not associated with changes in the perception of WFE. Concerned with organizational performance, Awan, et al. (2021) carried out a study in the banking sector in which they found that FWE positively affects subjective success, mediated by engagement at work. This consequently improves personal and organizational performance, thus paying attention to work policies that improve workers' conditions in favor of better professional development.

On the influence of gender, Kim et al. (2021) carried out a survey of several studies that used the gender variable combined with numerous organizational or personal variables, thereby suggesting that the nature of an individual's gender can influence work-family enrichment, and may even transform it into a study construct.

That said, it is observed that work-family enrichment is influenced by several factors and behavioral variables that affect individuals’ professional performance. For this study, the objective was to introduce gender as an influencer of the relationship between WFE and Organizational Citizenship Behaviors, since few studies have used gender moderation and the few that have done so are not in consensus. Additionally, there are Organizational Citizenship Behaviors. They are positive individual behaviors that aid the development of the work environment, which can relate to and be influenced by gender.

Organizational citizenship comprises “discretionary individual behaviors, not direct or explicitly recognized by the formal reward system, and which in aggregate promote the effective functioning of the organization” (Organ, 1988, p. 4). The contributions of several scholars, such as Barnard (1938) and his concept of “the willingness to cooperate”, Roethlisberger & Dickson (1939) and the concept of “informal organization”, Blau (1964) with studies on the dynamics of “social exchanges”, and Katz and Kahn (1973), who developed the concept of “innovative and spontaneous behaviors”, are part of the foundation of the concept of organizational citizenship proposed by Organ (1988).

Two important features were pointed out by Organ (1988). The first is that the manifestation of such behavior is not associated with a lack of rewards on the part of the organization, but rather that there is no contractual guarantee of any retribution. The second is the lack of a direct relationship between some OCB instances and the modification of organizational results. Therefore, there may be rewards, but they will be indirect or uncertain and not expected by those who behaved like this, but resulting from the employees' devotion to the company in order to voluntarily cooperate with their colleagues and the organization (Na-Nan; Kanthong & Joungtrakul, 2021).

Such behaviors were divided into three dimensions, in the view of Bastos, Siqueira & Gomes (2014). The first dimension concerns behaviors that develop creative actions or suggestions for the smooth running of the organization. The second dimension includes behaviors related to the positive dissemination of the organization's image to external individuals. The third dimension is characterized by behaviors that promote cooperation among co-workers. All these behaviors will always be spontaneous and prioritize the smooth running of the organizational environment.

Gomes (2011) explained that it is not easy to demonstrate organizational effectiveness in practical terms and that not every dimension of OCB would make a difference in the performance of organizations. However, factors such as engagement, commitment and job satisfaction mediate the formation of behaviors favorable to the good of the organization, organizational citizenship behaviors (Na-Nan; Kanthong & Joungtrakul, 2021 corroborating with Liu, Zhao and Sheard, 2017), who understood that even though organizational citizenship behaviors cannot be directly recognized, they can be identified as a source of sustainable competitive advantage for organizations.

Recently, Organ (2018), pointed out that the main guidelines of studies involving organizational citizenship behaviors are related to constructs of job satisfaction, justice at work and personality traits. However, new efforts are being made to go beyond such analyses and identify new causes for the emergence of organizational citizenship behaviors. Studies are expected to test new connections and the use of moderating variables (Organ, 2018).

In this sense, studies aimed to identify the relationship between organizational citizenship behaviors and work-family conflict/enrichment. For example, Baral & Bhargava (2010) identified that work-family enrichment has a positive relationship with organizational factors, such as OCB. In more recent studies, several variables were tested.

Özlük & Baykal (2020) intended to analyze the relationship between CCO, job satisfaction and job confidence among a sample of nurses from Istanbul, Turkey. On that occasion, the authors identified a significantly positive relationship between OCB and the other tested dimensions, and through a regression, trust at work can be explained by 13.5% of the OCB levels and satisfaction was related to 80.9% of the OCB levels in the studied sample (Özlük & Baykal, 2020). The notable contribution of the job satisfaction construct is observed as being directly related to citizenship behaviors, emphasizing the need for a good relationship between workers and organizations to achieve a high level of satisfaction with work.

Hollanda (2020) investigated, in the Brazilian context, whether conflict and work-family enrichment, in a longitudinal variation, was related to the occurrence of organizational citizenship behaviors. In this case, Holanda (2020) found that OCB CCO is more closely related to the work-family conflict due to the demands of time, and by dedicating more to the organization, less time is available to spend with the family. Thus, it was found that OCB can be a predictor of both conflict and work-family enrichment.

Finally, by relating OCC and gender, Gao (2020) intended to identify whether there are gender differences in organizational citizenship behaviors, as some studies have claimed that there are and others have not identified any differences. Thus, the author studied workers from different private sectors in China and found that men tend to exhibit more citizenship behaviors than women in the organizational context. However, Gao (2020) worked with the eastern scenario, culture and behaviors, which are totally different from those observed in the west and more specifically in Brazil, the context of the present study. Consequently, this study proposes to go deeper and identify the influence of gender in the relationships between the OCB and WFE constructs, as this relationship has not yet been studied among Brazilians.

As already noted, the relationship between gender-moderated FEW and OCB has not yet been studied in the Brazilian context or in the organizational environment of military barracks. There is a gradual change in the occupation of positions in the Brazilian Armed Forces, and the gradual increase in the admission of women is becoming apparent (Giannini, Folly & Lima, 2017). However, the challenges are still great and of the same proportion as those of other women in different public or private careers. Some studies have investigated relationships between gender and FWE and others between gender and OCB, with relevant results and deeper discussions of the themes, which aided the construction of the hypotheses of the present study.

Analyzing the issue of gender as a moderator of work-family enrichment also involves cultural aspects and gender expectations in which different behaviors between professional and family life are expected from women and men (Kim, et al., 2021). However, in a meta-analytic study, Lapierre et al. (2018) did not identify gender differences in relation to FWE. In this respect, Kim et al. (2021) denoted that gender issues may be tracing two different paths of interpretation: first, the similarity between men and women with regard to the FWE may mean that the beliefs in gender differences are less evident; or, it could mean that this similarity of experiences may be a tendency towards new social behavior, ranging from new personality traits to new experiences between work and family (Kim, et al., 2021).

In a longitudinal study, Kim et al. (2021), when testing the gender relationship with FWE in a sample of the American employed population, identified that the stability of work-family enrichment over the years does not differ by gender. However, other studies found some significant differences regarding gender, emphasizing that different contexts, cultures and samples also influence the results (Junker et al., 2020).

Mishra & Bhatnagar (2019) found two results on gender differences in a sample of Indian workers. On culture and work-family enrichment, gender had no influence. However, when dealing with more specific aspects, such as the relationship between the salience of the family role and community support influencing enrichment in the work-family direction, these were more strongly perceived by male workers. This led the authors to suggest that it is necessary for managers and organizational policy makers to understand the importance of family and community for the organizational environment (Mishra & Bhatnagar, 2019).

Bansal & Agarwal (2019) used the ideology of gender as a moderator of FWE, thus identifying that among workers who think about gender equally, there was greater work-family enrichment, unlike those who still perceive gender with different roles between men and women. For these individuals, the perception of FEW was lower (Bansal & Agarwal, 2019).

In studies that adopted the gender relationship as a moderator of organizational citizenship behaviors, it was possible to identify contrasts between the results of different surveys. Gao (2020) was one of the authors who noticed that in several studies, gender was significantly different between men and women, but not in others. Therefore, in his research in the Chinese context, the author found that men were more likely to act with organizational citizenship behaviors. In contrast, Aftab, Ali Shah & Khan (2020) studied organizational commitment and OCB among a sample of Pakistani university professors, with women scoring significantly more in both dimensions compared with men, showing that in this case women are more prone to organizational citizenship behaviors. It also shows that organizational commitment is a predictor of OCB (Aftab; Ali Shah & Khan, 2020).

Finally, in a sample of American public servants, the relationship between gender and OCB mediated by work motivation was tested. It was found, once again, that women were significantly different, indicating a higher level of motivation at work and, consequently, greater propensities for organizational citizenship behaviors compared with men (Alanazi, 2020).

That said, it is evident that several studies have addressed the issue of gender in different cultural and organizational contexts. However, there are few studies involving military workers and in a Latin American context, more precisely in Brazil. Thus, among so many possible relationships, a gap was identified and the possibility of testing the relationship between gender-moderated FWE and OCB. It is understood that FWE can have consequences for the performance of workers and that this may be related to organizational citizenship behaviors in different ways. Additionally, there is the variable of gender as a moderator because, as mentioned in several studies, gender still needs to be studied, since it can have a direct influence on interpersonal and organizational relationships.

This discussion gave rise to the hypotheses of this study, highlighting that as such a relationship had not been tested in previous studies, it was decided to compare genders in the relationships and moderations between the dimensions of the FWE scale and the OCB scale. These hypotheses will be presented below:

H1: Work-Family Enrichment (WFE) moderates the relationship between Work-Family Enrichment (WFE) and Organizational Image Disclosure (OID);

H2: Work-Family Enrichment (WFE) moderates the relationship between Work-Family Enrichment (WFE) and Creative Suggestion (CS);

H3: Work-Family Enrichment (WFE) moderates the relationship between Work-Family Enrichment (WFE) and Colleague Cooperation (CC);

H4a: Work-Family Enrichment (WFE) is related to Organizational Image Disclosure (OID);

H4b: Work-Family Enrichment (WFE) is related to Creative Suggestions (CS);

H4c: Work-Family Enrichment (WFE) is related to Colleague Cooperation (CC);

H5a: Work-Family Enrichment (WFE) is related to Organizational Image Disclosure (OID);

H5b: Work-Family Enrichment (WFE) is related to Creative Suggestions (CS);

H5c: Work-Family Enrichment (EFT) is related to Colleague Cooperation (CC).

Initially, a path model was prepared (Figure 1) for the dimensions of the WFE and CBO scales influenced by gender, that is, a diagram design that illustrates the research hypotheses and shows the relationships between the variables that will be analyzed (Hair et al., 2017).

Figure 1
Path model of the dimensions of the WFE * FWE – CBO
Source: Smart PLS® software v. 3.3.3 (Ringle, Wende and Becker, 2014).

This stage of structural equation modeling deals with the construction of the measurement model, representing the relationships between dimensions (latent variables – LV's) and their corresponding observed variables (observed variables – OV's) (Hair et al., 2017). The model is based on theory, and this condition is necessary to obtain useful results from the PLS-SEM. Hypothesis tests involving structural relationships will only be reliable and valid if the measurement model explains how these dimensions are measured. According to Nascimento & Macedo (2016), in structural equation modeling, there are two types of measurement scale, reflexive and formative. For the purposes of this study, the formative was used. The path diagram and its relationships between gender-moderated latent variables and observed variables are presented in Table 1.

Table 1
Path diagram for structural equations of the WFE model * Moderator - OCB

Source: Survey data based on Hair et al. (2014)

As seen in Table 1, the measurement model has nine parameters (®’s) with three moderations and six structural coefficients that connect the two exogenous dimensions with the three endogenous dimensions, with their 24 observed variables. Following the structuring of the measurement model, the process of estimating the PLS path model began.

Partial regression models are estimated by PLS-SEM algorithms in two steps: in the first, dimension scores are calculated and in the second, the weights, path model coefficients and R2 values resulting from the predictive dimensions are calculated (Hair et al., 2009).

The Smart PLS® algorithm was configured for 7 completion criteria. The number of iterations was set at 300, representing the maximum number of iterations used to calculate PLS results. Initial weights for external indicators were set at 1.0. When running the model on Smart PLS®, it stabilized after 2 interactions.

4.2.1 Evaluation of the Measurement Model

To assess the measurement model, the internal consistency assessment was used, consisting of Cronbach's Alpha (α), composite reliability (ñ.) and convergent validity by Average Variance Extracted (AVE). Cronbach's alpha and composite reliability range between 0 and 1, where values between 0.70 and 0.90 are considered good and efficient (Hair et al., 2017). Values above 0.90 are not desirable (and definitely undesirable above 0.95), as they may indicate that respondents may have demonstrated redundancy or duplicity in their responses, while values below 0.60 indicate a lack of reliability of internal consistency (Lopes et al., 2020).

The model's discriminant validity was determined using the Fornell and Larcker criterion, which determines that the square root of the AVE’s must be greater than the values of the correlation matrix (Fornell & Larcker, 1981). Another and more robust criterion is the HTMT criterion (Heterotrait - Monotrait Ratio), this measure is an estimate of what would be the true correlation between dimensions if they were perfectly measured (Hair et al., 2017). Henseler, Ringle & Sarstedt (2015) suggested that using the bootstrapping process for 5,000 subsamples, the upper limit of the HTMT confidence interval for 95% confidence should be below 1.0, indicating that the model has discriminant validity. Table 2 shows the evaluation of the model's internal consistency and Table 3 the discriminant validity.

Table 2
Cronbach's alpha, composite reliability and convergent validity for the measurement model

Source: Smart PLS® software v. 3.3.3 (Ringle, Wende and Becker, 2014).

Cronbach's alphas and composite reliability for the WFE and OCB instruments presented values between 0.868 and 0.928 and the AVE values from 0.623. Therefore, the model presents internal consistency and convergent validity. Table 3 shows the discriminant validity of the model:

Table 3
Analysis of discriminant validity using the Fornell-Larcker and HTMT criteria

Source: Smart PLS® software v. 3.3.3 (Ringle, Wende and Becker, 2015).

4.2.2 Structural Model Assessment

This step aims to assess whether the moderation model represents the underlying theories of the path model, allowing an analysis of the predictive capacity of the model and the relationships (hypotheses) between dimensions. The systematic approach to evaluating the structural moderation model was performed by analyzing the collinearity of the model (Variance Inflation Factor - VIF); by the coefficient of explanation of the predictive dimensions (R.) and the size of the effect .. between the exogenous and endogenous dimensions. Finally, the predictive relevance (Q.) was evaluated using the blindfolding method (Table 4):

Table 4
VIF, f2, R2 and Q2 values for moderation model dimensions

Source: Smart PLS® software v. 3.3.3 (Ringle, Wende and Becker, 2015).

As shown in Table 4, the VIF values for the dimensions are all below 5, so it can be stated that the collinearity does not reach critical levels due to the dimensions, not being a problem in the estimation of the moderation model. As for the explanation coefficient, the dimensions Cooperation with Colleagues (CC) (R. = 0.309) and Organizational Image Disclosure (OID) (R. = 0.203) had a strong effect, while Creative Suggestions (CS) (R. = 0.072) had a weak effect with regard to the explanation of the final model (Cohen, 1988; Lopes et al., 2020).

As for the size of the effect between the dimensions, it was observed that: FWE → (CC (.. = 0.034; p = 0.186), OID (.. = 0.005; p = 0.638) and CS (.. = 0.047; p = 0.134) presented non-significant effects, but did not prevent a possible relationship between dimensions. FWE → (CC (.. = 0.003; p = 0.774) and CS (.. = 0.019; p = 0.400)) also had non-significant effects, only FWE → OID (.. = 0.309; p = 0.000) had a large and significant effect (p < 0.05) (Cohen, 1988; Lopes et al., 2020). If the OID dimension with a value of Q. > 0.25 is considered a value of strong degree, the others are considered of weak degree Q. < 0.075 (Chin, 2010; Lopes et al., 2020).

Following the analysis, the significance of the coefficients of the structural model was evaluated. In the case of the structural model the betas must be statistically different from zero for the hypotheses to be confirmed. For Hair et al. (2017) the SmartPLS® software calculates the Student's t test between the original values of the data and between the values obtained from the subsamples through the bootstrapping technique for each relationship between the exogenous dimensions and the endogenous dimensions, forming the model with the structural coefficients (β's).

To compare Hypotheses H4(a-c) and H5(a-c), the multigroup analysis technique (MGA) was applied to compare the genders. Before performing the MGA, the measurement invariance of composite models (MICOM) must be tested using the three-step procedure, a necessary requirement (Sininkovics et al., 2016). Nguyen-Phuoc et al. (2021) suggest three steps: 1) assessment of configuration invariance; 2) the evaluation of compositional invariance; and 3) test for means and variances (Table 5).

Table 5
Invariance measurement testing using permutations

Source: Smart PLS® software v. 3.3.3 (Ringle, Wende and Becker, 2015). DBM = Difference between Means; DVO = Difference between Variances; and CI = Confidence Interval

Analyzing Table 5, the results show that the measurement invariance of both groups was fully established. A minimum requirement to assess the significant difference between the two groups using MGA (Hair et al., 2017). The MGA results are presented in Table 6. The Henseler MGA method (non-parametric method) and the Permutation Test were used as comparative tests between men and women. In the Henseler MGA method, a p value greater than 0.05 does not demonstrate significant differences between the specific path coefficients for the groups at a significance level of 5%. The same criteria were used for the Permutation Test (Nguyen-Phuoc et al., 2021).

Table 6
Multigroup analysis results: male (M) x female (F)

Source: Smart PLS® software v. 3.3.3 (Ringle, Wende and Becker, 2015). CC = path coefficient; N = no

Table 6 shows that there was no significant difference in the relationships between genders. This does not support the main assumption of this study, which was to analyze the influence of gender on the relationship between WFE and OCB in FAB military personnel. This result cannot be directly compared with other studies, as these relationships have never been tested before. Therefore, it was an unprecedented result that gender among Brazilian Air Force military personnel does not interfere in their behavior with regard to the dimensions of WFE and OCB. However, there is a similarity to different results from previous studies that identified the gender variable as non-significant. The gender variable can be influenced by cultural aspects and the individuals' conception of gender ideology (Bansal & Agarwal, 2019; Kim, et al., 2021). With that, their experiences and experiences in the family and professional environment become different. The non-significance of gender in this study corroborates the findings of Lapierre et al. (2018), when they do not identify differences between men and women in the WFE dimension, as well as the results of Kim et al. (2021).

However, when comparing these results with the OCB dimension, it was observed that the present study shows a different result from other studies mentioned here. Several authors were able to identify a gender difference in organizational citizenship behaviors. Gao (2020) found that men are more likely to have organizational citizenship behaviors, whereas Aftab, Ali Shah & Khan (2020) found that women are more inclined to this type of behavior towards the organization. The work of Alanazi (2020) also shows that women are more likely than men to show this behavior.

Considering the context of the current research, the non-differentiation of gender may be related to the masculine environment that characterizes military matters. Thus, there may be changes in the behavior of individuals, or even a homogenization of behaviors, as the environment has strict behavioral rules and follows a culture that is far removed from the reality of other organizational environments. This result leaves room for further investigations that focus on the context and culture of the military work environment.

That said, Table 7 presents the test of the hypotheses presented in the study, in which genders are already analyzed in a clustered manner, as there was no gender differentiation within the studied sample. Table 7 presents the values of the relationships between dimensions considering the original sample and the mean of the subsamples, standard deviation, t statistic and p values.

Table 7
Relationships between latent variables of the structural model

Source: Smart PLS® software v. 3.3.3 (Ringle, Wende and Becker, 2015). SD = Standard Deviation

It was observed in Table 7 that in Hypotheses 1 to 3, with interference from the moderating variable (FWE), hypothesis H1: WFE→ DIO was the only one not confirmed (p > 0.05), that is, Work-Family Enrichment (WFE) does not moderate the relationship between Work-Family Enrichment (WFE) and Organizational Image Disclosure (OID). This means that regardless of whether the individual is comfortable with the family dimension, it does not assume an effect on enrichment in the work-family direction, which in turn does not modify behavior in the organization, regarding the dimension of dissemination of organizational image. In other words, an individual even when satisfied with family relationships will not overflow this feeling to the dimension of work and the desire to positively publicize the organization to people outside it. Addressing the results of Özlük & Baykal (2020), due to the level of job satisfaction, impacts and is strongly related to the OCB. This result is gender-independent and reveals to what extent the family dimension may affect the organizational environment, a relevant factor for the improvement in the area of people management.

As for direct relationships without interference from the moderating variable, it was observed that the WFE dimension is significantly related only to the OID dimension (p < 0.05). This result shows that when there is no relationship with FWE, that is, the family interfering with work, there is a different result. Hypothesis H4a shows that positive levels of enrichment in the work-family direction affect organizational behaviors to the point that workers feel like talking about the company, disclosing the organizational image to other people. It has been shown that the feeling that work also enriches the family dimension results in positive returns for the organization, corroborating, to a certain extent, the study by Kim et al. (2021), since these authors identified that organizational interventions in favor of workers' psychological well-being also increased the perception of WFE.

Conversely, there are hypotheses related to the FWE dimension. In this case, hypothesis H5a was not confirmed, in which a relationship between FWE and OID is supposed. Unlike Hypothesis H4a, the FWE-OID relationship indicates that enrichment towards the famly in work does not result in voluntary behaviors of dissemination of the organizational image. Hypotheses H5b and H5c revealed a significant relationship between FWE-CS and FWE-CC. It is inferred that when positive aspects of the family dimension affect the work dimension, the individual is more prone to organizational citizenship behaviors in the dimensions of creative suggestions and cooperation with colleagues. These results are in keeping with the findings of Silveira-Rodrigues, Chambel & Carvalho (2021), who also studied the military context. At the time, the authors found that family-work enrichment, through the support of the spouse, improves the perception of well-being and satisfaction with work, in other words, being satisfied with the family dimension has an impact on good actions at work. This is positive for FAB military personnel, since such work demands a great deal from its workers and being well in one dimension will lead to positive results to the other dimension.

To finalize and better visualize the analysis of the results, Table 8 and Figure 2 present the final path diagram for the relationship between the WFE * OCB models:

Table 8
Final path diagram model for WFE * FWE – OCB

Source: Survey data based on Hair, et al. (2014).

Figure 2
Final model
Source: Smart PLS® software v. 3.3.3 (Ringle, Wende and Becker, 2015).

With the presentation of the final model of the initially proposed relationships, it is evident that changes occurred between the expected (Figure 2) and final results (Figure 2) for the sample of military personnel from the FAB in Rio Grande do Sul, Brazil. The main difference between the initial and final models is the issue of gender, as it was not possible to identify the influence of gender on the perception of the respondents. Now that the results have been presented and discussed, the final considerations of the study can be given.