Situational antecedents to organizational identification and the role of supervisor support

Purpose – The purpose of this paper is to examine the effects of procedural justice, training opportunities and innovation on job satisfaction and affiliation commitment via the mediating effect of organizational identification. The authors also explored the moderating role of satisfaction with supervisor on the relationship between the antecedents and organizational identification as well as its moderating effect on the mediational chain. Design/methodology/approach – The authors used structural equation modeling techniques, using MPLUS 7.4, to analyze data collected from 247 full-time employees who were recruited by undergraduate students attending a private university in the Southeast region of the USA. Findings – Results demonstrated that the indirect effects for procedural justice and training opportunities as predictors were significant, while none of the paths for innovation as a predictor were significant. Satisfaction with supervisor moderated the relationships between procedural justice and organizational identification and innovation and organizational identification. Originality/value – This research expands the nomological network concerning antecedents and consequences of organizational identification. It also explores the role of satisfaction with one’s supervisor, as this can affect identification with the organization. This research provides support for the notion that stronger employee–organization relationships lead to positive individual and organizational outcomes.

. Individuals who identify with their organization are more likely to work hard for the organization, are less likely to be absent, and are less likely to leave the organization (Ashforth, Harrison, & Corley, 2008;Dutton, Dukerich, & Harquail, 1994;Riketta, 2005). Employees whose sense of self is tightly connected to the social status and organizational prestige which the employing organization conveys are more satisfied and more involved in their jobs (He & Brown, 2013). This cognitive linking (Dutton, Dukerich, & Harquail, 1994) between the organization and the self is generally seen as the key to ensuring organizational effectiveness (Pratt, 1998).
The practical implications of increased OID among an organization's employees are clear. Organizations can expect to reap the benefits of a more highly engaged workforce through increased citizenship behaviors and higher levels of organizational support (Ashforth & Mael, 1989;Edwards, & Peccei, 2007). More fundamentally, increased OID causes individuals to exert greater effort on organizationally appropriate tasks that generate value (e.g. increased employee creativity: Hirst, van Dick, & van Knippenberg, 2009) while decreasing the incidence of behaviors which destroy value (e.g. the costs associated with turnover: Author, 2006). Framed in this way, the job of creating and maintaining a sense of connection between employees and organizations has become a leadership challenge (Scott & Stephens, 2009) as well as a leadership imperative.
Therefore, the purpose of this study is to investigate the impact of procedural justice, training opportunities and innovation on job satisfaction and affiliation commitment via the mediating effect of OID. Our intent here is to focus on the work environment or situational antecedents to OID which have not yet received sufficient attention in the OID literature. Additionally, we incorporate the moderating effect of satisfaction with the supervisor to further explore the nature of the environment in which identification is either enhanced or modulated.
The mediating effect of organizational identification OID is conceptualized as the "perception of oneness with or belongingness to" the organization (Ashforth & Mael, 1989, p. 34), and generally as a congruence of organizational and individual values (Pratt, 1998). Based on social identity theory (Haslam, 2004;Tajfel & Turner, 1986), OID posits that individuals classify themselves into different social categories to locate their place in the social environment (Stinglhamber, Marique, Caesens, Desmette, Hansez, Hanin et al., 2015). The model presented in Figure 1 positions OID as the center of our conceptualization of the relationships between the antecedents and outcomes in the OMJ 17,3 model. We do this for two reasons. First, Ashforth and Mael (1989) suggest that understanding OID is central to understanding the relationships between antecedents and individual attitudinal and behavioral outcomes. The implication that OID is a cognitive linking mechanism is supported by Bandura's (1986Bandura's ( , 1989 work in social cognitive theory. Bandura suggested that human behavior is regulated by forethought, such that cognitions about actions and potential outcomes help shape the level, direction and persistence of effort on tasks. Bandura's concept of emergent interactive agency (1986) suggests that people make conscious and reciprocal contributions to action by considering their motivations in light of various antecedents and outcomes. Additionally, Dutton and Dukerich (1991) suggest that OID is deeply dependent on perceptual (cognitive) processesthe more attractive an organization's image, the stronger the identification. Second, although Meyer, Becker and Van Dick (2006) review research that supports three perspectives on the relationship between identification and commitment (that identification is synonymous with commitment, a greater whole of which commitment is a part, or an antecedent of commitment), we believe that OID and affective commitment are indeed distinct, and that OID is an antecedent of affective commitment (Bergami & Bagozzi, 2000;Meyer et al., 2006;Riketta, 2005;Stinglhamber, Marique, Caesens, Desmette, Hansez, Hanin et al., 2015). Therefore, we position OID as a mediator in our model.

Situational antecedents to organizational identification
Previous research investigations of OID antecedents included analyses of OID attributes (Dukerich, Golden, & Shortell, 2002), leadership factors (Carmeli, Atwater, & Levi, 2011), and social exchange factors (Blader & Tyler, 2009), among others (He & Brown, 2013). Among the correlates of OID in Riketta's (2005) meta-analysis were those related to demographics (e.g. tenure, age, job level, gender, education). He and Brown (2013) suggest that dispositional differences in OID development is also under-researched. We believe that the inclusion of certain environmental and situational factors affecting OID is overdue. Specifically, we investigate the effects of procedural justice, training opportunities and an innovative culture on OID and its outcomes.
Procedural justice refers to the perceptions an employee has about the policies and procedures administered by an organization and whether they have a voice in the process (Konovsky, 2000). Employees can form these perceptions through daily interactions with their supervisors. Procedural justice has been found to motivate employee cooperative behavior and enhance job-related performance (Aryee, Chen, & Budhwar, 2004). Procedures that are perceived to be fair should increase employee perceptions of attachment and identity. Because individuals are generally motivated by self-enhancement needs, they tend to identify with organizations that confer positive qualities upon them (Dutton, Dukerich, & Harquail, 1994).
One of the mechanisms discussed by Fuller, Hester, Barnett, Frey, Relyea and Beu (2006) that acts as a signal to employees that they are valued is the development of high commitment management (HCM) practices within organizations. HCM practices include recognition, opportunities for training/development, participation in decision-making and problem-solving and pay for performance (Fuller, Hester, Barnett, Frey, Relyea, & Beu, 2006). In this research, we focus on one particular HCM practice: opportunities for trainings/ development. The extent to which individuals believe that the organization provides developmental opportunities for growth should be a tangible reminder that the organization values the individual's contribution (Allen, Shore, & Griffeth, 2003). Such programs provide the employee with opportunities to develop their skills and knowledge and communicate that the organization is willing to invest in the employee (Allen, Shore, & Griffeth, 2003). As Role of supervisor support a signal indicating that the employee is a valued member of the organization, the availability of training/development opportunities should serve to increase identification with the organization (Reade, 2001). And finally, external prestige is a significant contributor to self-concept orientations (Hameed et al., 2013;He & Brown, 2013) and OID, as individuals tend to identify more with organizations perceived as important and prestigious by external stakeholders (Fuller, Hester, Barnett, Frey, Relyea, & Beu, 2006). The group engagement model (Tyler & Blader, 2003) suggests that people evaluate their status in a group, in part, based on the perspective of those outside the group. Fuller, Hester, Barnett, Frey, Relyea and Beu (2006) suggest that organizational visibility and prestige is based upon characteristics (e.g. size, growth, distinctiveness) that are likely to result in the organization having a "publicly recognized name." (p. 820). Organizations which are distinctive via their success with innovation (e.g. meeting innovation goals, Fuller, Hester, Barnett, Frey, Relyea, & Beu, 2006) would appear to provide more visibility, status and prestige to their employees. Therefore, we believe that an organization which fosters a culture of innovation is attempting to provide something distinctive which may be reflected by external agents as conferring greater status or prestige on organizational participants. Consistent with this perspective, Vijayakumar and Padma (2014) examined four types of cultures for their impacts on OID. The culture that was most strongly related to OID was adhocracy cultures. These cultures are characterized by innovation and risk taking and can be further described as entrepreneurial in nature. Members of these organizations are free to experiment and challenge the current ways of doing things. Given this discussion, we hypothesize the following: H1. OID mediates the relationship between procedural justice and job satisfaction and affiliative commitment.
H2. OID mediates the relationship between training opportunities and job satisfaction and affiliative commitment.
H3. OID mediates the relationship between innovation and job satisfaction and affiliative commitment.
The moderating effect of satisfaction with the supervisor Satisfaction with the supervisor may enhance perceptions of support and OID in the context of effective procedural justice mechanisms, training opportunities and innovation initiatives. Immediate supervisors are considered the most influential in the daily work lives of their employees. They communicate expectations, provide rewards and training opportunities, and communicate policies and procedures (Kim, Hur, & Jun, 2017). DeConinck and Stilwell (2004) argued that those who feel they have been treated fairly and have voice will experience a higher level of satisfaction with their supervisors. In their study, perceptions of procedural justice were strongly related to satisfaction with one's supervisor across two data sets (DeConinck & Stilwell, 2004). Employee perceptions of training have been linked with satisfaction with supervisors as well. Schmidt (2007) found that those who are satisfied with workplace training report higher levels of overall job satisfaction, including satisfaction with one's supervisor. He highlighted the importance of training in attracting quality applicants as well as predicting overall job satisfaction and influencing career decisions. Finally, satisfaction with one's supervisor has been positively related to innovative organizations (Chen, Lee, Tsui, & Yu, 2012). In an examination of the cosmetics industry, Chen, Lee, Tsui and Yu (2012) found that climates that are innovative, open and flexible are highly correlated with satisfaction with one's supervisor. They stressed the OMJ 17,3 importance of supervisors in building and supporting innovative climates. Therefore, we predict when satisfaction with the supervisor is high, the job context factors previously discussed will be more strongly and positively related to OID: H4. Satisfaction with the supervisor moderates the relationships between procedural justice, training opportunities and innovation and OID such that these relationships are stronger when satisfaction with supervisor is high.
H5. Satisfaction with the supervisor moderates the indirect relationships between procedural justice, training opportunities and innovation and job satisfaction through OID such that these relationships are stronger when satisfaction with supervisor is high.
H6. Satisfaction with the supervisor moderates the indirect relationships between procedural justice, training opportunities and innovation and affiliation commitment through OID such that these relationships are stronger when satisfaction with supervisor is high.

Method
Participants Subjects (full-time employees) were recruited by undergraduate students attending a private university in the Southeast region of the USA. Thirty-four students were asked to provide contact information (name, email address) of no more than ten family and friends who were at least 21 years of age and who worked full-time (i.e. 35 h or more per week). Each student submitted ten names/contact information; each potential participant was then sent an email invitation with a link to the survey. In exchange for their help in recruiting subjects, students received a nominal amount of course credit. This method of data collection has been successfully used in several studies (Liu, Perrewé, Hochwarter, & Kacmar, 2004;Treadway, Hochwarter, Kacmar, & Ferris, 2005).
To ensure the integrity of the data, we found and excluded erroneous cases. For instance, respondents who self-reported their age below 21 years of age were eliminated as were respondents who provided the same response ("flat lining") across 65% or more of the survey. In addition, respondents who took less than six minutes to complete the survey were dropped. We included several "attention check" questions (e.g. "The answer to this question is 2.") to confirm that the participants were reading the questions. We removed all cases in which the participants did not respond correctly to these questions. In all, 247 of 340 individuals responded to the electronic invitation and successfully completed the survey instrument (72% yield rate).
A total of 56% of the respondents were female and 76% were Caucasian. The majority (59%) were at least 36 years of age and 42% had worked for their organization at least five years.

Materials
Unless otherwise noted, the items in the measures described below were scored using a fivepoint scale (5 = high and 1 = low): Procedural justice. We assessed procedural justice using the seven-item measure (a = 0.82) developed by Colquitt (2001). We asked participants to refer to the procedures their immediate supervisor uses to make decisions about pay, rewards, evaluations, promotions, assignments, etc., and to respond on a scale where 1 = "to a Role of supervisor support small extent" and 5 = "to a large extent." A sample item is "To what extent can you influence the decisions arrived at by those procedures?" Training opportunities. We used the five-item measure (a = 0.91) developed by Hung-Wen (2013), as adapted by Taormina (1994), to capture organizational training opportunities. A sample item is "This organization has provided excellent job training for me." Innovation. We used the nine-item measure (a = 0.86) developed by Dobni (2008) to assess organizational innovation. A sample item is "We have an innovation vision that is aligned with projects, platforms, or initiatives." Satisfaction with supervisor. We used Hatfield, Robinson and Huseman's (1985) measure of satisfaction with supervisor. This measure included four items (a = 0.88) with anchors that denoted high and low supervisor qualities (e.g. Unfriendly/Friendly). OID. We used the six-item measure (a = 0.84) developed by Mael and Ashforth (1992) to capture OID. A sample item is "This organization's successes are my successes." Job satisfaction. We used Cammann, Fichman, Jenkins, and Klesh's (1983) three-item (a = 0.85) measure of job satisfaction. A sample item is "All in all, I am satisfied with my job." Affiliation commitment. We used a three-item measure (a = 0.75) developed by Balfour and Wechsler (1996) to assess affiliative commitment. A sample item is "I feel a strong sense of belonging to this organization." Procedure Using MPLUS 7.4, we used structural equation modeling techniques to analyze our data. We began by estimating a measurement model. Second, we reran the model after adding an uncorrelated method factor to the measurement model to analyze the impact of common method variance in our study. Next, we added paths to the measurement model to estimate the fully mediated model shown in Figure 1 to test for indirect effects. Finally, we added satisfaction with supervisor as a first stage moderator and estimated the model again to test for interactive effects.

Results
The means, standard deviations and correlations among the variables used in our study are shown in Table 1. As expected, the variables in our study were positively correlated, but none of the correlations exceeded 0.60. 3.82 0.68 0.42 *** 0.31 *** 0.24 *** 5. Satisfaction with the supervisor 4.11 0.92 0.56 *** 0.45 *** 0.34 *** 0.38 *** 6. Job satisfaction 3.90 0.74 0.49 *** 0.36 *** 0.28 *** 0.48 *** 0.47 *** 7. Affiliation commitment 3.75 0.74 0.46 *** 0.37 *** 0.24 *** 0.56 *** 0.59 *** 0.55 *** Notes: N = 247; * p < 0.05; ** p < 0.01; *** p < 0.001 OMJ 17,3 Our first step was to estimate a measurement model to confirm that our scales demonstrated discriminant validity. The model fit the data (X 2 (608) = 1005, CFI = 0.97, TLI = 0.97, RMSEA = 0.051, SRMR = 0.065) and all path loadings were significant. Given the way our data were collected, we next tested for common method variance by adding an uncorrelated method factor to the measurement model. According to Williams, Cote, and Buckley (1989), if model fit is significantly improved with the addition of an uncorrelated method factor, common method variance may be of concern. The fit statistics after adding an uncorrelated method factor improved only slightly (CFI = 0.98, TLI = 0.98, RMSEA = 0.041, SRMR = 0.056). However, the chi-square difference test between these two models indicated a significant difference (X 2 diff (37) = 421, p < 0.05). Based on the significant chi-square test, we next calculated the variance explained by the method factor by summing the squared loadings. In our case, common method variance accounted for 10% of the total variance which is much smaller than the 25% observed by Williams, Cote and Buckley (1989). Overall, the results of the common method variance analyses suggest that the measurement model does benefit from the addition of an uncorrelated method factor, but the gain in fit is quite small, and more importantly the method factor appears to account for little variation in the data.
Following Muthén's (2004) recommendation, we next added paths to our measurement model to estimate the fully mediated model shown in Figure 1 prior to testing for moderation. As suggested by Taylor, MacKinnon and Tein (2008), we used a bias corrected bootstrapping approach to determine the significance of the indirect effects in our models (Shrout & Bolger, 2002). Results for the indirect effects are shown in Table 2. As can be seen there, the indirect effects for procedural justice and training opportunities as predictors were significant, while none of the paths for innovation as a predictor were significant. Thus, H1a, H1b, H2a and H2b were supported, while H3a and H3b were not.
In our final step, we added satisfaction with supervisor as a first stage moderator and reestimated the model. As can be seen in Figure 2, two of the interactions were significant. Thus, H4b was not supported. To determine if the form of the significant interactions aligned with our predictions, we graphed the significant interactions. In Figure 3, while both relationships between procedural justice and OID are positive, simple slope analyses revealed that the high satisfaction with supervisor line is stronger (t = 4.50, p < 0.00) than the low satisfaction with supervisor line (t = 2.55, p < 0.05). Thus, H4a was supported. Turning to Figure 4 we see that only the slope for the low satisfaction with supervisor line is significant (low t = 2.37, p < 0.05 high t = 0.70, p = 0.49). Thus, H4c was supported.

Role of supervisor support
Finally, we explored the moderated mediation results shown in Table 3. The indirect effect from procedural justice to job satisfaction through OID when satisfaction with supervisor was high was 0.43 and 0.18 when low. Further, the difference between these two effects was significant (0.24, p < 0.01) indicating a significant strengthening of the indirect effect when individuals worked for a supervisor with whom they were satisfied. A similar effect was found for the indirect effect from procedural justice to affiliative commitment through OID. The indirect effect when satisfaction with supervisor was high was 0.56 and 0.24 when low. Again, the difference between these effects was significant (0.32, p < 0.01) indicating a strengthening in the indirect effect when satisfaction with one's supervisor was high. Thus, H5a and H6a were supported. As none of the other moderated mediation results were significant, no support was found for H5b, H6b, H5c or H6c.

Discussion
The purpose of this study was to examine the effects of procedural justice, training opportunities and innovation on job satisfaction and affiliation commitment via the mediating effect of OID. Our intent was to focus on the work environment or situational antecedents to OID which have not yet received sufficient attention in the OID literature. If identification with the organization is indeed a good thing, organizations which take steps to ensure that organizational procedures and processes are fair and equitable, and training opportunities plentiful, will increase OID and the subsequent satisfaction and commitment of employees. Additionally, supervisors who work with employees to build positive, supporting relationships will find enhanced identification as a result. To the extent that increased OID ultimately and positively impacts top line growth while reducing the costs of dysfunctional turnover, the structuring of job context to support procedural justice perceptions and employee training would seem to be useful interventions.
There are several strengths of this research that bear mentioning. First, this research advances theory in a meaningful way by integrating OID as a mediator and including a new moderator in a model of OID. By confirming full mediation of procedural justice and training opportunities (job context factors) on job satisfaction and affiliation commitment through the influence of OID, we offer support for OID as a useful cognitive linking variable between individual perceptions of identity/self-worth and outcomes. We also believe that the data analytic approaches undertaken by this research provide confidence in our results and  Role of supervisor support conclusions. We investigated the validity of our measurement instruments and analyzed the relational structure of our model and the factor structure of our measurement scales. We also tested for common method variance (CMV) using a procedure outlined by Williams, Cote, and Buckley (1989). Among the limitations of our research is the fact that we used both supervisor and organizational referents for the independent variables in our model. This was done for two reasons. First, the procedural justice scale used the supervisor as the referent, so we did too. The procedural justice items were designed around the "procedures used by employees' supervisors in making such decisions" (Colquitt, 2001, p. 395) and much of the procedural justice research uses the supervisor as referent (De Cremer & Sedikides, 2005;Scott & Colquitt, 2007). Although the supervisor is the referent, he/she is the conveyor of organizational policies (Naumann & Bennett, 2000). The essential determinant of procedural justice is whether the employee has voice in the decision-making process and feels as though the outcome can be influenced (Thibaut & Walker, 1975). Second, we wanted variety in the referent of our predictors to ensure alignment with our outcomes as one was at the individual level (job satisfaction) and other was at the organizational level (affective commitment).
We failed to find an effect for innovation as a predictor of OID. Perhaps our measure of innovation did not adequately capture the extent to which organizational prestige due to innovation and/or an innovative culture was reflected in our subject's identification with the organization. An additional limitation is that our data were collected by self-report. It would have been beneficial to temporally separate the predictor and criterion variables to reduce CMV, but as we performed a post hoc analysis to rule out CMV, we are confident that the model relationships as investigated are robust. We also understand that there may be multiple ways in which satisfaction with supervisor could be causally involved in the process (as opposed to being a moderator, as we propose in our model). It is also possible to hypothesize reversed paths in our model. As a post hoc robustness check, we estimated a reversed paths model in which the independent variables became the dependent variables and vice versa. This model fit significantly worse than the hypothesized model (DX 2 (2) = 23.43, p < 0.001). Therefore, while the hypothesized model presents just one representation of how the constructs could be theoretically related, our post hoc analysis offers additional support for this model. And finally, while we believe that the high response rate is a significant strength of this research and may be due to the strong personal relationship between the faculty/research team member and the students who recruited participants, it is equally likely that the opportunity to earn extra credit contributed to the strong response.

Conclusion
Research has consistently demonstrated that stronger employee-organization relationships lead to positive individual and organizational outcomes. Our results confirm the central and important role that supervisors play in enhancing subordinate identification with the organization. As has been demonstrated in prior research, individuals who identify with their organization are more likely to work hard for the organization, are less likely to be absent and are less likely to leave the organization. In this research, we found that individuals who identify with their organization will also be more satisfied and more committed to their work. These are all positive and valued outcomes, and OID is seen as central to understanding those relationships. We also found that perceptions of fairness (procedural justice) and training opportunities for employees enhance OID. Given the changing nature of the employment contract, and the rise of a more contingent workforce, the days of binding employees to the organization for life long since passed. However, the goal of enhancing OID should not be to develop great employees for the long term, but to develop great employees for however long they choose to stay with the organization. Given these new labor market realities, organizations which work to provide positive, supportive, and fair environments for employees will attract the best employees both short and long term. If there is such a thing as a sustainable competitive advantage for human resource leaders in organizations, it might just be in developing systems and processes to support and enhance OID regardless of which individuals occupy organizational positions. Great organizations are composed of great people, but great people often come and go. Perhaps, then, what makes organizations great is that they continue to attract and develop great people. Building identification with the organization would seem to be a useful place to start that process.