Organizational Socialization and Collective Self-Esteem as Drivers of Organizational Identification

Organizational identification can explain various employees’ attitudes, so that it is thought to be a root construct in organizational phenomena. Although organizational identification means process, no previous studies have focused on the process by which organizational identification is developed. This study reveals the mechanisms involved in developing organizational identification. For this paper, a model was constructed from the perspective of Bandura’s modeling theory. Three hypotheses emerged: (1) Organizational socialization influences organizational identification positively, (2) collective self-esteem influences organizational identification positively, and (3) organizational socialization influences collective self-esteem positively. These were tested by means of SEM. Through this model, this paper demonstrates a new process of organizational identification via a new perspective.


Introduction
In this paper, a model to explain the process about how organizational identification (OID) is developed, is proposed and is empirically tested.This model will help to draw out efficiently various attitudes which result from OID.
Although OID is a process (Foote, 1951;Rotondi, 1975), past researchers have focused primarily on the factors determining OID rather than on the process itself (Pratt, 2000).Accordingly why OID develops is still vague, and it is difficult to enhance OID efficiently because the OID process is "black box."Previous researchers have shown various effects (e.g., affective commitment, cooperation, psychological contrast, organization citizenship behavior).These variables are respectively important for organization and need to be managed through OID.However even if OID is important for organizational phenomena, it is not easy that organization manages employees' OID because the process is still vague.Thus demonstrating the OID process has important implication.The present paper aims to detail the process by which OID is developed.Dutton, Dukerich, and Harquail (1994) defined OID as "the degree to which a member defines him-or herself by the same attributes that he or she believes define the organization," and cognitive connection.Recent researchers have been more likely to emphasize OID as a cognitive process (e.g., Mael & Ashforth, 1992;Rousseau, 1998), but Freud (1922) originally defined OID as an affective process.Harquail (1998) and Van Dick (2001) argued that OID is related to both cognitive and affective factors, and that these factors should not be separated from each other.In fact, social identity theory, which is central to OID research, includes both cognitive and affective factors (cf.Tajfel & Turner, 1979;Hogg & Abrams, 1988).Therefore, focusing on both types of factors, this paper defines OID as both a cognitive and an affective process, and as the process to get the perception and feeling of oneness with the group (Johnson, Morgeson & Hekman, 2012).
A review of the literature on the OID process reflects the impact of Lewin's (1947) work.Researchers such as Fiol (2003) and Ashforth (1998) have based their models on Lewin's.Although Pratt (2000) is not directly influenced by Lewin's work, he seems to have paid attention to it.To explain how people change their social habits, Lewin proposed a three-stage model with the phases "unfreezing," "moving," and "re-freezing."The aforementioned studies propose a relatively similar process because they are based on a simple model.Bandura and Huston (1961) believed that the repertoire of children's actions increases through identification with adults, and they viewed identification as a learning process that affects attitude formation.Previous researchers have also thought that identification is the process which forms attitude and that attitude formation can be explained by learning theory (Bandura, 1977;Miller & Dollard, 1941;Rotter, 1954).In fact, Bandura interpreted modeling as involved in identification, and his modeling theory is thought to express the process of identification (Bandura, Ross & Ross, 1963;Sukemune et al., 1985).Therefore, Lewin's model alone does not necessarily explain the process of identification.Modeling theory as the process of identification may be applied to developing a theory of the OID process.Thus, this paper considers the OID process on the basis of modeling theory and aims at developing a theory of the process from a new perspective.This paper is structured as follows: Section 2 provides a literature review on OID process that consists of three parts based on Bandura's modeling theory.In the end of this section, theoretical model is constructed through a literature review above.This theoretical model is tested using data gathered from a Japanese firm.Then section 3 explains about the firm, sample property, and each variable.The result is shown in section 4. Furthermore, a discussion on the result is carried out and theoretical/practical implication is described in section 5. Finally, section 6 concludes the main findings and proposes future research.

Literature Review
Bandura attempted to explain the process through a model having the following four stages: "attention," "retention," "motor reproduction," and "motivation" (Bandura, 1971;Sukemune et al., 1985).The attention process is characterized by memorization.People learn or model many actions from others who are frequently in contact with them, and these characteristics and actions are termed modeling stimuli.The distinctive and evident accessibility and the complexity of modeling stimuli affect the attention process.
The retention process involves remembering the information.People code modeling stimuli in two ways-iconic and linguistic representation-and these two approaches help them retain information.Rehearsal of the coded action leads to longer retention of the modeling stimuli.
The motor reproduction process is characterized by the transformation of the symbolic representation in the retention process into action.When people reproduce an action, they do not always act as the model did.In such cases, they need to self-adjust the difference between the model's and their action.To achieve this, people observe and understand their actions on their own.The motivation process refers to the fact that the learned action is not always practiced intentionally, even if people could reproduce the modeling stimulus.For people to reproduce this action, they need to understand the benefit of practicing.Bandura categorized the above four stages into two segments: "acquiring" and "practicing" (Sukemune et al., 1985).The former includes attention, retention, and motor reproduction, and the latter includes motivation.The present paper adopts this categorization to simplify the discussion.

The Acquiring Stage and Organizational Identification
Certain values and attitudes are categorized as central to the organization, and employees are not incorporated as members of the organization if these values and attitudes are not inculcated into them (Schein, 1971).After organizational socialization (OS), employees are better able to distinguish the difference between values that are central and non central to the organization.When these central values and attitudes are assumed to be part of the organizational identity-which refers to the beliefs employees share about the central, enduring, and distinctive features in an organization's character (Albert & Whetten, 1985), this study assumes that people form organizational identity after socialization.Wanous, Reichers and Malik (1984) conceive OS as the process involved with acquiring an acceptable role in social learning theory.Synthesizing both thoughts, the present study considers OS a proxy variable of the acquiring stage.Socialization refers to "the process by which persons acquire the knowledge, skills, and dispositions that make them more or less effective members of their society" (Brim & Wheeler, 1966) and "the process whereby an individual acquires the knowledge, skills, attitudes, mores, and patterns of behavior that combine to shape his social personality and enable him to become a functional part of society" (Eli & Shuval, 1982).OS, which is a sub-concept of socialization, is defined as the process by which an individual acquires the attitudes, behavior, and knowledge that he or she needs to participate as an organizational member (Van Maanen & Schein, 1979).
There are two routes by which employees may identify with their organizations; "through the recognition of an organization deemed similar to one's self," or "through changes in one's self to become more similar to an organization" (Pratt, 1998).Similarly, Mussen (1966) describes identification as the result of identity formation.Chao et al. (1994), Katz (1980), Schneider andReichers (1983), andWanous et al. (1984) argue that the stages of "pre-entry" and "adaptation" considerably influence self-identity formation.Newcomers face different uncertainties in their organization.Then while they establish and clarify organizational identity through OS, they are motivated to reduce uncertainties by changing their identities (Katz, 1980).According to Brim and Wheeler (1966), the function of socialization is to transform the "gemstones" of society into good workers; the content of socialization includes understanding the society's status structure and the role of prescriptions and behavior associated with the different positions in this structure.That is to say, it is believed that as the person understands the organization's status structure and adapts to the organization, his or her identity also changes toward organizational identity because OS is the process that forms the social personality (Eli & Shuval, 1982).
As Pratt (1998) and Mussen (1966) argued, identification is the result of identity formation.Thus, the causality between OS and OID is assumed because OS is the cause of identity-change.In addition, Ashforth and Mael (1984) argued that newcomers need to learn the policy, status structure, and behavioral norms in order to understand the organization and the activation within it.Newcomers construct self-definition in the process.Most of self-definition is composed of social identity.

The Practicing Stage and Organizational Identification
The practicing stage (i.e., motivation process) in Bandura's modeling theory focuses on reproducing the acquired action oneself.According to social identity theory, people are motivated to enhance self-esteem by discriminating between ingroup and outgroup (Tajfel & Turner, 1979).This paper replaces the motivation for self-reproduction with the motivation of enhancing self-esteem.Abrams and Hogg (1990) suggest two corollaries concerning self-esteem and intergroup discrimination.Corollary 1 (C1) refers to the idea that successful intergroup discrimination strengthens social identity (i.e., develops social identification) and enhances self-esteem.Corollary 2 (C2) posits that threatened self-esteem encourages intergroup discrimination to satisfy self-esteem needs.This second corollary relates to the motivation to enhance self-esteem.Thus, this paper focuses on C2 below.
Previous studies have used the self-esteem derived from personal identity (personal self-esteem; PSE) rather than that from social identity (e.g., Brockner & Chen, 1996;Crocker, Thompson, McGraw & Ingerman, 1987;Hogg & Sunderland, 1991;Wagner, Lampen & Syllwasschy, 1986).However, in deliberating on the issue of intergroup discrimination, self-esteem in terms of social identity-that is, collective self-esteem (CSE)-is more appropriate than PSE (Long & Spears, 1997).Therefore the relationship between CSE and intergroup discrimination is discussed here.
There are two types of studies about C2.On the one hand, in support of the C2 theory, some studies show that low CSE influences intergroup discrimination positively (e.g., Long, Spears & Manstead, 1994;Rubin & Hewstone, 1998;Ruttenberg, Zea & Sigelman, 2001).On the other hand, and contrary to C2, the studies show that high CSE influences intergroup discrimination positively (e.g., Crocker & Luhtanen, 1990;Luhtanen & Crocker, 1992).Crocker, Blaine, and Luhtanen (1993) thought that persons with high CSE had high self-enhancement motives and that persons with low CSE had high self-protective motives.On this basis, Long and Spears (1997) thought it more likely that a person's having high CSE led to deriving a positive image from the ingroup and, thus, that he or she did not need to enhance social identity through intergroup discrimination.In fact, Long and Spears demonstrated that a person with high PSE and low CSE was likely to discriminate between groups.Rubin and Hewstone (1998) showed a similar result.Although Long, Spears, and Manstead (1994) hypothesized that a person with high CSE would have an enhanced social identity, the result showed that persons with low CSE engaged in intergroup discrimination.Ruttenberg, Zea, and Sigelman (2001), contrary to Crocker, Blaine, and Luhtanen (1993), proposed that a person with high CSE would have a strong self-protective motive and that a person with low CSE would have a strong self-enhancement motive.The result of the analysis demonstrated that persons with low CSE do engage in intergroup discrimination.Nonetheless, these studies do not theoretically deny that a person with high CSE is likely to enhance social identity.Crocker and Luhtanen (1990) and Luhtanen and Crocker (1992) argued that CSE mediated the degree of protecting or enhancing social identity when it is threatened.That is, CSE may influence self-protective and self-enhancement motives.People with high CSE, compared to those with low CSE, are more likely to believe that they are successful and to avoid the criticism of failure.Therefore, people with high CSE may engage in ingroup-enhancing bias or distortion when they face fear of their collective identity.On the other hand, persons According to Sachdev and Bourhis (1987), members of high-status groups favor the ingroup more than do members of low-status groups.Member of low-status groups also favor the outgroup instead of the ingroup.
Many studies have supported C1 to date (c.f., Rubin & Hewston, 1998)-that is, that intergroup discrimination or social identification enhances self-esteem.However, the opportunities for social comparison continue even after CSE has developed.Thus, it is likely that further intergroup discrimination will be done on the basis of self-esteem.We may conclude, therefore, that there is a looping causality between self-esteem and social identification.Then if C2 is correct, as causality loops, does the degree of intergroup discrimination gradually decrease?It is actually more likely that attachment for the ingroup increases and that ingroup favoritism is encouraged.

The Acquiring and Practicing Stages
Previous research has often focused on the idea that socialization enhances a person's self-esteem.Gecas and Schwalbe (1986) proposed that parental behavior (control, support, and participation) and adolescents' perceptions of their parents influenced adolescents' self-evaluation (self-worth, self-efficacy, and PSE).As a result, adolescents develop their sensitivity, interpersonal relations, and self-worth through parental behavior (socialization support).Furthermore, adolescents' socialization experiences emphasized their mastery, agency, exploratory activities, and self-efficacy.Both self-worth and self-efficacy are sub-concepts of self-esteem.Therefore, Gecas and Schwalbe (1986) argued that socialization support and adolescents' socialization experiences develop their self-esteem.Harris-Britt, Valrie, and Kurtz-Costes (2007) hypothesized that racial socialization functions as a moderator between discrimination and PSE.However, the result of the analysis showed that the main effect of racial socialization influenced PSE significantly in hierarchical multiple regression, though the main effect of discrimination did not.This result may imply that racial socialization influences PSE directly and this result accords with previous studies that PSE is enhanced by messages about cultural knowledge and ethnic pride (i.e., the message of racial socialization; cf. the review of Harris-Britt et al., 2007).According to Constantine and Blackmon (2002), people strengthen prejudice toward outgroups through racial socialization and as a result raise ingroup consciousness.Thus, they thought that racial socialization might enhance PSE.
The above studies mentioned that racial socialization influenced PSE, though Yoon (2001) demonstrated a critical result concerning this simple causality.He assumed that CSE functions as a mediator between racial socialization and PSE.That is, a person develops racial identity (i.e., enhances CSE) through racial socialization.On the basis of racial identity, people reevaluate themselves and improve their self-definition, resulting in positive PSE.This relationship between racial socialization and CSE is correspondent with that expressed in other studies.Katz, Joiner, and Kwon (2002) focused on members of devalued groups and analyzed the mechanism that led to emotional distress.Katz et al. explained that members of devalued groups attained membership through socialization and that members may lower CSE on the basis of the group evaluation.This study considered women to be a devalued group and men as a valued group.This logic implies that if a person is male, he would enhance CSE through socialization.Furthermore, Yoon (2004) asserted that socialization leads to ethnic consciousness, translates into positive evaluation, and relates to CSE.In fact, his results demonstrated that parental support in terms of ethnic socialization influenced children's CSE.
Few studies exist on the relationship between OS and CSE.However, OS is thought to derive from socialization, and the essence of both is same (Hayashi, 2013).Thus, this paper's hypotheses below are based on the studies referred to above.

Method
The present study analyzed a Japanese semiconductor-manufacturing company that has 21 production bases worldwide.The HR department of this company was contacted and questionnaires were distributed during induction sessions for promoted employees.They answered the questionnaire in the presence of the HR staff.Japanese companies, in particular manufacture companies, tend to make much of "make" strategy to their employees, compared to western companies which tend to make much of "buy" strategy.According to Rousseau (1998), the company using only "buy" strategy may hinder employees' OID.That is, when we investigate the process of OID, such Japanese companies are likely to be appropriate.

Sample
Although 250 employees answered the questionnaire, the number of valid responses was 236.The sample was composed of 133 senior staff, 65 assistant managers, and 38 managers.There were 232 males and 4 females.Ages ranged from 28 to 50 years (M = 34.32,SD = 4.99).Tenure ranged from 1 to 27 years (M = 10.37,SD = 5.81).

Measures
The questionnaire contained three parts, covering OID, CSE, and OS.

Organizational Identification (Note 1)
For this study, the 6-item OID scale from Mael and Ashforth (1992) was used.Respondents indicated on a 7-point Likert scale the extent to which they agreed or disagreed with each given statement: (1) when someone criticizes my company, it feels like a personal insult; (2) I am very interested in what others think about my company; (3) when I talk about this company, I usually say "we" rather than "they"; (4) this company's successes are my successes; (5) when someone praises this company, it feels like a personal compliment; and (6) if a story in the media criticized the school, I would feel embarrassed.The coefficient alpha for this scale was .805.

Collective Self-Esteem
A 10-item CSE scale from modified Pierce et al. (1989) was employed.Respondents indicated on a 7-point Likert scale the extent to which they agreed or disagreed with each given statement: (1) my company counts for me; (2) my company take me seriously; (3) my company is important; (4) my company is trusted; (5) there is faith in my company; (6) my company can make a difference; (7) my company am valuable; (8) my company is helpful; (9) my company is efficient; and (10) my company is cooperative.The coefficient alpha for this scale was .847.

Organizational Socialization
The third instrument used was the 35-item OS scale from Haueter, Macan, and Winter (2003).Respondents indicated on a 7-point Likert scale the extent to which they agreed or disagreed with each given statement: (1) I know the specific names of the products/services produced/provided by this organization; (2) I know the history of this organization (e.g., when and who founded the company, original products/services, how the organization survived tough times); (3) I know the structure of the organization (e.g., how the department fit together); (4) I understand the operations of this organization (e.g., who does what, how sites, subsidiaries and/or branches contribute); (5) I understand this organization's objectives and goals; (6) I understand how various departments, subsidiaries, and/or sites contribute to this organization's goals; (7) I understand how my job contributes to the larger organization; (8) I understand how to act to fit in with what the organization values and believes; (9) I know this organization's overall policies and/or rules (e.g., chain of command, who is influential, what needs to be done to advance or maintain good standing); (10) I understand the internal politics within this organization; (11) I understand the general management style (e.g., top-down, participative) used in this organization; and ( 12) I understand what is meant when members use language (e.g., acronyms, abbreviations, nicknames) particular to this organization.The coefficient alpha for this scale was .834.

Control Variable
Respondents were asked to answer when they entered this company.And tenure was calculated using present year and this number.This tenure was treated as control variable, confounding variable.

Results
Structural Equation Modeling (SEM) was conducted with AMOS to determine if the proposed model fitted the data.Model fit was evaluated by using several fit indices, including χ2, CFI, GFI, AGFI, and RMSEA.Table 1.Factor loading of confirmatory factor analysis First, confirmatory factor analysis using SEM was conducted and the factor loading was shown in Table 1.Almost all items showed the high factor loading.Next, path analysis using SEM was conducted (see Figure 2).The significance level of the path coefficient was set at 1%.The fit indices were χ2 = 586.619,CFI = .945,GFI = .845,AGFI = .809,and RMSEA = .055.This model showed high fit to the data.As the result revealed, the factors of OS affected both CSE and OID positively and significantly; the path coefficients were .777and .342,respectively.These results supported H1 (OS influences OID positively) and H3 (OS influences CSE positively).H2 (CSE influences OID positively) was also supported because CSE affected OID positively and significantly: the path coefficient was .522.We can think this model is a partial mediation model because of good fit indices.Furthermore, tenure affected OS and CSE significantly.This means that the longer an employee works in an organization, the more OS and CSE develop.However, tenure did not affect OID significantly.Therefore, tenure indirectly influences OID through OS and CSE.

Discussion and Implications
The three hypotheses were supported: (1) OS influences OID positively, (2) CSE influences OID positively, and (3) OS influences CSE positively.This means that two routes were demonstrated.On the one hand, as employees enter the organization and socialize, their identities will be gradually changed toward the organizational identity.This change of identity, in turn, leads to OID.On the other hand, employees become members of their organizations through socializing.As a result, they will regard their organizations more positively because they recognize themselves as members of the organizations and develop an increasing awareness of their organization's evaluation.Employees wish to regard themselves positively; thus, they identify with their organization, which in turn gains a positive evaluation because of their self-enhancement motive.
However, this paper posited that the organization gains a positive evaluation from its surroundings.This assumption supports the above logic-that employees will regard their organizations more positively because they recognize themselves as members of the organizations and become increasingly aware of their organizations' evaluation.Previous research has also supported the importance of this assumption (e.g., Ashforth & Mael, 1989;Carmeli, Gilat & Weisberg, 2006;Dukerich, Golden & Shortell, 2002;Dutton, Dukerich & Harquail, 1994;George & Chattopadhyay, 2005;Mael & Ashforth, 1992;March & Simon, 1958;Smidts, Pruyn & Van Riel, 2001), while for example Tajfel and Turner (1979) and Elsbach and Kramer (1996) argued about the possibility that organizational identity might be threatened and gain the low evaluation.It means this assumption is not always supported.Thus, in the present study, this process of OID may be applied only to an organization with a positive evaluation.
This paper contributes to an understanding of the OID process.Although previous researchers have used Lewin's model (1947) to explain the OID process, the present paper used Bandura's model (1971).This difference between the models leads to some different results.For example, in Lewin's model, the first step is to "unfreeze" the habit.This step may be useful to employees who have changed their jobs and have experienced organizational change because they already have organizational identity in the previous organization.This step may not be useful, however, to newcomers who begin to work in an organization for the first time because they do not have a previous organizational identity to unfreeze.On the other hand, in Bandura's model, the first step is to acquire the way to work and the organizational identity.This step may occur after Lewin's "unfreezing."Thus, this step may be useful to newcomers and or to employees who have already unfrozen their previous organizational identity.In other words, Bandura's model explains the process by which employees identify with their organizations more fully than Lewin's model (which instead explains the process by which employees "disidentify" with their organizations).Therefore, the present paper may explain the OID process better than previous studies have.
This study has also demonstrated the relationship between OS and OID.From the beginning, it has been observed that the two concepts have the following difference.Because the environment and the person are both supposed to be subjects in OS (e.g., Brim & Wheeler, 1966), aspects of education and control have always existed in the study of OS.However, because only the person is assumed to be the subject in OID (e.g., Freud, 1922), the aspects of education and control have not been predominant in the study of OID.From the results of this study, the possibility of educating people so as to develop their OID was demonstrated by presenting the relationship between both concepts.Thus, this paper has demonstrated the possibility of managing OID.
Previously, many researchers have showed the positive relationship between OID and tenure (e.g., Foote, 1951;Hall & Schneider, 1972;Hall, Schneider & Nygren, 1970;Lee, 1971;Mael & Ashforth, 1992;March & Simon, 1958).They didn't think tenure affect OID directly, but indirectly through satisfying employee's needs (March & Simon, 1958), making certain investments in their careers (Hall & Schneider, 1970), and so on.We supported their assertions because tenure affected OS and CSE significantly, but not OID significantly.That is, this result emphasizes that OS is important to develop OID.
In addition, this paper also contributes to the practical side.First, the source of a company's reputation is organizational identity (Fombrun & Van Riel, 2004).Therefore, employees play an important role in forming reputation.Friedman (2009) argued that employees' OID becomes a mediator when human resource functions influence reputation formation.Other researchers argue that OID is important for the influence of organizational identity on reputation (Fombrun & Van Riel, 2004;Lichtenstein, Netemeyer & Maxham, 2010).Therefore, if the organization can manage employees' OID, it would lead to forming the favorable reputation.The present paper demonstrated the process of OID, so it is likely that the organization can form the favorable reputation.
Second, when an organization wishes to develop OID through its employees, it first needs to develop OS tactics (Van Maanen & Schein, 1979).At the same time, it needs to differentiate itself from other organizations because it needs to be evaluated positively by its surroundings, which include the employees and applicants.As mentioned above, the core concept of reputation is organizational identity, and OID is important if organizational identity is to influence reputation.According to the definition of organizational identity, an organization that has organizational identity perceives the ways in which it is different from other organizations.Thus, the organization needs to self-evaluate to know itself, to be aware of its evaluation from its surroundings, and to gain a positive evaluation.This implication is applied to the formation of a favorable reputation.Thus, having a good understanding of the process of OID development may influence the formation of a favorable reputation.
Third, according to Rousseau (1998), "making" strategy increases the degree for employees to identify with their organization.We support this argument and explain how "making" strategy develops employee's OID.When a firm recruits employees from all over the world, it becomes difficult to retain them because their value and culture are diverse.In terms of resource based view (RBV; Barney, 1991), employees' retention is important for the firm's competitive advantage.Therefore, this study suggests the way to retain employees because OID leads to attachment to their organization, and in turn turnover reduction.

Conclusion
The aim of this research was to develop new OID process model on the basis of Bandura's modeling theory.The model consisted of OS, CSE, and OID.Then the relationship between these concepts was tested using SEM.As a result, we demonstrated how employees identified with their organization.OS influenced OID both directly and indirectly.This means that OS is important for OID and the possibility that organization manage employees' identification with it.If organization manages employees' OID successfully, it may lead to favorable organizational reputation.Favorable reputation draws many talents to the organization, and it functions to decrease employees' turnover intention.That is, OID can be the source of organizational competitive advantage.
This study used cross-sectional data and self-completion design for an analysis.This data is not well suited to the consideration of causal processes.So this model needs to be tested using panel data.Furthermore, this data was gathered from one company which is within one industry.It is likely to be influenced by industrial characteristics.Thus, future work which uses cross-industrial data and cross-cultural data would be expected.

Figure 1 .
Figure 1.A model of this research This model provides various determinants of OID, which were researched by previous studies, with the framework which can classify them.Based on this model, it can be demonstrated which steps (OS, CSE, or OID) are influenced by factors.Future works need to focus on what kind of determinants have larger influence over OID.