Attachment and caregiving (Bowlby, Reference Bowlby1982) are core psychological constructs that contribute to understanding human relationships (Mikulincer & Shaver, Reference Mikulincer and Shaver2017). In Bowlby's (Reference Bowlby1982) terminology, attachment represents the individual's motivation to ensure emotional bonds between people by seeking proximity in times of need, whereas caregiving represents the individual's motivation to ensure emotional bonds by means of providing support to others. Although the mutual contributions of attachment and caregiving have been extensively investigated in social contexts (e.g., Mikulincer & Shaver, Reference Mikulincer and Shaver2017; Reizer, Ein-Dor, & Shaver, Reference Reizer, Ein-Dor and Shaver2014), they have rarely been examined empirically in an organizational setting.
In the last decade, as part of the growing tendency in the organizational domain to move from an individualistic perspective to a more relational-based perspective, there has been a growing interest in the way that individual differences in attachment styles affect organizational outcomes (for reviews, see Harms, Reference Harms2011; Paetzold, Reference Paetzold, Simpson and Rholes2015; Yip, Ehrhardt, Black, & Walker, Reference Yip, Ehrhardt, Black and Walker2018). Organizational studies have provided robust evidence concerning the role of attachment personality styles in predicting organizational citizenship behavior (OCB) (Little, Nelson, Wallace, & Johnson, Reference Little, Nelson, Wallace and Johnson2011) and employee job performance (Ronen & Zuriff, Reference Ronen and Zuroff2017). This stream of research continues to flourish, with nearly 50% of the papers on attachment at the workplace published after 2010 (Yip et al., Reference Yip, Ehrhardt, Black and Walker2018: 186).
While prior research has established the nature of the relationships between attachment dimensions and workplace outcomes, very little is known regarding the context by which these associations occur and whether group variables might moderate these associations. And indeed, recent integrative reviews advocated that the next stream of empirical studies should focus on identifying the moderators for attachment at the workplace (Harms, Reference Harms2011; Harms, Bai, & Han, Reference Harms, Bai and Han2016) in general, and regarding group variables in particular (Yip et al., Reference Yip, Ehrhardt, Black and Walker2018). Therefore, the current study sought to examine the joint contribution of attachment and caregiving in predicting the organizational equivalent of prosocial behavior, termed OCBs (Bolino & Grant, Reference Bolino and Grant2016). In addition, we examined the moderating role of group cohesion in the associations between attachment, caregiving, and OCB.
OCB can be defined as ‘individual behaviors that are discretionary, not directly or explicitly recognized by the formal reward system and promote the effective functioning of the organization' (Organ, Reference Organ1988: 4). Organ (Reference Organ1988) referred to OCB as a multidimensional concept consisting of altruism, courtesy, sportsmanship, civic virtue, and conscientiousness. Reference Williams and AndersonWilliams and Anderson's (1991) dimensions were largely based on Reference OrganOrgan's (1988) five-dimensional taxonomy. They suggested a parsimonious two-factor conceptualization, comprised of organizational citizenship behavior-individual (OCBI) and organizational citizenship behavior-organizational (OCBO). OCBI is more individually targeted and comprises behaviors targeting other individuals in the workplace (e.g., willingness to help new colleagues adjust to their work environment), whereas OCBO is more impersonal and encompasses behaviors directed at the organization (e.g., arriving early at the office to start work; Organ, Podsakoff, & Podsakoff, Reference Organ, Podsakoff, Podsakoff and Zedeck2011). Previous research suggested that OCBI and OCBO may have different antecedents (e.g., Ilies, Fulmer, Spitzmuller, & Johnson, Reference Ilies, Fulmer, Spitzmuller and Johnson2009) and moderating processes (Ozer, Chang, & Schaubroeck, Reference Ozer, Chang and Schaubroeck2014), recommending that there may be value in examining OCBI and OCBO separately. As noted, the current study adopted the multidimensional perspective of OCBs by investigating both OCBI and OCBO from a relational personality perspective.
As OCBs are considered to be discretionary, a substantial body of research has explored relationships between dispositional indicators, such as the Big Five personality dimensions (Chiaburu, Oh, Berry, Li, & Gardner, Reference Chiaburu, Oh, Berry, Li and Gardner2011), dispositional affect (Imer, Kabasakal, & Dastmalchian, Reference Imer, Kabasakal and Dastmalchian2014), and OCBs. However, less attention has targeted the relational dispositions that underlie citizenship behaviors (Little et al., Reference Little, Nelson, Wallace and Johnson2011). Consistent with previous research, this study investigates OCBs from a relational personality perspective, relying on Reference BowlbyBowlby's (1982) attachment and caregiving conceptualization.
Individual Difference in Attachment Style
Attachment theory is currently recognized as one of the most influential approaches in developmental, personality, and social psychology (Mikulincer & Shaver, Reference Mikulincer and Shaver2017). Our study is grounded in the underlying assumptions of attachment theory (Bowlby, Reference Bowlby1982). Attachment theory maintains that early interactions between children and their primary caregivers may explain the presence of individual differences in how people connect with and relate to others during adulthood and in social and interpersonal situations (Mikulincer & Shaver, Reference Mikulincer and Shaver2017). Whereas some individuals grow to feel safe and secure in the world and in their relationships with others, some develop a chronic sense of insecurity, either by exhibiting overt signs of anxiety or by becoming withdrawn and avoidant.
During adulthood, these individual differences in attachment can be conceptualized as regions in a two-dimensional space: anxiety and avoidance (Brennan, Clark, & Shaver, Reference Brennan, Clark, Shaver, Simpson and Rholes1998). Anxiously attached individuals tend to manifest a lack of confidence regarding others' reactions, perceiving themselves as unworthy of love, and holding negative images of themselves in relationships. These individuals often become preoccupied with their relations, adopting behaviors aimed at eliciting affection and support from others. Conversely, avoidant individuals tend to be self-reliant, favoring emotional distance or detachment in day-to-day interpersonal relationships as a way to cope with relational tensions and disappointments. People scoring low on both the anxiety and avoidance dimensions are considered securely attached and typically express trust and confidence in the goodwill of others during interpersonal interactions (Mikulincer & Shaver, Reference Mikulincer and Shaver2017) and provide more support to their spouses (Reizer, Ein-Dor, & Possick, Reference Reizer, Ein-Dor and Possick2012).
Acknowledging the relational nature of work and the significance of interpersonal interactions at work (Blustein, Reference Blustein2011) has recently led organizational scholars to investigate organizational outcomes through the lens of Bowlby's framework (Harms, Reference Harms2011; Paetzold, Reference Paetzold, Simpson and Rholes2015; Wu & Parker, Reference Wu and Parker2017; Yip et al., Reference Yip, Ehrhardt, Black and Walker2018). Specifically, it has been noted that the two-dimensional attachment conceptualization predicted organizational outcomes beyond those predicted by the Big Five personality dimensions (Richards & Schat, Reference Richards and Schat2011). These investigations have revealed several additional advances that have expanded our understanding of the contribution of the attachment avoidance and anxiety dimensions to leadership (e.g., Thompson, Glasø, & Matthiesen, Reference Thompson, Glasø and Matthiesen2018; Wu & Parker, Reference Wu and Parker2017) and employment relationships (Albert, Allen, Biggane, & Ma, Reference Albert, Allen, Biggane and Ma2015; Crawshaw & Game, Reference Crawshaw and Game2015). In the following sections, we will specify how these two dysfunctional attachment styles relate to OCBs.
The Association Between Attachment and Ocb
Attachment styles are believed to be fundamental for understanding interpersonal processes at work as they represent the personal perceptions and motivation to engage in close, supportive interactions (Harms, Reference Harms2011; Harms, Bai, & Han, Reference Harms, Bai and Han2016). Therefore, previous work has suggested that attachment may be associated with OCB, which is considered an interpersonal and relational organizational outcome (Harms, Bai, & Han, Reference Harms, Bai and Han2016; Paetzold, Reference Paetzold, Simpson and Rholes2015). Based on their previous experiences, attachment-avoidant individuals tend to possess a negative and detached approach toward others (Mikulincer & Shaver, Reference Mikulincer and Shaver2017). They do not doubt their self-worth but rather believe that the other person is unworthy of their attention. As they are nontrusting of others, they prefer to keep their distance from others and do not rely upon them (Brennan, Clark, & Shaver, Reference Brennan, Clark, Shaver, Simpson and Rholes1998). Specifically, avoidant individuals are unlikely to help others or engage in OCBI. Furthermore, as they are less alert to social cues and social interactions, they tend to miss opportunities for engaging in OCBI (Harms, Bai, & Han, Reference Harms, Bai and Han2016; Little et al., Reference Little, Nelson, Wallace and Johnson2011; Richards & Schat, Reference Richards and Schat2011; Syna Desivilya, Sabag, & Ashton, Reference Syna Desivilya, Sabag and Ashton2006). Finally, avoidant individuals try to avoid social meetings and social interactions (Hazan & Shaver, Reference Hazan and Shaver1990), which ultimately decreases their OCBO as well (Syna Desivilya, Sabag, & Ashton, Reference Syna Desivilya, Sabag and Ashton2006).
Individuals characterized by anxious attachment lack the expectations that other individuals will be there for them when they need help, but they are still eager for close relationships. Their lack of self-worth, along with their intensified feelings of being underappreciated, affect their tendency to engage in extreme behaviors that keep the other person close to them, thus satisfying their need for constant reassurance that they are accepted and loved (Mikulincer & Shaver, Reference Mikulincer and Shaver2017) and seek proximity to others (Richards & Schat, Reference Richards and Schat2011).
These tendencies can suggest two contradicting expectations regarding anxiously attached individuals and OCB. On the one hand, anxious individuals may express some motivation – even without possessing some of the necessary social skills – to engage in helping behavior (Collins & Ford, Reference Collins and Ford2010), thus, suggesting a positive association between anxiety and OCBs. On the other hand, anxious individuals tend to feel underappreciated at work, experience interpersonal conflicts (Hazan & Shaver, Reference Hazan and Shaver1990; Paetzold, Reference Paetzold, Simpson and Rholes2015), and lack the ability to regulate negative emotions and distress at work (Harms, Reference Harms2011; Reizer, Reference Reizer2015; Yip et al., Reference Yip, Ehrhardt, Black and Walker2018), thus possibly driving them to reduce OCB.
Indeed, previous research has reported inconclusive findings regarding the relationship between anxiety and OCBs. Whereas some studies have indicated that attachment anxiety was negatively associated with OCBI and OCBO (Syna Desivilya, Sabag, & Ashton, Reference Syna Desivilya, Sabag and Ashton2006), others found nonsignificant associations (Harms, Bai, & Han, Reference Harms, Bai and Han2016; Little et al., Reference Little, Nelson, Wallace and Johnson2011; Richards & Schat, Reference Richards and Schat2011). However, studies examining attachment and OCBs have typically used a cross-sectional design (Harms, Bai, & Han, Reference Harms, Bai and Han2016; Little et al., Reference Little, Nelson, Wallace and Johnson2011; Richards & Schat, Reference Richards and Schat2011) or measured self-reported OCBs (Syna Desivilya, Sabag, & Ashton, Reference Syna Desivilya, Sabag and Ashton2006). Furthermore, although attachment literature usually acknowledges the need to examine attachment impacts at different stages of relationship forming, such as during the relationship's early stages, rather than during more stable stages of acquaintance (Mikulincer & Shaver, Reference Mikulincer and Shaver2017), most studies have focused on relatively long-term employees and have thus established stable relationships. We presume, then, that to clarify the association between attachment and OCB, attachment personality dimensions should be examined over time. Thus, a time lag study design would allow us to overcome the shortcomings of the cross-sectional nature of previous studies. Accordingly, in order to expand on previous investigations, the current study focuses on attachment dimensions during the early stages of personal acquaintance, with OCBs measured after 2 months by an external trainer. Thus, we posited the following hypothesis:
Hypothesis 1
Attachment avoidance and anxiety will predict OCBI and OCBO.
Individual Differences in Caregiving
According to Bowlby (Reference Bowlby1982), the ability to establish secure emotional bonds with others, such as a caregiver and a care seeker, comprises a major indication of an optimally functioning person. Indeed, caregiving dimensions are recognized as a significant construct in the social and clinical literature, serving as a complementary conceptualization to attachment dimensions (Reizer, Ein-Dor, & Shaver, Reference Reizer, Ein-Dor and Shaver2014; Shaver, Mikulincer, Gross, Stern, & Cassidy, Reference Shaver, Mikulincer, Gross, Stern, Cassidy, Cassidy and Shaver2016). However, distinct from attachment, caregiving dimensions have received less research attention in the organizational domain, thus indicating a need to derive insights from research on caregiving in close relationships to the organizational context (Reizer & Hetsroni, Reference Reizer and Hetsroni2015).
Bowlby (Reference Bowlby1982) conceptualized caregiving as an inborn capacity to care for others. This basic ability is activated when another person needs our protection and support (Bowlby, Reference Bowlby1982; Meneghini et al., Reference Meneghini, Romaioli, Nencini, Pagotto, Zermiani, Mikulincer and Shaver2015). Parallel to the two-dimensional attachment model individual differences in caregiving are similarly conceptualized (Meneghini et al., Reference Meneghini, Romaioli, Nencini, Pagotto, Zermiani, Mikulincer and Shaver2015; Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010). The hyperactivated (compulsive) caregiving style reflects an exaggerated appraisal of others' needs, which translates into an intrusive and ineffective tendency to provide help aimed at controlling others. The hyperactivated caregiver expresses positive evaluations of others' needs for help yet tends to underestimate himself or herself as an effective caregiver. For this caregiver type, willingness to help is amplified, with a tendency to be involved in others' problems on a level of enmeshment, well beyond what appears reasonable, but not necessarily successfully, mainly because they express lower levels of perceived self-efficacy in helping situations and express difficulties in self-regulation processes (Meneghini et al., Reference Meneghini, Romaioli, Nencini, Pagotto, Zermiani, Mikulincer and Shaver2015; Shaver et al., Reference Shaver, Mikulincer, Gross, Stern, Cassidy, Cassidy and Shaver2016). Such caregiving efforts are often perceived as controlling and may even be experienced by the needy person as intrusive or aggressive (Meneghini et al., Reference Meneghini, Romaioli, Nencini, Pagotto, Zermiani, Mikulincer and Shaver2015; Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010).
In contrast, a deactivated caregiving style reflects an underestimation of a person's need for care, low empathy and compassion toward others, and tends to entail egoistic and cynical perceptions toward helping others. Deactivation caregiving strategies are characterized by a systematic dismissal or misinterpretation of the information that signals the other's needs, insufficient empathy, emotional distance, and limited involvement in caregiving (Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010). This tendency involves deactivating caregiving in order to avoid the frustration resulting from failing to provide effective care to a needy other. In this case, caregiving efforts are dramatically reduced, and sensitivity and responsiveness to others' needs are inhibited (Meneghini et al., Reference Meneghini, Romaioli, Nencini, Pagotto, Zermiani, Mikulincer and Shaver2015). Individuals low on both deactivation and hyperactivation scales are marked by sensitive, altruistic, and responsive care and tend to be characterized as good caregivers.
The Association Between Caregiving and Ocb
Caregiving has been viewed as the personality dimension representing one's helping and caring perceptions (Reizer & Mikulincer, Reference Reizer and Mikulincer2007; Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010). Therefore, it seems likely that caregiving dimensions would be associated with helping behaviors. Indeed, accumulating evidence suggests that hyperactivated and deactivated caregiving dimensions are associated with helping and support giving in different contexts, such as parent–child relationships (e.g., Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010), couple relationships (e.g., Reizer, Ein-Dor, & Shaver, Reference Reizer, Ein-Dor and Shaver2014), and social contexts (Reizer, Dahan, & Shaver, Reference Reizer, Dahan and Shaver2013; Shaver et al., Reference Shaver, Mikulincer, Gross, Stern, Cassidy, Cassidy and Shaver2016).
Only a few studies have focused on hyperactivation and deactivation caregiving in the organizational domain. Two studies examined volunteering, suggesting that deactivated caregivers were involved in fewer volunteer activities and were detached from social responsibilities (Meneghini et al., Reference Meneghini, Romaioli, Nencini, Pagotto, Zermiani, Mikulincer and Shaver2015; Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010). Conversely, while hyperactivation presented more self-centered motives (Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010), they did not reduce their volunteering activity levels. To the best of our knowledge, only a single study (Reizer & Hetsroni, Reference Reizer and Hetsroni2015) focused on caregiving dimensions and prosocial behaviors in the organizational domain. This cross-sectional investigation indicated that, for call center workers, hyperactivated caregiving was negatively associated with lower levels of self-reported OCBO, while deactivation was associated with both self-reported OCBI and OCBO. For the current study, we expand the knowledge of how caregiving dimensions can contribute to the workplace in several ways. First, we will reexamine the effect of caregiving on OCB by using other-reported measures and a time lag research design rather than self-reported measures and a cross-sectional design that minimizes exposure to common method bias (Podsakoff, MacKenzie, & Podsakoff, Reference Podsakoff, MacKenzie and Podsakoff2012). Second, research has indicated that caregiving dimensions are correlated with attachment dimensions, as both are presumably developed through early interactions in the course of childhood experiences (Reizer, Ein-Dor, & Shaver, Reference Reizer, Ein-Dor and Shaver2014; Shaver et al., Reference Shaver, Mikulincer, Gross, Stern, Cassidy, Cassidy and Shaver2016). Thus, there is a need to explore the unique contribution of caregiving above and beyond attachment dimensions.
Specifically, we hypothesized that both deactivated and hyperactivated caregiving would be negatively associated with OCBs. Based on previous findings, we presume that a deactivated caregiver will try to evade any request for OCBs, such as assisting colleagues at work (OCBI) or ‘going the extra mile' to meet organizational norms and attending meetings (OCBO). Hyperactivated caregiving would be negatively associated with OCBI and OCBO, as these individuals tend to have difficulties in emotion regulation processes and in performing adequate and effective discretionary helping behaviors, irrespective of their motivation to help.
Hypothesis 2
Deactivated and hyperactivated caregiving will be negatively associated with OCBI and OCBO.
The Moderating Role of Group Cohesion
Numerous studies have reported that organizations increasingly rely on teams for carrying out important activities and processes (e.g., Mathieu, Hollenbeck, van Knippenberg, & Ilgen, Reference Mathieu, Hollenbeck, van Knippenberg and Ilgen2017). Group cohesion is defined as team spirit and indicates the level of coordination, cooperation, support, and consensus existing among group members that motivate them to stick together and feel solidarity and unity (Forsyth, Reference Forsyth2018). In a cohesive group, members may experience greater enthusiasm and collective positive mood convergence. In turn, this leads to a greater willingness to act prosocially toward their colleagues and perform OCBs (Chen, Tang, & Wang, Reference Chen, Tang and Wang2009). In contrast, individuals who do not perceive a sense of cohesion with their group (whether due to distrust, dislike, disinterest, or other factors) may fail to benefit from the positive group advantages (Salas, Grossman, Hughes, & Coultas, Reference Salas, Grossman, Hughes and Coultas2015).
Most organizational studies investigating Bowlby's relational dimensions, in general, and attachment, in particular, have focused solely on the individual. Indeed, Yip et al., (Reference Yip, Ehrhardt, Black and Walker2018), in their recent review, concluded that future organizational research should involve the examination of attachment in group settings. Our study sought to accommodate this call and examine the moderating role of the group setting on the impact of both attachment and caregiving dimensions on OCB. The moderating effects of a group-level study would advance our understanding beyond the inconclusive findings which have characterized previous investigations of Bowlby's relational processes at the workplace.
Group cohesion can serve as a moderating process to attachment's relational dimensions. Relationships within groups can fulfill personal relational needs (Mikulincer & Shaver, Reference Mikulincer and Shaver2017; Rom & Mikulincer, Reference Rom and Mikulincer2003). The greater the group's cohesiveness, the more likely its members would feel protected, comforted, supported, and encouraged by the group. As anxiously attached individuals have been shown to experience an extreme need for acceptance (Mikulincer & Shaver, Reference Mikulincer and Shaver2017), they may benefit from group cohesion, as it supplies their basic needs for acceptance, love, and closeness to others (Lavy, Bareli, & Ein-Dor, Reference Lavy, Bareli and Ein-Dor2015; Rom & Mikulincer, Reference Rom and Mikulincer2003).
Only two studies have considered the effects of Bowlby's relational dimensions on task-oriented groups. While the first study (Rom & Mikulincer, Reference Rom and Mikulincer2003) suggested that group cohesion moderated the effects of attachment anxiety over performance among new army recruits, a more recent investigation (Lavy, Bareli, & Ein-Dor, Reference Lavy, Bareli and Ein-Dor2015) indicated that group cohesion moderated the effects of group heterogeneity in attachment and avoidance as well as the final grade of undergraduate students who were engaged in a group academic project. However, these studies did not refer directly to OCBs, nor did they examine the effects of caregiving on prosocial behavior. The current study, then, expands on the cited promising research by suggesting that group cohesion would moderate the contribution of attachment on OCB over a 2-month period.
Specifically, based on previous studies, we assume that group cohesion would moderate attachment anxiety and OCB associations. We posited that attachment anxiety may be positively related to OCBs among cohesive group members. High group cohesion activates a group-specific sense of contextual security, which would inhibit anxious strategies and enable the deployment of psychological resources (Rom & Mikulincer, Reference Rom and Mikulincer2003) that can eventually be used for proactive behaviors such as OCBI and OCBO. Under low levels of group cohesion, the absence of a cohesive group would increase the need to rely upon internal models of security. In these conditions, anxious individuals, who tend to report lower levels of self-efficacy or recognition at work (Hazan & Shaver, Reference Hazan and Shaver1990; Wu & Parker, Reference Wu and Parker2017), may be less willing to go the extra mile to perform OCBs.
Avoidant people tend to be poor group players and collaborators (Rom & Mikulincer, Reference Rom and Mikulincer2003). Indeed, previous work has shown non-significant interactions between avoidance and group cohesion in predicting task performance (Lavy, Bareli, & Ein-Dor, Reference Lavy, Bareli and Ein-Dor2015; Rom & Mikulincer, Reference Rom and Mikulincer2003). Based on these findings, we posit the following hypothesis:
Hypothesis 3
The relationship between attachment anxiety and OCB will be moderated by group cohesion. Specifically, group cohesion will provide a supportive context for a positive association between attachment anxiety and OCB, while lack of group cohesion will increase the negative contribution of attachment anxiety to OCB.
The moderating role of group cohesion on the associations between caregiving and OBCs has yet to be examined, and the contributions of contextual and group dynamics still require investigation. However, based on the relational perspective (Bowlby, Reference Bowlby1982; Mikulincer & Shaver, Reference Mikulincer and Shaver2017), we assume that group cohesion as a source of relational support may also offer hyperactivated caregivers a supportive and encouraging context, providing them with acceptance and assurance (Lavy, Bareli, & Ein-Dor, Reference Lavy, Bareli and Ein-Dor2015; Rom & Mikulincer, Reference Rom and Mikulincer2003). This supportive context would offer several potential benefits, such as satisfying their motivational and self-serving needs, diminishing personal distress regarding their efficacy as a help provider, and perhaps eventually enhancing their inclination to provide help to their colleagues and to express prosocial behaviors. In contrast to hyperactivated caregivers, deactivated caregivers tend to be less responsive in helping situations and are likely to act cynically toward help requests (Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010). Therefore, we assume that supportive group cohesion would not necessarily enhance deactivated caregiver OCB performance. In sum, we posited the following hypothesis:
Hypothesis 4
The relationship between hyperactivated caregiving and OCB is moderated by group cohesion. Specifically, hyperactivated caregiving is associated with OCB under higher levels of group cohesion.
In summary, the current study contributes to the existing literature in several ways. First, by expanding previous work on the associations between attachment styles and OCB (Harms, Bai, & Han, Reference Harms, Bai and Han2016; Little et al., Reference Little, Nelson, Wallace and Johnson2011), we sought to demonstrate that group cohesion moderates the association between attachment and OCB. Second, as there remains a lack of understanding of the effects of caregiving in a workplace setting, we examined the potential contribution of caregiving to OCB variance, beyond what can be explained by attachment. Previous research in this area relied on self-report data and did not control for attachment styles (Reizer & Hetsroni, Reference Reizer and Hetsroni2015). Third, as Reizer & Hetsroni (Reference Reizer and Hetsroni2015) argued, potential moderating mechanisms should be included in examining caregiving and OCB associations. As such, we incorporated group cohesion as a potential moderator in the research model. Finally, the current investigation used a time lag study design and measured the effect of the predictors (attachment and caregiving) on the outcome (OCB) over a 2-month period. We also obtained measures of OCB from training instructors (rather than from self-ratings) in order to decrease common method bias effects (Podsakoff, MacKenzie, & Podsakoff, Reference Podsakoff, MacKenzie and Podsakoff2012). It is noteworthy that in addition to being the first study to examine attachment, caregiving, and OCB in an organizational group setting, our research may help us better understand the unstable findings that have characterized attachment caregiving and OCB in previous research.
Method
Participants and procedure
This research was carried out in a large security organization in Israel. The sample included 147 employees (68 women, 79 men) who participated in a 9-week course for training of professional and technical skills. All participants held non-management positions, and their organizational tenure averaged 15.39 months (sd = 7.12). During the first day of the training, participants were randomly assigned to eight groups, each comprising approximately 15–20 employees. Each group's trainer (i.e., supervisor) accompanied the group for 5 days a week for 9 weeks. Furthermore, the training was conducted in the organization's training dormitory settings. The trainers observed the participants in different situations and engaged in various tasks related to work and non-work behaviors during the entire length of the course. Questionnaires were administered at two different time points. Time 1 was administered at the beginning of the course, with participants completing attachment and caregiving questionnaires. Time 2 was administered 2 months later, at the end of the training course, when group cohesion was assessed. Furthermore, to avoid common-source bias, the trainers assessed OCBs at Time 2.
The collected data allowed us to use a time lag research design in a group context, consistent with the methodology adopted by previous investigators (Lavy, Bareli, & Ein-Dor, Reference Lavy, Bareli and Ein-Dor2015; Rom & Mikulincer, Reference Rom and Mikulincer2003). These studies tracked the effects of attachment on performance behavior during a group training process. We decided to use a similar methodology adopted by the previous group investigators. We monitored several groups of employees who participated in a 2-month intensive course as part of their job requirements. The employees did not know each other and were affiliated with various organizational departments in different regions. The employees completed the personality questionnaires during the course's first few meetings, and the trainers assessed their trainees' OCBs at the conclusion of the course. The ability to track the effects of personality on personal behavior during a relatively short period of acquaintance allowed us to examine attachment, caregiving, and OCB during the formative stages of the relationships.
Measures – time 1
Attachment insecurity
Attachment anxiety and avoidance were assessed with the Experiences in Close Relationships scales (ECR; Brennan, Clark, & Shaver, Reference Brennan, Clark, Shaver, Simpson and Rholes1998). Participants rated the extent to which each item was descriptive of their general experiences in close relationships, on a 7-point Likert-type scale, ranging from 1 (not at all) to 7 (very much). Twelve items assessed attachment anxiety (e.g., I worry about being abandoned), and 12 items assessed avoidance (e.g., I prefer not to show a partner how I feel deep down). Cronbach's alphas were.87 for the attachment anxiety items and .84 for the avoidance items.
Caregiving
Caregiving dimensions were measured with a 20-item self-report instrument designed by Shaver, Mikulincer, & Shemesh-Iron (Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010) to measure caregiving-related deactivation and hyperactivation. Participants were asked to rate the extent to which each item accurately describes their attitudes, feelings, beliefs, and motives in social interactions on a 7-point Likert-type scale, ranging from 1 (not at all) to 7 (very much). Ten items measured hyperactivated (compulsive) caregiving (e.g., Sometimes I feel I force help on another person), with 10 items measuring deactivated caregiving (e.g., Sometimes I feel that helping others is a waste of time). Cronbach's alphas were .78 for hyperactivated caregiving and .80 for deactivated caregiving.
Measures – time 2
Perceived group cohesion
This 10-item scale (Lavy, Bareli, & Ein-Dor, Reference Lavy, Bareli and Ein-Dor2015; Rom & Mikulincer, Reference Rom and Mikulincer2003) assessed the basic definitional components of group cohesion, such as interpersonal attraction and task commitment (e.g., the group members are committed to the group; the group works in a coordinated fashion). Participants were asked to rate to what extent each item characterized their group on a scale of 1 (not at all) to 7 (very much). The scale's Cronbach alpha for the current sample was high (.94).
Organizational citizenship behaviors
The 16-item OCB scale (Lee & Allen, Reference Lee and Allen2002) was used by trainers to rate trainee participants on OCBI and OCBO. Trainers rated the target individual on each item on a 7-point Likert-type scale, ranging from 1 (never) to 7 (very often). Eight items assessed OCBI (e.g., Gives up time to help others who have work or non-work problems), and eight assessed OCBO (e.g., Defends the organization when other employees criticize it). Both the OCBI and OCBO scales achieved very high Cronbach alpha reliabilities for the current sample (.93 and .92, respectively).
Control variables
Our study employed two control variables: organizational tenure (measured in months) and gender (0 = female, 1 = male). It was posited that OCBs would appear stronger in more senior individuals, who, during their service, acquired skills and knowledge of the organizational environment (Ng & Feldman, Reference Ng and Feldman2010). We designated gender as a possible antecedent of OCBs due to social norms emphasizing the performance of communal helping behaviors among women, following Reference Lovell, Kahn, Anton, Davidson, Dowling, Post and MasonLovell et al.'s (1999) suggestion.
Results
Descriptive statistics for study variables are presented in Table 1. Correlations indicated that attachment avoidance was negatively associated with both OCBI and OCBO, whereas attachment anxiety was positively associated with OCBI. Group cohesion was positively associated with OCBO, while deactivation was negatively associated with OCBI. Baron & Kenny (Reference Baron and Kenny1986) suggested that the moderator should be uncorrelated with the predictor to provide a clearly interpretable interaction term. Following their suggestion, our findings indicated that group cohesion did not correlate with relational personality dimensions or with OCBI.
Means, standard deviations, and bivariate correlations
Note: *p < .05, **p < .01, ***p < .001. Results in parentheses are reliabilities. Values on the diagonal represent internal consistency, where applicable. N = 147.
The trainers provided multiple ratings of OCBs, which created the risk of dependency on the responses. To address this possibility, we analyzed the data using hierarchical linear modeling techniques (HLM; Raudenbush, Bryk, Cheong, & Congdon, Reference Raudenbush, Bryk, Cheong and Congdon2004). Our HLM analyses comprised multiple steps, using grand-mean-centered variables. In Step 1, we entered the control variable of organizational tenure and gender. In Step 2, we entered the grand-mean-centered independent variable of caregiving hyperactivation, deactivation, attachment anxiety, attachment avoidance, and perceived group cohesion. In Step 3, we entered the grand-mean-centered interaction of personality dimensions with group cohesion. The HLM analysis, as well as the inclusion of the control variables in the statistical analyses, were conducted appropriately to preclude any confounds and artifacts.
Prior to testing our hypotheses, we ran a null-model equation in HLM to determine the degree of nonindependence in trainer rating of OCBs. This test, equivalent to a one-way analysis of variance (ANOVA) of trainer effects on OCB ratings, was significant: χ2(11, N = 148) = 34.6, p < .001, for OCBI, and χ2(11, N = 148) = 24.4, p < .05, for OCBO. The analysis indicated systematic differences between trainer variance in OCBO and OCBI ratings, respectively. The intraclass correlation coefficients (ICC; Hofmann, Griffin, & Gavin, Reference Hofmann, Griffin, Gavin, Klein and Kozlowski2000) showed that the variance in OCB performance ratings between trainers was .24 for OCBI and .12 for OCBO. This finding supported our decision to use HLM for conducting our analysis. HLM results for the final step are shown in Table 2. Our findings demonstrated that gender was a significant predictor of OCBI. Specifically, results indicated that women tended to perform more OCBIs.
HLM correlations predicting organizational citizenship behaviors
Note. Dummy variable was used for gender (0 = female, 1 = male); SEɤ = standard error of ɤ; *p < .05; **p < .01.
Hypothesis 1, which examined whether attachment predicted OCBI, was supported for attachment avoidance, γ = −.26, p < .05. Results indicated that attachment avoidance also predicted OCBO, γ = −.31, p < .05. The association between attachment anxiety and OCBI was positive, γ = .22, p < .05. Caregiving dimensions were not significant predictors of OCBs, though deactivated caregiving was negatively correlated with OCBI, as presented in the correlation matrix. Thus, this finding provides partial support for Hypothesis 2.
Significant interactions were identified among the variables. Group cohesion moderated the associations between attachment anxiety and OCBI, γ = .24, p < .01. The significant two-way interaction plot is provided in Figure 1. To further probe the interaction, we conducted a simple-slope test. A simple-slope analysis, using Preacher, Curran, and Bauer's (Reference Preacher, Curran and Bauer2006) method indicated that the relationship between attachment anxiety and OCBI was significant for those high in group cohesion, γ = .47, se = .14, p < .01. However, the slope for those low in group cohesion was nonsignificant, γ = −.01, se = .17, p = .90. Thus, the findings provide partial support for Hypothesis 3.
Interaction between attachment anxiety and group cohesion predicting OCBI
Furthermore, group cohesion moderated the associations between hyperactivated caregiving and OCBI, γ = −.25, p < .01. To determine whether the significant interaction supported Hypothesis 3, we plotted a slope at one standard deviation below and above the mean, using Preacher, Curran, and Bauer's (Reference Preacher, Curran and Bauer2006) method. Unexpectedly, a simple-slope analysis showed that the relationship between hyperactivated caregiving and OCBI was stronger for those low in group cohesion, γ = .40, se = .16, p < .05, than for those high in group cohesion, γ = −.10, se = .16, p = .50. This indicates that under low group cohesion levels, OCBI of hyperactivated individuals were more elevated than under high group cohesion levels.
Finally, group cohesion moderated the associations between deactivated caregiving and OCBI, γ = −.26, p < .01. A simple-slope analysis indicated that the relationship between deactivated caregiving and OCBI was positive for those low in group cohesion, γ = .34, se = .15, p < .01. Conversely, this relationship was negative for those high in group cohesion, γ = −.25, se = .12, p < .05, demonstrating that deactivated caregivers perform fewer OCBI in high-cohesive situations. An additional HLM analysis examined the same statistical model without including the control variables, producing similar results.
Discussion
Our research confirmed that individual differences in relational perceptions may be related to OCB, with individual differences addressing the unique moderating effect of group cohesion. In general, the current research findings support the growing tendency in the organizational field to acknowledge the growth-enhancing contribution of Bowlby's relational perspective in the organizational domain (Wu & Parker, Reference Wu and Parker2017; Yip et al., Reference Yip, Ehrhardt, Black and Walker2018).
By expanding on previous work, which raised concerns regarding common method variance (e.g., Reizer & Hetsroni, Reference Reizer and Hetsroni2015; Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010; Syna Desivilya, Sabag, & Ashton, Reference Syna Desivilya, Sabag and Ashton2006), the predictor variables (e.g., attachment and caregiving) and the outcomes (OCBs) were measured at separate time points. In addition, we followed Podsakoff, MacKenzie, & Podsakoff (Reference Podsakoff, MacKenzie and Podsakoff2012) recommendation to obtain measures of predictor and criterion variables from various sources.
The contribution of attachment to OCBs
Previous work has suggested that attachment avoidance can predict OCBIs and OCBOs (Syna Desivilya, Sabag, & Ashton, Reference Syna Desivilya, Sabag and Ashton2006). Expanding on this, we focused on these associations within a prospective design and multiple-source report. As noted, avoidant individuals use work behaviors to eschew social contact and prefer to focus on their own work-task performance (Hazan & Shaver, Reference Hazan and Shaver1990). They tend to rely heavily on themselves and on their personal resources (Ein-Dor et al., Reference Ein-Dor, Coan, Reizer, Gross, Dahan, Wegener, Carel, Cloninger and Zohar2015) and shun extra-role behavior involving interaction with others (Little et al., Reference Little, Nelson, Wallace and Johnson2011). Such an inclination can also impair their proclivity for attending social functions and showing deep concern for the organization. In line with numerous studies relating to prosocial relationships (for a review, see Shaver et al., Reference Shaver, Mikulincer, Gross, Stern, Cassidy, Cassidy and Shaver2016), our results indicate that avoidance is a stable predictor of non-supportive behaviors, whereas the association between anxiety and helping behaviors is less consistent.
Regarding OCBO, our research findings indicated that avoidant individuals tend to impede these extra-role behaviors. As for attachment anxiety, our research results indicate that attachment anxiety is positively associated with prosocial behaviors toward colleagues (OCBI). Several explanations can be offered for the current positive association between anxiety and OCBI. First, our study examined attachment and OCBs among group members, as opposed to previous work that focused on independent work environments (e.g., Richards & Schat, Reference Richards and Schat2011). It was suggested that the anxious individual may be more affected by group dynamics and desires to be included in the group (Rom & Mikulincer, Reference Rom and Mikulincer2003). Therefore, it seems reasonable that anxious individuals will try to provide more apparent help to group members in order to satisfy their personal need for acceptance by the group members. Though this is the first study to examine anxiety and OCBI among group members, previous research has supported these positive associations by suggesting that anxious people tend to help others when they can identify with the help receiver (Kogut & Kogut, Reference Kogut and Kogut2013).
The positive associations between anxiety and OCBI may also be explained by social defense theory (Ein-Dor & Hirschberger, Reference Ein-Dor and Hirschberger2016), which comprises an extension of attachment theory. This theory posits that insecure dimensions have domain-specific advantages that promote the effectiveness of groups. Accordingly, individuals high in attachment anxiety are beneficial to group efficacy because they are likely to demonstrate sensitive behaviors, such as noticing ambiguous signs of threat and warning others about them (Ein-Dor & Hirschberger, Reference Ein-Dor and Hirschberger2016). Our study expands on these findings by claiming that the positive contribution of anxiety to prosocial behaviors in groups can be evaluated not only in life-threatening situations but also during organizational settings such as training.
The moderating role of group cohesion in the association of attachment, caregiving, and OCB
In confirming the third hypothesis, our results demonstrated that group cohesion moderated the associations between anxiety and OCBI, supporting the notion that high group cohesion facilitates a situational-supportive environment. While high-anxious individuals are likely to possess at least a modicum of basic motivation and skills for helping others (Monin, Feeney, & Schulz, Reference Monin, Feeney and Schulz2012), they may lack sufficient psychological resources for actually engaging in helping situations (Collins & Ford, Reference Collins and Ford2010). Thus, group cohesion provides them with a complementary psychological resource. High group cohesion promotes a group-specific sense of attachment security, which, in turn, could inhibit anxious individuals' basic distress strategies. Thus, the deployment of psychological resources to more relational-oriented behaviors is enabled (Rom & Mikulincer, Reference Rom and Mikulincer2003). A situation of high group cohesion could also signal that closeness and trust (Mach, Dolan, & Tzafrir, Reference Mach, Dolan and Tzafrir2010), typically the main objectives of anxious persons, could be attained in the course of group interaction, thereby freeing up resources for the fulfillment of OCB goals.
Although our research findings did not support the moderating role of group cohesion on the association between attachment and OCBO, the theoretical reasoning supports this finding. OCBI is more likely to transpire in social and interpersonal contexts, reflecting a greater interpersonal commitment to other individuals (Bolino & Grant, Reference Bolino and Grant2016) and more affected by the group dynamic context than by the organizational aspect of citizenship behavior.
In addition, our findings suggest that the links between caregiving and OCBs are complex. Consistent with Shaver, Mikulincer, & Shemesh-Iron (Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010), who reported nonsignificant associations between hyperactivation and engagement in volunteer activities, we found no association between hyperactivated caregiving and OCB. In line with previous findings (Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010), we found that deactivated caregiving was negatively correlated with OCBI. Our results did not support a direct linkage between hyperactivated caregiving and OCBI, but rather a moderated association.
Contrary to expectations, the findings indicated that hyperactivated caregiving is more apparent in low cohesive groups. A hyperactivating strategy is intrusive to the prospective target, but it can also include some feelings of commitment and an expression of a need to invest in social relationships (Reizer, Ein-Dor, & Shaver, Reference Reizer, Ein-Dor and Shaver2014). As lower levels of group cohesion are reflective of weak and infrequent social connections (Mathieu et al., Reference Mathieu, Hollenbeck, van Knippenberg and Ilgen2017), we can suggest that these weak social ties facilitate helping behaviors among hyperactivated individuals, such that even their ineffective efforts to provide help are observed, appreciated, and recognized more positively by others in the nonsupportive environment.
As for the deactivated caregiver, we unexpectedly found that deactivation is positively associated with OCBI under lower levels of cohesion and negatively associated with OCBI under higher levels of group cohesion. The deactivated and less sensitive caregivers' reluctance to help others is in part due to their preference for emotional distance (Shaver, Mikulincer, & Shemesh-Iron, Reference Shaver, Mikulincer, Shemesh-Iron, Mikulincer and Shaver2010). Lower levels of group cohesion supported their need for ‘distanced protection’ (George & Solomon, Reference George, Solomon, Cassidy and Shaver2008), thus increasing their willingness to provide help to their colleagues. However, with higher levels of group cohesiveness, deactivated caregivers tend to present lower levels of OCBI. It may be suggested that these close relationships may threaten their need for emotional distance from the needy person, as previously proposed (Collins & Ford, Reference Collins and Ford2010) or consistent with the bystander effect, possibly counting on others to offer help to the needy (Fischer et al., Reference Fischer, Krueger, Greitemeyer, Vogrincic, Kastenmüller, Frey, Heene, Wicher and Kainbacher2011).
Theoretical implications
Our research findings provide further evidence for Bowlby's relational perspective in the workplace. Bowlby (Reference Bowlby1982) proposed that both attachment and caregiving allow us to employ a rich array of strategies to foster our own personal development and interpersonal functioning. This research addresses this basic theoretical assumption by integrating both attachment and caregiving in understanding organizational phenomena. In the same vein, the current results offer some empirical support for the assumption that attachment and caregiving, although potentially influenced by childhood experiences, remain discrete constructs with different outcomes and unique environmental triggers (Reizer, Ein-Dor, & Shaver, Reference Reizer, Ein-Dor and Shaver2014; Reizer & Mikulincer, Reference Reizer and Mikulincer2007). In addition, we showed that group cohesion provides a mechanism of compensation for anxiously attached individuals, contributing to their sense of security.
Conversely, in less cohesive groups, hyperactivated caregivers are more likely to engage in OCBs. These research findings expand on previous work (Lavy, Bareli, & Ein-Dor, Reference Lavy, Bareli and Ein-Dor2015; Rom & Mikulincer, Reference Rom and Mikulincer2003) indicating that security-enhancing groups play a role in shaping effective performance. However, more research is encouraged to investigate the mechanisms of change by which security-enhancing leaders and groups can produce positive changes in the workplace as well as generate group-related variables that can strengthen or inhibit these changes.
The unexpected moderating role of group cohesion in the associations between caregiving and OCBI may imply that trait-incongruent interventions can compensate for low dispositional levels of the caring personality. Along with previous findings (Grant, Reference Grant2009), our results highlight the need for integrative theoretical frameworks to identify the boundary conditions for these patterns. Identifying these conditions will help us understand the circumstances in which we should anticipate OCBs in contexts that are congruent versus incongruent.
Practical empirical implications
Acknowledging the relational perspective of individual differences in the workplace may be of benefit to managers and organizational practitioners. These professionals increasingly recognize that jobs, roles, tasks, and projects are embedded in interpersonal relationships. Internal relationships in the workplace are more pervasive and vital than in the past, and employees carry out their tasks and responsibilities increasingly interdependently (Blustein, Reference Blustein2011). Thus, incorporating the relational perspective in the processes of assessment, training, and team building – along with understanding interpersonal relational needs – may be beneficial. These research findings can also help management in selecting workers and assigning them to work teams.
Future directions
We hope our findings encourage future investigations into the role of attachment and caregiving in a time lag research design. For example, we speculate that hyperactivated and deactivated caregiving can create undesirable outcomes in the long term. Insensitive care may cause recipients to feel incompetent, dependent, and helpless, creating antagonism toward the helper (Banki, Reference Banki2010), though these negative outcomes may not be immediately apparent. In addition, our findings follow previous research suggesting that OCBI and OCBO may have different antecedents (e.g., Ilies et al., Reference Ilies, Fulmer, Spitzmuller and Johnson2009) and different moderating processes (Ozer, Chang, & Schaubroeck, Reference Ozer, Chang and Schaubroeck2014). Thus, we recommend that it may be advantageous to investigate OCBI and OCBO separately in future studies. Finally, our study addresses Reference Organ, Podsakoff, Podsakoff and ZedeckOrgan et al.'s (2011) recommendation to include contextual moderators to understand the weak dispositional effects found in previous work. Supporting their conclusion, we argue that future work can integrate additional moderators, such as job resources (e.g., perceived organizational support and managerial support) or job demands (e.g., work overload or physical demands) as potential moderators of the associations between relational personality dimensions and OCB.
In addition, from a theoretical perspective, a key component of self-esteem lies in the interplay of individuals and their social environment (see Leary, Reference Leary2005). According to sociometer theory (SMT; Leary & Baumeister, Reference Leary and Baumeister2000), self-esteem serves as a sociometer, an internal gauge of others' evaluations of the individual. Due to the importance of social inclusion for survival, humans developed this psychological warning system that monitors and responds to cues relevant to the individual's relational value, such as cues connoting liking and disliking. SMT acknowledges that individuals may possess more than one mechanism monitoring relational value, such that people can gauge their relational value to collectives and even to small groups (Leary, Reference Leary2005). Indeed, a sense of belongingness to groups and work teams is regarded as self-definitional and is of high personal import (Schilpzand & Huang, Reference Schilpzand and Huang2018).
Integrating the social-relational framework, we nuance our predictions by suggesting that the positive effect of anxiety on OCB under high levels of group cohesion (or high levels of liking) may derive from increased levels of self-esteem in the anxious individual. We recommend that future study include a self-esteem measure to identify the meditating processes that facilitate positive effects. We believe that incorporating SMT principles in attachment research can offer a new direction for examining Bowlby's relational dimensions in the organizational domain.
Limitations
The current study is not without limitations. First, in line with previous meta-analyses examining the association between personality and OCB (Ilies et al., Reference Ilies, Fulmer, Spitzmuller and Johnson2009; Organ, Podsakoff, & Podsakoff, Reference Organ, Podsakoff, Podsakoff and Zedeck2011), the correlations in our study were modest. However, it has been suggested that the associations between self-rated personality measures and other-rated OCB may be biased downwards due to raters' limited observational opportunities in contrast to self-reported OCB measures (Ilies et al., Reference Ilies, Fulmer, Spitzmuller and Johnson2009). Furthermore, Organ, Podsakoff, & Podsakoff (Reference Organ, Podsakoff, Podsakoff and Zedeck2011) argued that the associations between personality and OCB tend to be relatively weak in magnitude.
Second, similar to previous work focusing on the moderating role of attachment on task performance among students (Lavy, Bareli, & Ein-Dor, Reference Lavy, Bareli and Ein-Dor2015) or in military recruitment processes (Rom & Mikulincer, Reference Rom and Mikulincer2003), our study was carried out during a training process in an organization, thus offering us the opportunity to examine attachment and caregiving at the beginning of the group formation process. However, the training context limits the generalizability of our findings.
Finally, the current study used a common measure of group cohesion derived from responses collected at the individual level (Salas et al., Reference Salas, Grossman, Hughes and Coultas2015). As our research focuses on the dispositional contribution of individual differences, we did not want to neglect the within-group variability. However, future studies should use more current group cohesion measures, such as micropsychological processes and actual group interactions, which could provide a more substantiated model over time (Srikanth, Harvey, & Peterson, Reference Srikanth, Harvey and Peterson2016).
Conclusion
Bowlby's theoretical perspective is a very popular and influential framework in contemporary personality and social psychology literature (Mikulincer & Shaver, Reference Mikulincer and Shaver2017). Nonetheless, for many years, it has been ‘largely ignored by researchers investigating the role of individual differences in applied workplace settings’ (Harms, Reference Harms2011: 287). The current study contributes to the emerging empirical interest among organizational scholars by including caregiving, a basic relational construct in Blowby's theory, to the study of attachment at the workplace. We examined whether caregiving personality styles might contribute, above and beyond attachment styles, to citizenship behaviors at the workplace. In addition, by addressing the recent call to identify moderating processes (Harms, Reference Harms2011) in general and group variables in particular (Yip et al., Reference Yip, Ehrhardt, Black and Walker2018), we focused on the moderating role of group cohesion as it can satisfy emotional needs. Our findings indicate that the associations between attachment anxiety and OCBI are stronger under higher levels of group cohesion, whereas the associations between hyperactivating caregiving and OCBI are stronger under lower levels of group cohesion. Based on these findings, we recommend that both aspects of attachment and caregiving be investigated in future organizational work because both aspects can provide valuable knowledge concerning the antecedents of employee functioning at the workplace. In addition, our study highlights the importance of broadening the current investigation of attachment and caregiving styles – beyond their interpersonal implications – to workplace domains. Both researchers and practitioners would do well to attend more closely to the relational personality perspectives of caregiving, as well as attachment, in the organizational domain to achieve optimal human functioning at the workplace.
Author ORCIDs
Abira Reizer, 0000-0002-0148-8613.
