Teaming up with temps: the impact of temporary workers on team social networks and effectiveness

ABSTRACT

Temporary workers offer immediate benefits to the bottom line; yet, it is unclear how incorporating temporary workers into teams affects how they function. We apply social identity theory to propose that temporary workers significantly reduce individual- and team-level networks and team effectiveness but that commitment to the leader and intergroup competition can help temporary and permanent employees work together more effectively. Using a sample of employees nested in teams (Study 1, n = 312), we found that status differences affected member interactions resulting in sparser advice and friendship networks for temporary workers compared to their permanent counterparts. At the team level (Study 2, n = 58), these team member differences or contract diversity impacted team functioning through advice networks, such that, teams with greater contract diversity had sparser networks and were less effective. Further, commitment to the leader was found to moderate the negative impact of contract diversity on advice and friendship network density. With the increasing use of temporary worker and the prevalent use of teams, these findings have broader implications for HR functions and present possible avenues to mitigate the negative consequences of temporary workers.

KEYWORDS:

• Temporary employment
• teams
• team effectiveness
• social networks
• contract diversity

Temporary work is increasingly being used to afford organizations greater flexibility and reduce employment costs (Bidwell, Briscoe, Fernandez-Mateo, & Sterling, 2013). Temporary workers are employed at organizations for a particular length of time, typically with short-term contracts, and are hired directly or recruited through an agency, whereas permanent employees are employed indefinitely (Kalleberg, 2000). The staffing flexibility and cost reduction associated with temporary work (e.g., not having to hire, train, and fire) is attractive to organizations, but this perspective focuses on immediate benefits to the bottom line and ignores the potential downside to hiring temporary workers for team functioning.

Numerous studies on contract work find that temporary workers negatively impact permanent employees’ attitudes and behaviours (e.g., Banerjee, Tolbert, & DiCiccio, 2012; Chattopadhyay & George, 2001; Davis-Blake, Broschak, & George, 2003; von Hippel & Kalokerinos, 2012). These findings highlight the negative consequences that temporary workers have on permanent employees, yet in order for teams to be effective, employees must find ways to work together such as share knowledge (e.g., advice networks) and offer support (e.g., friendship networks), which are central to team effectiveness (Kozlowski & Bell, 2003).

To understand how differences among team members in contract type (temporary or permanent) affect teams, we test a model using two studies that examines whether, how, and when temporary and permanent workers can work together effectively. First, we test within-group differences in organizational teams by comparing the social networks (i.e., advice and friendship networks) of team members with different contracts, specifically temporary workers and their permanent counterparts in blended teams and permanent employees in non-blended teams who work only with other permanent employees. Second, we build on the within-group findings by examining how team-level contract diversity influences team effectiveness through social networks and factors in the broader context as moderators that may contribute to better team functioning (e.g., team commitment to the leader, intergroup competition; see Figure 1 for depiction of the proposed model).

Figure 1. Proposed model exploring contract diversity and social networks at the individual and team level.

Thus, our examination of the impact of blended workgroups on team functioning attempts to make several important contributions that are particularly salient given the increase in blended workgroups. First, we contribute to the temporary employment literature by examining (a) whether the inclusion of temporary workers in teams relates to how well members work together, (b) a mechanism through which blended workgroups hinder or enhance team effectiveness, and (c) ways we can capitalize on diversity in employment contracts (e.g., garner commitment). We lend further insight into temporary work research by incorporating structural elements as team processes that relate to contract differences.

Second, we contribute to the team diversity literature through our timely exploration of contract diversity, a relatively unexplored, status-diversity construct. We build on findings that status differences are likely recognized and reinforced in social networks (e.g., Bingham, Oldroyd, Thompson, Bednar, & Bunderson, 2014) by providing an account of how contract diversity can enhance or inhibit team functioning through our proposed model. Further, we contribute methodologically and substantively to findings linking other forms of diversity to social networks (e.g., Klein, Beng-Chong, Saltz, & Mayer, 2004; Reagans, Zuckerman, & McEvily, 2004), which focus on single organizations or use one level of analysis, by using a multi-level approach at the individual (Study 1) and team levels (Study 2).

Finally, with the increasing use of temporary workers and the prevalent use of teams, uncovering interactional differences in employment contracts and their association with team effectiveness is essential for modern organizations and managers interested in leveraging their human capital in the face of contract diversity. By identifying how this previously unexplored factor impacts team effectiveness, our findings have the potential to impact job design, talent management, and training and development, and consequently, have broader implications for organizational performance (Brown & Eisenhardt, 1995).

Theoretical background

The past few decades have seen a large increase of studies examining diversity in work teams and the association between work team diversity and team performance in particular (van Dijk, van Engen, & van Knippenberg, 2012). More recently, status-related processes have been proposed as an important mechanism explaining the effects of work team diversity on team performance (van Dijk & van Engen, 2013). Status is the degree of influence, prominence, and respect that others perceive an individual to have that determines one’s social standing (Ridgeway & Walker, 1995) and is based on perceived access to resources (e.g., knowledge; Bingham et al., 2014). Individuals view status as a sign of competence and as a means to obtain resources (e.g., information) through a better hierarchical position (Huberman, Loch, & Önçüler, 2004). Those lower in status tend to have a higher risk of being harmed, less likely to be facilitated, and attributed with lower levels of warmth (van Dijk, Meyer, van Engen, & Loyd, 2017), which could impact cohesion, cooperation, and eventually performance of the work team.

In this study, status hierarchies are reflected by the ties between organizational members through immediate barriers to team interactions. The level of centrality in a network can be considered as a sociometric measure of status (e.g., Friedkin, 1991); being central in a network implies many sources for knowledge and information, and having many friends adds more status in a network. The determinants of network centrality are consistent with theories of similarity-attraction (Byrne, 1971) and social identity (Tajfel & Turner, 1986). According to similarity-attraction, individuals are attracted to others based on important similar aspects such as attitudes and values. Individuals categorize others based on these perceived similarities and differences, where similar others are categorized as in-group members, while those who are different are considered out-group members. Consistent with social identity theory, in-group members are assigned positive attributes and treated with favouritism, meanwhile outsiders are stigmatized (Tajfel & Turner, 1986) and may become targets of social exclusion (Scott, Restubog, & Zagenczyk, 2013). These theories suggest that status differences based on employment type may prompt high status members to be attracted to one another and form in-groups, while categorizing those with low status as out-group members. Because attraction likely results in increased communication and interaction (Tsui, Egan, & O’Reilly, 1992), similar employees may develop closer ties with each other than with dissimilar others. As such, we propose that differences in status due to employment type may impact the extent to which team members ask each other for advice (advice networks) and see each other as friends (friendship networks; Van Emmerik & Brenninkmeijer, 2009).

Hypotheses

Due to the short-term nature of temporary employment contracts, status differences may be particularly salient in diversity associated with employment contracts. Although some temporary workers perform highly-skilled work, they typically receive less pay compared to equally skilled permanent workers, have a transient employment contract, have been described as being inferior and having a weak work ethic (Boyce, Ryan, Imus, & Morgeson, 2007; Marler, Woodard Barringer, & Milkovich, 2002), which suggests the presence of status differences between temporary and permanent workers (Boyce et al., 2007; Kalleberg, 2000). These observable differences in pay and existing stereotypes are an outward manifestation of underlying status differences where permanent employees have more influence, prominence, and respect in the workplace.

Status and specifically, perceptions of human capital, are directly linked to the employee-organization relationship such that employees who are perceived to be fulfilling relationship obligations (more enduring relationships; invested in organizational goals) acquire more status than employees who are perceived to fulfil transactional obligations (well-defined roles, short-term relationships; Bingham et al., 2014). This disparity in obligations results in perception differences between permanent employees and their temporary counterparts. Permanent employees are typically perceived as more invested, competent, influential, and thus, having greater access to resources and acquiring higher status (e.g., Bingham et al., 2014; Hogg & Terry, 2000) compared to temporary workers who are stereotyped as less skilled, intelligent, and knowledgeable and thus, having fewer resources being given lower status (Parker, 1994).

Perceptions of temporary workers are likely formed early on as employees are typically introduced to others as temporary members causing others to attribute status-related stereotypes solely based on the nature of their employment. Organizations further perpetuate these status differences by investing fewer resources (e.g., benefits; career planning; training) in low status members (e.g., temporary workers) once hired. Providing support for these contentions, these differences in status are arguably the most salient disparity in blended workgroups (Boyce et al., 2007). Although status differences are inherent in other diversity constructs, contract diversity is unique in that (a) other time-related characteristics (e.g., tenure) do not have the same stigma attached (e.g., a less skilled new permanent employee will not experience the same stigma as a new temporary worker) and (b) the temporal nature of employment contracts differentiates it from other status-related diversity constructs (e.g., gender) for which time is not a factor in the formation of status hierarchies in teams.

Type of contract and advice network centrality

Advice networks, which capture the exchange of resources such as information, assistance, and guidance among team members, help members acquire and use relevant knowledge from their teammates to perform tasks (Balkundi & Harrison, 2006). Members’ access to these resources is indicated by their level of centrality within the advice network (Sparrowe, Liden, Wayne, & Kraimer, 2001) as those with a greater number of connections obtain more resources. Drawing on the similarity-attraction (Byrne, 1971) and social identity (Tajfel & Turner, 1986) frameworks, both permanent and temporary team members are more likely to turn to their permanent counterparts, who are perceived to be high-status in-group members with greater access to resources (e.g., information) for advice over their temporary counterparts as they lack influence and are thought to offer little value. Although permanent members could benefit from integrating the knowledge, perspectives, or skills of temporary members (Tempest, 2009), status hierarchies are likely to be reinforced once formed. Even in light of contradictory information (e.g., temporary workers are more skilled), permanent workers may be prone to ignore data that does not adhere to stereotypes or misperceptions. For example, even if organizations employ highly-skilled temporary workers for short-term projects (i.e., functional flexibility), they may still be stereotyped as outsiders who are less skilled and intelligent (Marler et al., 2002) and discounted as a potential source of advice. Instead, members are motivated to protect or gain status/resources within the status hierarchy. In the interest of protecting their status, permanent members may avoid asking temporary members for advice to not risk a drop in status (Agneessens & Wittek, 2012), whereas temporary members have only to benefit from asking their high-status permanent counterparts for advice, and not from other low-status temporary members, as they gain access to resources without their status being threatened (see Figure 2). Findings support that status hierarchies create non-reciprocal relationships where high status individuals or advice givers are less likely to ask for advice from low status individuals or advice seekers (e.g., Agneessens & Wittek, 2012).

Figure 2. Examples of hypothesized advice and friendship networks of blended and nonblended teams. (a) Advice network of blended team, (b) advice network of nonblended team, (c) friendship network of blended team, (d) friendship network of nonblended team.

T: temporary worker; BP: blended permanent worker; NP: nonblended permanent worker; dotted arrow: unidirectional ties; solid arrow: reciprocal ties.

In a related vein, differences in advice networks are likely present between permanent employees in blended and nonblended teams. Though research suggests that high-status (e.g., permanent) group members form coalitions when there are status conflicts (e.g., Wittenbaum & Bowman, 2005), permanent employees in blended groups, who are unlikely to ask advice from lower status members, have fewer team members from whom they can seek advice. Conversely, permanent employees in nonblended groups can defer to all other team members without risking a drop in status. In terms of advice seeking, all team members are likely to seek advice from high-status permanent employees consistently in both blended and nonblended groups, yet blended permanent employees may receive more requests for advice because there are overall fewer permanent members in the group from whom to ask advice. As such, we predict that:

H1:

Temporary workers will receive fewer requests for advice and seek more advice from their team members than permanent employees.

H2:

Permanent employees in blended teams will seek less advice and receive equal amounts of requests for advice from team members than permanent employees in nonblended teams.

Type of contract and friendship network centrality

Categorization based on contract type may also impact friendship networks, which are affect-laden connections that develop over time through shared experiences that are characterized by support, trustworthiness, and affection (Lincoln & Miller, 1979). Because friendship networks are formed based on perceived similarities and shared identities (Mehra, Kilduff, & Brass, 1998), we leverage similarity-attraction and social identity to suggest that perceived differences based on employee contract reinforce status hierarchies and the creation of sub-groups based on perceived similarity to others (i.e., in-group: permanent vs. out-group: temporary). Indeed, scholars note that these status hierarchies and segregation affect primary ties (i.e., friendship networks; Lincoln & Miller, 1979). Employees (e.g., permanent employees) are likely to favour developing closer ties with those perceived similar to them (e.g., other permanent employees) compared to with dissimilar others (e.g., temporary employees). Favouritism also causes the segregation of others, which further reduces opportunities for communication with non-group members and increases in-group ties and a shared identity. Non-reciprocal advice ties (i.e., low-status employees seek advice, high-status employees give advice; Agneessens & Wittek, 2012) also limit the amount of interaction occurring between temporary and permanent workers, which further exacerbates group divisions.

Like advice ties, studies show that friendship is, in almost half of the cases, non-reciprocal (e.g., Almaatouq, Radaelli, Pentland, & Shmueli, 2016; see unidirectional and reciprocal ties in Figure 2). People are generally poor at perceiving the reciprocal nature of friendships, as individuals may have an egocentric view that puts them more in the centre of friendship networks compared to reality (Kumbasar, Rommey, & Batchelder, 1994). This implies that the in-degree and out-degree friendship network centrality do not necessarily have to be equal; for example, a team member may consider all of her fellow team members as friends, but only one of the fellow team members may consider her as a friend in return.

Temporary members as low-status out-group members are likely affected the most. As proposed above, unlike advice networks, temporary workers will seek friendships with similar others with whom they share experiences (i.e., other temporary workers). Though they may desire developing additional friendship ties, we expect that they will seek them with individuals outside of the workgroup, especially when they are the only temporary group member, rather than with permanent workers, with whom a friendship may seem one-sided, lacking trust, not feasible, or not supportive. Permanent members may visibly express favouritism for other permanent members or not make any effort to communicate/connect with temporary workers, causing feelings of isolation and alienation.

Likewise, rather than reaching out to temporary members, permanent employees will focus on building relationships with other permanent members, with whom they feel they have more in common. As they share similar status, the relationships that develop can prove beneficial over time (e.g., increased social support). The finite nature of their contracts may make it seem like a fruitless endeavour to invest friendship resources in temporary workers. Thus, permanent workers expect to gain little by forming friendship ties with low status individuals and may even fear a drop in status or threat to in-group friendships. Though both blended and nonblended permanent employees will seek friendship from other permanent members, blended permanent employees will have fewer members from which to seek friendships and receive fewer requests due to having fewer permanent members in their group and temporary members who seek friendship elsewhere. In line with this theoretical rationale, we expect that:

H3:

Temporary workers will receive fewer and seek fewer friendships from their team members than permanent employees.

H4:

Permanent employees in blended teams will seek fewer friendships and receive fewer friendship requests from team members than permanent employees in nonblended teams.

Contract diversity and team functioning

We now build on the within-group findings to better understand how contract diversity impacts team functioning. Scholars have examined the impact of group differences on team performance and the results largely remain equivocal. Some studies suggest that team diversity enhances performance, whereas other studies find a negative relationship between team diversity and performance. Varied findings have been attributed to the “oversimplification of team diversity – an inherently complex construct” (Bell, Villado, Lukasik, Belau, & Briggs, 2011, p. 710) and an insufficient consideration of mechanisms or moderators linking diversity to team performance (van Dijk et al., 2012; van Knippenberg & Schippers, 2007). The previous perspective focusing on directly linking diversity and performance through status hierarchies or social network structures diverges from the current thinking in team research, which depicts teams as complex adaptive systems and focuses on broadening our understanding of process variables (e.g., input-mediator-output-input (IMOI); Ilgen, Hollenbeck, Johnson, & Jundt, 2005). The IMOI framework suggests that team processes are members’ “interdependent acts that convert inputs to outcomes through cognitive, verbal, and behavioural activities directed toward organizing taskwork to achieve collective goals” (Marks, Mathieu, & Zaccaro, 2001, p. 357). Although team processes have typically included content and temporal mechanisms, they have largely ignored structural configurations (e.g., status hierarchies, social networks; Crawford & Lepine, 2013; van Dijk & van Engen, 2013).

Few studies focusing specifically on social networks suggest that they mediate the relationship between demographic diversity constructs (e.g., education, gender) and team outcomes (e.g., Curşeu, Raab, Han, & Loenen, 2012; Reagans et al., 2004). Yet, other studies fail to find support for the link between certain diversity constructs and social networks. For example, Henttonen, Janhonen, Johanson, and Puumalainen (2010) found that education and age diversity did not impact team social networks; only gender diversity was related to social networks. Another study found that deep level differences (e.g., values similarity) were more instrumental in affecting social networks than surface level differences, as demographic similarity did not predict friendship networks (Klein et al., 2004).

Individual-level predictions that interactions among team members differ in blended and nonblended work teams suggest that contract diversity will impact the density of team social networks. Weak relationship ties among group members are characterized by a sparse network density (Sparrowe et al., 2001), which can restrict the exchange of knowledge and support, and the flow for different ideas and shared experiences (Henttonen et al., 2010). Because permanent employees in blended teams will be less likely to go to temporary workers for friendship and advice, blended teams will have networks that are less dense compared to nonblended teams.

As resources are exchanged in teams, members gain a better understanding of their colleagues and develop comradeship (Klein et al., 2004). An exchange of resources also leads to the sharing of knowledge and without this team interaction, members do not work as a unified team, resulting in reduced productivity and difficulties achieving team goals. The interaction that occurs may result in increased relationship conflicts among team members (e.g., von Hippel & Kalokerinos, 2012), which can inhibit performance. Teams that therefore fully utilize team members by exchanging resources (i.e., seeking their advice and friendship), or have denser social networks, are more likely to experience higher performance. Indeed, research indicates that team advice and friendship network density impact team performance (e.g., Henttonen et al., 2010). Moreover, Grund (2012) found that teams with only a few highly central members show lower team performance compared to teams with a more evenly distributed social network. As such, we expect that contract diversity affects advice and friendship network density, and as a result, team effectiveness through the structural mechanisms of social networks.

H5:

Contract diversity negatively affects team effectiveness through its negative impact on advice and friendship network density.

The moderating roles of commitment to the leader and intergroup competition

Though we contend that contract diversity is negatively related to team effectiveness, certain factors may mitigate its negative effect. We leverage social identity to focus specifically on two factors, commitment to leader and intergroup competition, as possible boundary conditions of the contract diversity – social networks – team effectiveness relationship. These variables were selected because they are most proximal to facilitating a sense of belonging to a higher-order group, and support the formation of a superordinate identity in teams (van Dijk et al., 2017). In doing so, we present a model that positions these factors as moderators of the mediated relationship between contract diversity and team effectiveness through social networks.

We proposed earlier that in blended teams, temporary and permanent members have a different status, driving members to limit or increase their advice and friendship relations with members of the other status group. However, research has also shown that collective team behaviour such as communication and cooperation is facilitated to the extent that members consider themselves in terms of higher-order social categories (i.e., as members of a common in-group; see Haslam, van Knippenberg, Platow, & Ellemers, 2003). As such, the extent to which teams develop dense advice and friendship networks is also dependent on the extent to which they develop a superordinate identity or a sense of belonging to a higher-order group (Kane, Argote, & Levine, 2005). There is an abundance of evidence that a superordinate identity decreases in-group favouritism and increases cooperativeness (see Gaertner et al., 2000 for a review). For example, Kane (2010) found that a superordinate identity facilitated the exchange of information between groups.

Drawing from social identity theory, one way to facilitate a superordinate identity is through leaders who garner commitment from their followers. Effective leaders use social attraction and charisma to acquire status to support them with the capacity to actively gain compliance with their requests (Hogg, 2001). If the team is committed to the leader, it is likely to collectively comply with the leader’s requests, resulting in improved cooperation, despite subgroup status differences. The leader encourages team members, albeit temporary or permanent, to work together or strive for a shared goal, which can facilitate a sense of belonging to a common in-group and increase the total number of advice/friendship ties within a team. Both temporary and permanent members are therefore more likely to identify as a common in-group when they are committed to their leader, which weakens the negative impact of contract differences on the ties between group members. Furthermore, being collectively committed to the leader can be considered a superordinate identity itself (Hornsey & Hogg, 2000). For example, Sluss and colleagues (2012) show that newcomers’ connection to an individual leader translates into a connection with the group he or she leads. Commitment to the leader, therefore, may reduce status differences between temporary and permanent members because all members are viewed as being a part of a common in-group, in turn facilitating cooperation among both temporary and permanent group members, which consequently increases their ties.

H6:

Commitment to the leader will moderate the mediated relationship of contract diversity on team effectiveness through advice/friendship network density such that the higher the commitment, the weaker the negative effect of contract diversity on team effectiveness through advice/friendship network density.

Another determinant of a superordinate identity that is consistent with social identity theory is intergroup competition, which refers to a “social situation in which the goals of different teams are linked in such a way that goal achievement by any one team reduces the ability of other teams to reach their respective goals” (Baer, Leenders, Oldham, & Vadera, 2010, p. 823). Intergroup competition prompts teams to consider the environment as threatening and stressful (van Oostrum & Rabbie, 1995). The resulting out-group threat and intergroup comparisons lead to more homogeneous identifications with the in-group, encouraging blended group members to categorize both permanent and temporary members within their teams as in-group members, and members from other teams (whether they are permanent or temporary) as out-group members. Intergroup competition can result in improved communication and cooperation (Haslam et al., 2003) through the formation of network ties within teams. What this means is that although group members may be temporary, they may be considered in-group members and be more of equal contributors in exchanging advice and friendship. Intergroup competition thus reduces status differences by encouraging groups to develop a superordinate identity, which increases advice and friendship ties.

H7:

Intergroup competition will moderate the mediated relationship of contract diversity on team effectiveness through advice/friendship network density such that the higher intergroup competition, the weaker the negative effect of contract diversity on team effectiveness through advice/friendship network density.

Overview of studies

Our hypotheses were tested using two studies. The individual-level hypotheses (H1–H4) were tested using a sample of temporary and permanent employees nested in organizational teams (Study 1). As such, in the aim of Study 1 was to provide a better understanding of the configuration of the social networks in the team. Then, in Study 2, we tested the team-level hypotheses (H5–H7) with a sub-sample of the teams used in Study 1, combining this team member data with data collected at the leader-level.

Study 1: individual-level relationships

Participants and procedure

We first tested our individual-level hypotheses using a field study conducted in three countries: the Netherlands, Romania, and Bulgaria. The second author was familiar with these countries, which allowed us to sample organizations in a broader geographical area. We collected data across 12 organizations in sectors that traditionally employ temporary workers and organize their processes around teams; the final sample included manufacturing companies, a hotel, call centres, a university, and a municipality. Teams were included in the sample if they met the criteria (e.g., two or more individuals, shared common goals, interact socially, exhibit task interdependencies) defined by Kozlowski and Bell (2003). A team of research assistants was responsible for the distribution and collection of the questionnaires. Team members completed the questionnaire during a break and the team leader completed the questionnaire in the office.

The final sample included employees of manufacturing companies (production sector, n = 124), a hotel and call centres (service sector, n = 71), and a university, a ministry, and a municipality (public sector, n = 117). The teams included in this study were all operational-level teams with low functional diversity including production teams, desk teams, and functional departments. By analysing the covariance matrices, we determined that it was appropriate to combine the data, as they were similar across the industries.

In total, we received team member (n = 312) data from 70 teams, ranging in size from 3 to 12 (average = 4.45). A slight majority (52%) of team members were male with an average age of 42.56 years old. About 53% of the team members had at least a college degree. The average within-team response rate was 91%. The dataset included 39 temporary workers (both fixed-term and temporary agency workers) and 273 permanent employees, which may seem disproportionate, but it is reflective of typical team composition, and therefore generalizable to how organizations structure blended workgroups. The average tenure was 15.67 months (SD = 14.26) for temporary workers and 81.20 months (SD = 92.46) for permanent employees (t = −4.41, p < .001). Moreover, both temporary and permanent workers had an equal average educational level (t = 1.18(307), p = n.s.). On average, temporary workers were younger (mean age = 36.67) compared to permanent workers (mean age = 43.42, t = −3.44(307), p < .01).

Measures

Type of contract

We measured type of contract at the individual level using a dichotomous variable (i.e., do you have a permanent or temporary contract with this organization?). Permanent employees were further categorized as blended (i.e., work in teams with temporary workers; n = 73) and nonblended (i.e., work in teams with only permanent employees; n = 200).

Advice/friendship network centrality

Individual social networks were measured using the roster method (Scott, 2000). For each team, we listed the team members in alphabetical order and asked each team member if they go to that member for work-related advice, and if they consider that member a friend (no = 0, yes = 1). For each team, a matrix was constructed, resulting in 70 friendship- and advice-matrices. We computed in-degree and out-degree centrality scores for each individual using UCINET (Borgatti, Everett, & Freeman, 2002), normed within each network (Borgatti et al., 2002) to allow for comparisons across groups of different sizes. In line with other studies on ego-network centrality (e.g., Lee, Qureshi, Konrad, & Bhardwaj, 2014; Sparrowe et al., 2001), out-degree friendship (and advice) network centrality was assessed by counting the number of fellow team members that were nominated as a friend by a focal team member. For in-degree friendship (and advice) network centrality, we counted the number of fellow team members that nominated the focal person as a friend.

Control variables

We controlled for country (reference = the Netherlands) and sector (reference = service sector) to consider the heterogeneity of the sample. We controlled for age, gender, educational level, and organizational tenure, but also for similarity in age, gender, education, and organizational tenure because similarity in demographic and work-related characteristics could be related to the position of team members in social networks (Klein et al., 2004). Additionally, age and education similarity can be used to control for the motives for hiring temporary workers because they could indicate if a temporary worker was hired for providing specialized skills or instead only as additional support. We calculated Euclidean distances to measure the respondents’ demographic similarity to their teammates.

Analyses

Considering the nested structure of the data (individuals nested in teams), we tested our first four hypotheses using multilevel analyses. We used the mixed models-procedure in SPSS version 22 (IBM Corp., Released 2013) because this allows for the use of post-hoc analyses in multilevel models. Post-hoc analyses allows for the analysis of our categorical independent variable (contract type). In our analyses, we used random intercepts and slopes to test the relationship between type of team member and the network centrality measures (both Level 1 variables) within the teams (Level 2).¹ However, because the slope did not show variance in most of the models, we chose the most parsimonious model and used random intercepts and fixed slopes to test our first set of hypotheses. To assess the differences between contract type on the network centrality measures, we used a Bonferroni post-hoc analysis.

Results

Tables 1 and 2 show the descriptive statistics and results of the multilevel analyses (top panel) and Bonferroni post-hoc analyses (bottom panel).²

Table 1. Means, descriptive statistics, and correlations of main variables at the individual level for study 2 (N = 312).

	Variable	Mean	SD	1	2	3	4	5	6	7	8	9	10
1	Type of team member^a	.87	.33
2	Advice in-degree centrality	59.39	32.65	.24**
3	Advice out-degree centrality	59.85	34.89	−.05	.03
4	Friendship in-degree centrality	47.58	34.45	.25**	.38**	.05
5	Friendship out-degree centrality	48.50	40.39	.10	.03	.24**	.42**
6	Age	42.56	11.64	.19**	.03	−.05	−.09	−.19**
7	Age similarity	4.88	3.11	−.08	.00	.01	−.17**	−.16**	.13*
8	Gender^b	.48	.50	−.01	−.07	.03	.18**	.22**	−.02	−.16**
9	Gender similarity	.17	.19	.12*	−.08	−.05	−.21**	−.16**	−.04	.03	.08
10	Educational level	4.46	.98	.07	.03	.02	.05	.04	−.18**	−.04	.16**	.10
11	Educational level similarity	.30	.33	−.17**	.03	.03	.04	.03	−.06	−.02	−.03	.18**	−.29**
12	Tenure	72.85	89.22	.25**	.03	−.12*	.06	−.08	.43**	.03	−.03	−.08	−.05
13	Tenure similarity	32.08	34.72	.01	−.07	−.01	−.13*	−.10	.33**	.26**	−.03	.06	−.02
14	Romania^c	.23	.42	.00	.12*	.11*	.26**	.22**	−.18**	−.06	.20**	−.17**	.10
15	Bulgaria^c	.27	.44	.23**	−.06	−.06	.29**	.26**	.05	−.20**	.34**	.03	.16**
16	Production sector^d	.40	.49	−.11	−.02	.01	−.04	−.04	−.11*	.02	−.25**	−.21**	−.25**
17	Public sector^d	.20	.40	.05	−.08	−.09	.12*	.10	.02	−.16**	.15**	−.12*	.05
	Variable	11	12	13	14	15	16
1	Type of contract^a
2	Advice in-degree centrality
3	Advice out-degree centrality
4	Friendship in-degree centrality
5	Friendship out-degree centrality
6	Age
7	Age similarity
8	Gender^b
9	Gender similarity
10	Educational level
11	Educational level similarity
12	Tenure	−.10
13	Tenure similarity	−.01	.62**
14	Romania^c	.15**	−.06	.10
15	Bulgaria^c	−.24**	.19**	.04	−.33
16	Production sector^d	.07	−.13*	−.15**	−.04	−.30**
17	Public sector^d	−.14*	.10	−.01	−.06	.32**	−.40**

*p < .05; **p < .01.

^a0 = Temporary.

^b0 = Male.

^cReference = the Netherlands.

^dReference = service sector.

Table 2. Results of multilevel analyses (Study 1, N = 312).

	Advice network centrality (in-degree)		Advice network centrality (out-degree)		Friendship network centrality (in-degree)		Friendship network centrality (out-degree)
	Within	Between	Within	Between	Within	Between	Within	Between
	ß (SE)	ß (SE)	ß (SE)	ß (SE)	ß (SE)	ß (SE)	ß (SE)	ß (SE)
a. Regressions
Age	−.11 (.17)	.05 (.30)	−.19 (.18)	−.17 (.35)	−.05 (.13)	−.39 (.41)	−.44 (.20)	−.55 (.40)
Age similarity	−.72 (.69)	.55 (1.83)	−73 (.68)	1.84 (1.86)	.00 (.51)	−.68 (1.67)	−.38 (.57)	−.26 (1.65)
Gender^a	−12.96 (4.86)**	4.57 (10.87)	4.35 (6.96)	−1.43 (10.69)	−2.89 (4.29)	4.31 (13.29)	5.13 (6.11)	2.98 (13.37)
Gender similarity	−16.77 (17.76)	−.80 (14.82)	12.19 (21.76)	−12.11 (14.89)	−20.41 (17.43)	−23.50 (19.51)	−14.51 (13.79)	−31.17 (19.60)
Romania dummy^b		4.93 (7.04)		5.33 (7.31)		27.74 (9.58)**		25.57 (9.37)**
Bulgaria dummy^b		−11.84 (7.03)		−6.74 (7.00)		27.31 (9.73)**		28.38 (9.67)**
Production sector^c		−2.50 (5.67)		−.68 (5.58)		8.27 (8.02)		6.41 (8.06)
Public sector^c		−4.65 (4.81)		−4.72 (4.93)		4.73 (6.81)		3.06 (7.15)
Type team member	22.14 (7.59)**	12.72 (4.15)**	−16.53 (7.83)*	14.06 (4.67)**	8.54 (4.34)*	13.15 (5.33)*	−8.47 (6.56)	14.27 (5.23)**
b. Means
Temporary worker		33.53		61.58		30.44		42.27
Blended permanent		53.38		40.98		39.55		34.28
Nonblended permanent		66.10		66.11		52.89		54.29

*p < .05; **p < .01.

^a0 = male.

^bReference = the Netherlands.

Reference = service sector. The top panel presents the unstandardized estimates of the random coefficient models. The bottom panel shows the results of the post-hoc-analyses comparing the three types of team members.

For our first hypothesis, we predicted that temporary workers receive fewer requests for advice (in-degree) and seek more advice (out-degree) from their team members compared to blended and nonblended permanent employees. We find that temporary members have a lower advice network in-degree centrality compared to blended permanents (MD = 19.85, p < .01) and nonblended permanents (MD = 32.57, p < .01), and a higher advice network out-degree centrality compared to blended permanents (MD = 20.60, p < .01), but not compared to nonblended permanents (MD = 4.53, p = n.s.), providing partial support for H1. Hypothesis 2 proposed that blended permanents seek less advice (out-degree) and receive equal amounts of requests for advice (in-degree) from team members compared to nonblended permanents. Our results show that blended permanents have a lower out-degree advice network centrality than nonblended permanents (MD = 25.13, p < .01), and there is no difference between blended and nonblended permanent members receiving requests for advice (MD = 12.72, p = n.s.), providing support for H2. Moreover, blended permanents have a lower out-degree advice network centrality than nonblended permanents (MD = 25.13, p < .01), and there is no difference between blended and nonblended permanent members receiving requests for advice (MD = 12.72, p = n.s.), providing further support for H2.

Our third hypothesis suggested that temporary workers receive fewer (in-degree) and seek fewer (out-degree) friendships from their team members than permanent employees. The results for Hypothesis 3 show a non-significant difference in in-degree friendship network centrality between temporary and blended permanent members (MD = 9.11, p = n.s.), but that as expected, temporary members have a lower in-degree friendship network centrality compared to nonblended permanents (MD = 22.45, p < .05). Our findings show no significant difference in out-degree friendship network centrality between temporary and blended (MD = 7.99, p = n.s.) and nonblended permanent members (MD = 12.02, p = n.s.). In conclusion, we find partial support for H3. Finally, Hypothesis 4 predicted that blended permanents seek fewer friendships (out-degree) and receive fewer friendship requests (in-degree) from team members than nonblended permanent members. The results show that blended permanent employees have a lower out-degree friendship network centrality (MD = 20.01, p < .05) than nonblended permanent employees but we find no difference in in-degree friendship network centrality (MD = 13.34, p = n.s.) between permanent employees, which offers weak support for H4.

To examine advice and friendship ties in more detail, we calculated the extent to which employees (temporary or permanent) adhere to their own social category for advice and friendship. Considering the under-representation of temporary workers, we needed an indicator of inter-group ties that took into account that permanent employees have more similar team members compared to temporary workers. We therefore used the adjusted homophily index (see Gower & Legendre, 1986), which ranges from −1 (indicative of extreme “heterophily”) to +1 (indicative of extreme homophily). We find a significant difference (F = 10.66, p < .01) with respect to advice ties between temporary (=−.08) and permanent employees (=.23), but no significant difference in friendship ties (F = .37, p = n.s) between temporary and permanent workers.

Discussion of study 1

The results are generally in the support for the within-group hypotheses. Contract type has significant implications for social networks such that when compared to permanent employees, temporary workers have sparser social networks. At the same time, permanent members ask for less advice and support when they are teamed up with temporary team members, providing them with fewer resources compared to nonblended permanent members. In Study 2, we build on the within-group findings to examine how contract diversity impacts social network density and effectiveness across teams.

Study 2: team-level relationships

Participants and procedure

The sample for the team-level analyses included a subset of our sample used in Study 1. The sample included employees from the production sector (n = 106), service sector (n = 71), and public sector (n = 63) with a total of 58 teams (n = 240 team members and n = 58 leaders), ranging in size from 3 to 9 (average = 4.22). The covariance matrices were similar across the industries, which suggested that it was appropriate to combine the data. The average within-team response rate was 95.5%. The average age was 42 years old and a slight majority (51%) of team members were female. The sample included 20 teams with temporary workers and 38 teams without temporary workers. The average proportion of temporary workers in blended teams was 32.66% (SD = 15.18) and there were no teams with only temporary workers. The average team tenure was 15.71 months (SD = 15.12) for temporary workers and 88.47 months (SD = 96.69) for permanent employees (t = −3.98, p < .001). Each team leader was a permanent employee and was responsible for one team. A slight majority (55%) of team leaders were female with an average age of 43 years old. About 73% of the team leaders had a college degree and their average tenure with the organization was 161.02 months (SD = 131.90).

Measures

Contract diversity

We measured contract differences using Blau’s diversity index, which is defined as one minus the sum of the squared proportions of each category, and is a common measure for calculating the level of diversity for categorical variables (Harrison & Klein, 2007).

Advice/friendship network density

Social networks were measured using network density by dividing the total number of advice/friendship ties within a team by the total number of possible ties within a team, which is consistent with other studies (e.g., Sparrowe et al., 2001),

Commitment to the leader

To measure commitment to the leader, adapted items of the Organizational Commitment Questionnaire (Mowday, Steers, & Porter, 1979) were used with the team leader as a focus of commitment. The five items were measured on a 5-point scale from “totally disagree” to “totally agree” (e.g., “I feel loyalty to my team leader”). In order to find out whether data aggregation from the individual to team level was feasible, Rwg_(j)s and ICCs were calculated (Rwg_(j) was .83 (median = .91), ICC(1) was .16, and ICC(2) was .47).

Intergroup competition

Team members were asked to complete an intergroup competition scale that included items of the Intergroup Effectiveness Scale developed by Richter, Scully, and West (2005). The scale had four items and was answered using a 5-point Likert scale ranging from “not at all” to “a very great extent” (e.g., “To what extent do you feel the relationship between your team and the other teams is competitive?”). For intergroup competition, Rwg_(j) was .80 (median = .82, ICC(1) was .39, and ICC(2) was .77).

Team effectiveness

Using a scale from Rousseau and Aube (2010), team leaders judged team effectiveness on the basis of five items: “This team achieved its goals”; “This team is very productive”; “The quality of the work performed by this team is very good”, “This team always respects the deadlines for our goals”, and “This team has a high respect for costs”. The team leaders had to rate the performance on these five criteria using a 5-point scale ranging from 1 (not at all) to 5 (to very large extent).

Control variables

We controlled for age, gender, education, and organizational tenure diversity, as both demographic and work-related diversity (e.g., Reagans et al., 2004) relate to network density and team outcomes. We used Blau’s index to calculate gender and educational level density. To measure age and team tenure diversity, we used the variance in each team.

Analyses

Hypotheses 5–7 were tested using a macro by Hayes (2013) to test for the significance of the indirect effect of contract diversity on team effectiveness through network density and the conditional indirect effects depending on different levels of commitment to the leader and intergroup competition. We used bootstrapping to calculate the 95% confidence intervals (5000 resamples) to assess the significance.

Results

Table 3 shows the descriptive statistics and the correlations of the main variables at the team level. The results of the OLS regressions are shown in Table 4. Because team size, educational level diversity, and tenure diversity did not correlate with any of the main variables, we excluded them from the final analyses.³

Table 3. Means, descriptive statistics, and correlations of main variables at the team level (Study 2, N = 58).

		Mean	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1	Contract diversity	.14	.20
2	Advice density	.56	.22	−.42**
3	Friendship density	.53	.31	−.62**	.47**
4	Team effectiveness	4.04	.56	−.30*	.36**	.25*	(.80)
5	Commitment to the leader	3.67	.46	.21	.24	−.09	.42**	(.84)
6	Intergroup competition	2.34	.66	−.12	.22	.22	−.06	−.21	(.73)
7	Age diversity	76	71	.33*	−.06	−.24	.01	.36**	−.04
8	Gender diversity	.17	.21	.12	−.16	−.20	.00	.29*	−.08	−.03
9	Educational level diversity	.31	.20	.09	.04	.01	−.04	.01	.32*	−.14	.06
10	Tenure diversity	482	851	.04	−.04	−.07	−.01	.26*	.19	.20	.06	.08
11	Team size	4.22	1.44	−.11	−.05	.02	−.23	−.13	.05	.01	.05	−.10	.31*
12	Romania^a	.31	.46	.32*	−.01	−.15	.09	.14	.34**	.13	−.01	.33*	.10	−.26*
13	Bulgaria^a	.36	.48	−.53**	.31*	.48**	.14	.14	.25	−.13	.11	−.05	.13	.21	−.49**
14	Production sector^b	.41	.50	.12	.07	−.14	−.34**	−.14	−.19	.10	−.23	−.10	−.21	.07	−.55**	.03
15	Public sector^b	.25	.44	−.03	−.08	.05	.28*	.01	−.22	−.10	.01	−.01	−.04	−.01	.21	−.19	−.48**

*p < .05; **p < .01.

^aReference = the Netherlands; ^breference = service sector. Cronbach’s alpha scores are reported in the diagonal.

Table 4. OLS regressions with unstandardized beta coefficients (Study 2, N = 58).

	Advice network density		Friendship network density		Team effectiveness
Control variables
Age diversity	.00 (.00)	.00 (.00)	.00 (.00)	.00 (.00)	.00 (.00)
Gender diversity	−.11 (.13)	−.12 (.12)	−.28 (.17)	−.27 (.18)	−.22 (.34)
Romania^a	.24 (.08)**	.06 (.09)	.08 (.11)	.02 (.14)	−.25 (.23)
Bulgaria^a	.13 (.07)	−.00 (.08)	.21 (.09)*	.17 (.13)	.01 (.19)
Production sector^b	.13 (.08)	.16 (.07*	−.03 (.11)	−.03 (.12)	−.55 (.22)*
Public sector^b	−.01 (.07)	−.01 (.07)	.03 (.09)	.04 (.10)	.12 (.17)
Independent variable
Contract diversity	−.52 (.16)**	−.59 (.16)**	−.66 (.22)**	−.67 (.24)**	−.35 (.47)***
Mediators
Advice network density					.88 (.38)*
Friendship network density					−.08 (.28)
Moderators
Commitment to the leader		.23 (.08)**		.03 (.12)
Intergroup competition		.09 (.05)***		.04 (.08)
Contract diversity × Commitment to the leader		.88 (.34)*		.22 (.53)
Contract diversity × Intergroup competition		.00 (.21)		.01 (.32)
Total R^2	.36	.50	.46	.46	.38
F	3.90**	4.13**	6.01**	3.58**	3.16**

***p < .10; *p < .05; **p < .01.

^aReference = the Netherlands.

^bReference = service sector.

Hypothesis 5 suggested that advice and friendship networks mediate the relationship between contract diversity and team effectiveness. The results show that contract diversity is negatively related to friendship network density (B = −.66 (.22), p < .01) and advice network density (B = −.52 (.16), p < .01), but only advice network density is significantly related to team effectiveness (B = .88 (.38), p < .05) and the bootstrapped indirect effect is significant (estimate = −.45(SE = .27), p < .05 [CI 95%_low = −1.15: CI 95%_high = −.07]), which partially supports H5.⁴

Our sixth and seventh hypotheses predicted that the effect of contract diversity on team effectiveness through advice and friendship network density would be moderated by commitment to the leader (H6) and intergroup competition (H7). For commitment to the leader, we find a non-significant interaction for friendship network density and a significant interaction for advice network density (B = .88 (.34), p < .05). The bootstrapped indirect effect of contract diversity on team effectiveness through advice network density for low commitment to the leader is negative (estimate = −.85(SE = .48), p < .05 [CI 95%_low = −1.93: CI 95%_high = −.02]), but for high commitment to the leader this indirect effect becomes non-significant (estimate = −.18(SE = .24), p = n.s. [CI 95%_low = −.94: CI 95%_high = .08]), which partially supports H6 (see Figure 3 for the interaction effect). Cross-level analyses further show that the differences between blended permanent members and temporary members with respect to their advice in-degree (estimate = 16.52(SE = 10.10), p = n.s. [CI 95%_low = −3.58: CI 95%_high = .36.64]) and out-degree centrality (estimate = 19.01(SE = 10.47), p = n.s. [CI 95%_low = −39.86: CI 95%_high = 1.83]) disappear under conditions of high commitment to the leader, supporting our claim that cooperation improves when they are collectively committed to the leader. For intergroup competition, the interaction terms of intergroup competition for network density are not significant, which rejects H7.

Figure 3. Leader commitment interaction between the contract type – advice network density relationship.

Discussion of study 2

The results at the team level suggest that advice network density mediates the relationship between contract diversity and team effectiveness. Although contract diversity is negatively related to friendship network density, it does not have a mediating role that impacts team effectiveness. Commitment to the leader helps to mitigate the negative impact of contract diversity on advice network density. Intergroup competition positively relates to advice network density, but the interaction was not significant. Overall, our hypotheses were generally supported with regard to advice network density but not friendship network density.

General discussion

Prior research has emphasized the negative impact that temporary workers can have on the attitudes and behaviours of permanent employees, but blended team members must interact with one another for a team to function effectively, which begs the question as to which mechanisms may enable temporary and permanent workers to thrive together in teams. Investigating these relationships at the individual and team levels with organizations in various industries enables us to examine the extent to which blended workgroups impact team effectiveness through social networks across two studies. Our findings suggest that temporary workers have sparser social networks compared to permanent employees with between-group differences emerging for permanent employees who work with temporary workers having sparser advice networks than their nonblended permanent counterparts. Finally, the results reveal that contract diversity has a negative impact on team effectiveness through advice networks at the team level, but that commitment to the leader helps to neutralize this negative impact. Taken together, the results advance new directions for theory and practice.

Theoretical implications

Foremost, the findings provide a valuable extension to research on the impact of temporary workers, which highlights the negative consequences of temporary workers on permanent employees’ attitudes and behaviours (e.g., Chattopadhyay & George, 2001; Davis-Blake et al., 2003), but has largely ignored how temporal issues affect team processes, which are paramount to team effectiveness. Our findings address this concern for more research on the impact of temporary workers on their permanent counterparts (e.g., Banerjee et al., 2012; Gallagher, 2005) and lagging research on studying temporary workers in the context of teams.

The current investigation shows that blended work groups have sparser social networks, which is evident by temporary workers and blended permanent employees asking for less advice and having fewer friendship ties. Temporary members are more likely to go to permanent members for advice and support, whereas permanent members prefer to exchange more resources with fellow team members with a similar work status, rather than temporary members of a lower status. Our results fit with studies that show that status influences network ties because high-status individuals are placed in more central positions (e.g., Lincoln & Miller, 1979) and that status uniquely relates to how the workplace operates (e.g., Boyce et al., 2007). We suggest that contract differences hamper the facilitation of social networks due to segregation based on status positions (e.g., training not offered to temporary workers).

Though our understanding of how temporary workers impact teams seems critical for organizations interested in truly capitalizing on a blended workforce, the research to date provides little insight, as most has focused on the individual-level impact of temporary workers. This oversight underscores the need for scholarship that expands the focus from the individual to the broader context. Our incorporation of individual- and team-level phenomena contributes to our theoretical understanding of the impact of temporary work on group dynamics. The findings reveal that advice networks are an explanatory factor linking contract diversity and team effectiveness. Rather than strengthening social networks by giving teams a richer pool of information and resources, the use of blended teams weakens social networks. Our focus on how temporary workers affect social networks responds to calls for conceptualizing structural elements (e.g., social networks) as team processes (e.g., Crawford & Lepine, 2013) and the impact of team structure on team performance (Carboni & Ehrlich, 2013). The results imply that temporary workers have detrimental consequences for team effectiveness, which is an important aspect to contextualizing our research (Johns, 2006) because it shows how the context in which workers are embedded (e.g., blended teams) influences their behaviour.

While we found that temporary workers have detrimental consequences for teams, leaders who garner commitment from team members help to mitigate this negative effect, which helps address calls for research on how leaders impact diverse work teams (Kearney & Gebert, 2009). This finding of commitment to the leader is important for the equivocal findings on the impact of group differences on team performance largely due to the insufficient consideration of salient moderators (van Knippenberg & Schippers, 2007). Namely, differences in employment contracts can have positive or negative implications for social networks depending on the level of commitment to the leader, which is consistent with research on the critical role of leaders in reducing status differences in diverse teams (Mitchell et al., 2015). While our findings elucidate the importance of commitment to the leader, the non-significant finding of intergroup competition suggests that outgroup threats are not enough to overcome within-group status differences and increase the homogeneity of group identification, suggesting that contextual factors operate differently in mitigating status differences.

We draw on status-related diversity research to explore contract diversity through the lens of existing diversity frameworks and theories, but we advocate that the impact of contract diversity may not simply be extrapolated from previously studied diversity constructs. The effect of contract diversity on teams may be more complex due to the temporal nature of employment contracts and nuanced differences between workers (e.g., temporary, blended, and nonblended). As the workforce continues to be more diverse in terms of employment types, these findings suggest that diversity research should include contract diversity as a predictor that affects team processes and outcomes through within-group status configurations. We help provide a more comprehensive understanding of the effect of diversity on team outcomes by using a multi-level approach across various organizations to explain how diversity in employment contracts affects networking behaviour. We positioned contract diversity as a salient construct that affected team functioning, which answers calls for research on how status-related differences influence team dynamics (Ravlin & Thomas, 2005; van Dijk & van Engen, 2013). Our findings emphasize the importance of social categorization based on status that is consistent with social identity (Tajfel & Turner, 1986). We extend previous studies (e.g., Klein et al., 2004; Reagans et al., 2004) by examining blended workgroups across multiple organizations in different sectors, which enhances our understanding of the role of status-related diversity.

Practical implications

Our findings offer important implications for HR professionals, managers, and organizations that utilize teams and temporary workers. First, our findings also have implications for employee training and development. Once temporary workers are hired, it is important for them to be integrated into teams to build trust through opportunities for interaction (e.g., team building), which has been shown to predict knowledge sharing (Connelly & Kelloway, 2003). Our findings show that blended permanent employees are less likely to ask their temporary counterparts for advice, which may be an oversight because although they may be stereotyped as lacking skills based on their work status, temporary workers may bring fresh approaches from previous work settings (van der Vegt, Bunderson, & Kuipers, 2010). When introducing temporary members of the team, leaders should make explicit how the temporary team member can help the team, and offer the temporary team member time to introduce himself/herself socially. Moreover, when forming teams, HR professionals and managers can include diversity training, which includes perspective taking to facilitate the development of advice networks and increase team effectiveness. Developing team members to be more aware and empathic, value other perspectives, share information, and assist others with tasks can influence their perceived evaluation of others and affect status differences related to employment contracts. Second, as commitment to the leader was found to mitigate the negative consequences associated with blending workgroups, HR professionals can select and develop managers who are able to garner commitment from their subordinates and foster a climate of commitment through effective management practices. This is especially helpful for managers who have limited ability to influence team composition.

Limitations and future research

Our results should be considered in light of several potential limitations. First, our research design was cross-sectional, but the potential for common method bias was attenuated because we used multi-source measures. The team diversity literature suggests considering the longitudinal impact of various types of diversity on team outcomes, although it is difficult to examine these relationships longitudinally because of the short-term nature of temporary work. Future studies should endeavour to explicitly test these constructs (e.g., superordinate identity) using a longitudinal design. For example, questions remain as to how long it takes for a superordinate identity to develop in blended workgroups. Future research should also focus on other mechanisms that foster superordinate identities. Finally, although examining structural elements and conceptualizing social networks as team processes was a contribution of our study, it does not delve into underlying processes such as communication, and as such we encourage future researchers to examine how contract diversity impacts other team processes such as conflict, information sharing, and mental models.

Future research is also needed to examine the extent that teams with sparser social networks seek advice and friendship from those outside their team. It may be that support outside of the team (e.g., high status expert) or other factors further mitigate some of the negative effects of sparse within-team social networks. It is also plausible that permanent and temporary workers consider each other only as acquaintances but not close friends, and as such, our measure of social networks does not capture the strength of ties, or how often a respondent goes to that member for advice or comradeship (e.g., number of interactions). The measure also does not capture the quality of the social relationships. We also encourage research on other types of social networks (e.g., hindrance ties) that could impact team effectiveness. For example, temporary workers could make it difficult for permanent group members to carry out their job tasks if they are responsible for their training.

Finally, our study only focused on temporary workers but research shows that contingent workers are not a homogenous group (e.g., Wilkin, 2013). Future research is needed to replicate our findings for other types of contingent workers (e.g., independent contractors who may be treated like “regular” employees; George & Chattopadhyay, 2005), as well as different experiences with and motives for accepting temporary employment. Future studies should also compare the impact of teams with differing levels of temporary worker diversity (see Marler et al., 2002) and different histories and skill levels of temporary employment on team outcomes. The impact of temporary contracts on social networks and performance might also depend on other moderators such as task complexity, where the impact may be more detrimental for complex tasks. Further work is needed to examine other salient job attitudes (e.g., perceived organizational support) and behaviours (e.g., organizational citizenship behaviours). Moreover, although our sample included organizations that operated in various regulatory frameworks (e.g., advanced to less regulation of temporary work), we encourage future researchers to replicate these findings in countries that limit temporary employment to shorter durations (e.g., three months), which could affect the extent to which temporary workers are integrated in social networks.

Conclusion

This study enriches our understanding of why contract diversity affects team outcomes through a multi-level structural understanding of the implications of blended teams. We build upon temporary work and team diversity research to answer calls for research by offering insight as to the underlying mechanisms of team effectiveness. The results imply that temporary workers can have detrimental consequences for how teams perform, but that leader commitment can help to mitigate this effect, and offer a better understanding of how to increase team effectiveness.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. We did not add a Level 3 (organization) to our analysis because the number of organizations (n = 13) did not meet the threshold of 30–50 units required for including an additional level in the multilevel analysis (Maas & Hox, 2005). Therefore, we included country and sector as control variables.

2. Because educational level and tenure diversity did not consistently correlate with the main outcome variables, we excluded them from the final analyses. The results were also analysed with all of the control variables and remained consistent.

3. The results were also analysed with all of the control variables and remained consistent.

4. To test the robustness of these findings, we also analysed Hypothesis 5 with (a) teams with only one temporary employee, (b) teams with more than three members, and (c) teams with less than six members. Our findings are consistent across these samples.