Introduction

Corporate social responsibility (CSR)—also known as corporate citizenship, responsible business, corporate responsibility, and social performance—is a form of corporate self-regulation integrated into business and organizational models (Wood 1991). CSR represents inter-organizational activities and status which are related to the organization’s societal and stakeholder obligations (Luo and Bhattacharya 2006; McWilliams and Siegel 2010). A growing interest in CSR is apparent for both practice and research, with specific focus on the way organizations incorporate social demands into their operations and also on anticipated benefits for the organizations (Maignan and Ferrell 2001). Particularly, CSR is developing rapidly across a variety of popular initiatives, such as providing employees’ with education, promoting training in ethical programs, adopting environmental-friendly policies and even sponsoring community events (Maignan and Ferrell 2000). Benefits for business organizations which stem from CSR are the ability to charge a premium price for their product, gaining a good business image, attracting investments, enhancing employees’ job satisfaction, and overall improvement of business performance (e.g., Carmeli et al. 2007; Maignan and Ferrell 2001). Current literature, though, does not provide plausible explication to clarify how such self-regulated CSR (i.e., social performance) affects employees’ performance (Pedersen and Neergaard 2008), a research gap that we aim to bridge in this study.

A potential explanation for the relationship between CSR and team performance can be provided by self-evaluations in which team performance is boosted with, strengthened team efficacy, team self-esteem (Gardner and Pierce 1998; Judge and Bono 2001; Tyran and Gibson 2008), team identification (Somech et al. 2009), and self-guidance training (Brown 2003). Self-evaluation is the overall perception of employees about themselves (Judge and Bono 2001). Motivation, behavior, and performance in teamwork settings are substantially influenced by the variety of self-evaluation components of the members of the team, such as team efficacy (e.g., Tyran and Gibson 2008), team self-esteem (e.g., Katz-Navon and Erez 2005), emotional self-awareness (Jordan 2001), self-estimation (Olszewska 1982), etc. Particularly, there is consistent evidence, suggesting that general self-efficacy and self-esteem strongly influence the way employees act and react in various settings (Gist et al. 1989; Judge et al. 1997; Williams and Williams 2010). It should be noted that while self-efficacy is an individual-level phenomenon, team efficacy (i.e., team self-efficacy) exists as a group-level attribute (Bandura 1997; Feltz and Lirgg 1998). Team efficacy is a judgment about the extent to which a team can successfully perform its work tasks, and previous research suggests that team efficacy significantly predicts team performance (Porter 2005).

Conventional wisdom has indicated that people with a strong sense of self (e.g., team efficacy or self-esteem) are more highly motivated, they are higher achievers, and they are more resilient in the face of adversity than those who have a weak self-concept (Gardner and Pierce1998). Whereas some research examines self-esteem and team efficacy (or self-efficacy) as antecedents that affect employee’s performance (Chen et al. 2004), others examine self-esteem and team efficacy as outcomes of organizational care, training, education, and ethics—all being considered forms of CSR (Kaler 2000; McAllister and Bigley 2002; Muafi and Gusaptono 2010; Yadav and Iqbal 2009). Put together the above mentioned studies point out the mediating mechanisms of team efficacy and self-esteem between CSR and team performance.

Our study differs from previous study by contributing to current knowledge in two important ways. First, previous study tends to integrate team self-efficacy and self-esteem into a single construct to predict work outcomes (e.g., job satisfaction and the management of organizational change) (e.g., Judge et al. 1998, 1999). In contrast, we explicate the theoretical distinctions between team self-efficacy and team self-esteem by simultaneously testing their mediating mechanisms. Team self-efficacy and self-esteem are two distinct constructs. Team self-efficacy is a malleable, task-specific belief (Bandura 1997), while self-esteem is a personality quality, which brings a trait-like, generalized competence belief (Chen et al. 2000, 2001, 2004). Specifically, self-esteem is defined as people’s belief regarding how well they are living up to the standards of value prescribed by the worldview (Harmon-Jones et al. 1997). Team self-efficacy, however, captures enduring differences of people in the tendency to predict if they are as capable of meeting specific task demands. A person can have a low level of self-esteem, yet he or she may have a high self-efficacy of performing a certain chore. We would like to stress that failure to distinguish between self-efficacy and self-esteem would reduce precision and validity, and deter from understanding the determinants of performance (Chen et al. 2004).

Second, we move beyond earlier study (e.g., Brammer et al. 2007) to evaluate three relevant dimensions of CSR at a team-level in regard to their influence on performance. It is important because some research has failed to take the multi-dimensional nature of CSR into account from a teaming perspective (e.g., De los Salmones et al. 2005). Predominantly, most CSR research is focused on either an organization-level (e.g., Longest and Lin 2004; Maxfield 2008; Shen and Chang 2009) or an individual-level (e.g., Lin et al. 2010; Maignan et al. 1999). Here we attempt to study the team-level. It is important to conduct team-level analyses for CSR issues since the perceptions of team members about ethical issues (e.g., CSR) have a significant impact on the team’s subsequent responses (Peek et al. 2007) and decision making (Hunt and Jennings 1997). The core values of CSR can determine employees’ preferences on what they consider to be good or bad and form the foundation for moral principles that would then translate into the team’s or the organization’s ethical system (Elm 2003; Fry 2004), substantially influencing the performance of their team or organization. Responsible business practices such as CSR are of high importance for issues like staffing, team building, and general improved motivation (Castka et al. 2004). At the same time, self-esteem has so far been mostly examined from organization-based angles rather than team-based perspectives (e.g., Forsyth et al. 2007; Sekiguchi et al. 2008). Our study complements previous research by evaluating the effect of CSR on team performance via the mediation of team efficacy and team self-esteem at the team-level analyses.

Theoretical Background and Hypotheses Development

In CSR, it is important to search mediating processes to better explain why certain inputs affect team performance. Our search for mediators has been well informed by previous attention to the link between team inputs and outputs (Ilgen et al. 2005). An important theoretical perspective of team performance is the input–process–output theory (I–P–O theory) (Littlepage et al. 1995). Classic research by Steiner (1972), McGrath (1984), and Hackman (1987) suggests models in which inputs lead to processes that eventually result in team performance (Ilgen et al. 2005). A range of models exists, suggesting that input and process variables have a key influence on outcomes such as team performance (Gladstein 1984; Hackman 1987; Hackman and Morris 1975; Littlepage et al. 1995). Unfortunately, research about team performance has rarely tried to examine the relationship among input, process, and outcome measures from a CSR perspective. For that reason, we explore the input–process–outcome relationships that eventually lead to team performance. Specifically, we examine the relationships linking CSR input variables (e.g., economic citizenship) and team processes (e.g., team efficacy) to an outcome (i.e., team performance). Figure 1 shows the conceptual model of this study.

Fig. 1
figure 1

Conceptual model

Full size image

Hypotheses Development

The evolution of the concept and definition of CSR have started in the early 1950s (Carroll 1999), though the construct has been used since interchangeably with other labels, such as corporate citizenship (Carroll 1998), sustainable development, or accountability (Henderson 2007; Kovacs 2006). It is considered to be a multi-faceted construct (Brammer et al. 2007; Carroll 1998), and it is widely taught in management and business context (Matten and Moon 2005).

A number of studies were conducted regarding the association between CSR and organizational performance, which was found to be positive but moderate (Orlitzky et al. 2003). The development in our understanding of CSR as a multi-faceted construct and the need to check for mediation relationships rather than simplistic direct impact have lead us to develop our theoretical model to include both elements, that is, exploring the various dimensions of CSR as antecedents to performance, and evaluate the possible mediation of major constructs in the management and behavioral sciences—efficacy and esteem, at the team-level, a subject fairly neglected in the literature (cf. Sonnentag and Volmer 2009). It should be noted that all these antecedents and mediators are perceptional. Earlier scholarly work has emphasized the need to test performance via objective measures rather than relying on perceptions, though perception measures for performance can be strongly associated with external measurements (Delaney and Huselid 1996; Huselid 1995).

Team Efficacy and Self-Esteem

Team efficacy is a team’s self-confidence (or belief) in its capability to successfully accomplish specific teamwork (Bandura 1997; Kreitner and Kinicki 2007). Team efficacy originates in individual team members, and through team processes of social interaction and mutual task experience the members’ self-efficacy jointly converges into a team-level factor (Kozlowski and Klein 2000). Team efficacy is based on an aggregated concept of self-efficacy from the individual level to the group level (Tyran and Gibson 2008; Katz-Navon and Erez 2005). Principally, team efficacy influences what team members choose to do, how much effort they invest in aiming to reach the team’s objectives, and their persistence when initial team efforts fail to obtain good performance (Bandura 1997). High team efficacy can lead to successful teamwork performance in a variety of achievement-related situations (Gardner and Pierce 1998; Schenkel and Garrison 2009). A strong link exists between team efficacy and performance (Judge and Bono 2001; Knight et al. 2001). Strong team efficacy helps to improve team performance following failures, and high team efficacy results in improved team performance (Bandura 2000; Feltz and Lirgg 1998; Gibson 1999; Seijts et al. 2000; Spink 1990).

A construct which is related to team efficacy but is still distinct from it is team self-esteem. It reflects general employees’ evaluation of their adequacy and/or worthiness as team members (McAllister and Bigley 2002). It is a collective opinion or appreciation of their own team (Swogger 1993). Self-esteem has been conceptualized hierarchically, possessing different levels of specificity and focus such as organization-based, team-based, and task-specific levels (Gardner and Pierce 1998). Positive self-perception is considered an important building block in human resource-based competitive strategies (Lawler 1992; McAllister and Bigley 2002; Pfeffer 1998). Above all organization- or team-based self-esteem has been linked to critical outcomes, such as job satisfaction, organizational commitment, organizational citizenship, and job performance (McAllister and Bigley 2002; Pepi et al. 2006). For example, an investigation into the relationship between self-esteem and performance at the individual level shows that those who have low self-esteem were rated as less competent, and were less successful in term of performance (Doherty 1980). Hence, team efforts to maintain positive team self-esteem need to be encouraged by management for facilitating work performance (Swogger 1993). Team members who possess high levels of team self-esteem have confidence in the ability of their team and become motivated, capable and empowered (Gardner and Pierce 1998), leading to increased task performance of their team. Collectively, the hypotheses can be stated as below:

H1

Team efficacy is positively related to team performance.

H2

Team self-esteem is positively related to team performance.

CSR, Team Efficacy, and Team Self-Esteem

CSR consists of four dimensions which are refined from previous literature focusing on employees as the stakeholders: (1) economic citizenship, referring to the firm’s obligation to bring utilitarian benefits to employees, such as quality working environment, training, and education (e.g., Maignan and Ferrell 2000; Zahra and LaTour 1987); (2) legal citizenship, referring to the firm’s obligation to fulfill its business mission within the framework of legal requirements; (3) ethical citizenship, referring to the firm’s obligation to abide by moral rules which define proper behavior in society; and (4) discretionary citizenship, referring to the firm’s obligation to engage in activities that are not mandated, not required by law, and are not expected of businesses in an ethical sense (Maignan and Ferrell 2000).

Unlike the first three, the fourth dimension of discretionary citizenship is not closely related to employees’ benefits (e.g., economic reciprocation) or to their job contents (e.g., business ethical or legal practices). We are not aware of previous research that has directly linked discretionary citizenship with job performance. Hence, we employed this dimension as a control variable rather than an antecedent of team efficacy and of team self-esteem. The justification regarding the other three dimensions of CSR is provided in detail as follows:

The necessary and primary social responsibility of any business organization is economic in nature, as an organization is a basic economic unit in our society (Carroll 1979). In addition, the organization has the role of taking care of its employees, due to their position as major stakeholders (Maxfield 2008; Turker 2009). As such, it has a responsibility to provide quality working conditions, training, education, and career development (i.e., forms of economic CSR/corporate citizenship) while producing goods (or services) and selling them at a profit (Weyzig 2009). Developing employees’ skills and training generates a positive influence on their self-efficacy and self-esteem (Baruch and Peiperl 2000; Muafi and Gusaptono 2010), suggesting a positive influence of economic citizenship on self-efficacy and self-esteem.

Current theory developments indicate that targeted training and education in workplaces (i.e., a form of economic citizenship to employees) play an important role in developing employees’ self-efficacy (Wilson et al. 2007). For example, corporate economic citizenship has positive effect on entrepreneurship self-efficacy through educational programs provided to employees (Chowdhury and Endres 2005; Cox et al. 2002). The typical teaching methods in most educational programs provided by business organizations (i.e., economic citizenship), which include guest speakers and case studies, can strengthen learners’ self-efficacy through the use of role models (Wilson et al. 2007). Meanwhile, in studying workplace antecedents of self-esteem, Tharenou (1979) indicates that determinants of self-esteem comprise facets in which the work is performed (e.g., safe and quality working environment), or extrinsic job characteristics, including the economic context (e.g., the development of skills and careers) (Bunker 1991; Ferris et al. 2005).

Creed et al. (1996) observe the importance of career development or job training programs (i.e., economic citizenship) in affecting self-esteem and self-efficacy. It is thus further concluded that employees with low self-efficacy should be given priority access to scarce behavioral-modeling training resources (Eden and Aviram 1993). Trainings, education or career development practices increase team members’ self-efficacy (e.g., Muafi and Gusaptono 2010). Certain training methods can boost self-efficacy in the self-management areas (Frayne and Lathan 1987), cognitive model (Gist 1989) and behavior model (Gist et al. 1989). Personal development courses improve self-esteem significantly (Creed et al. 2001; Yadav and Iqbal 2009). Based on the above rationale, we hypothesized:

H3

Economic citizenship is positively related to team efficacy.

H4

Economic citizenship is positively related to team self-esteem.

Legal issues often shape and influence people’s social identity (Bergami and Bagozzi 2000; Eskridge 2001). Once employees become aware that behavior performed by their corporation is illegal (moral awareness—see Butterfield et al. 2000), their degree of liking for their own team would be undermined (i.e., low team self-esteem). At the organizational level, company’s insincere attitude to the society can be intolerable by its employees (Shi and Cui 2010). Eventually, employees can lose confidence in their organization and in their team if such illegal behavior is consistent (i.e., reduced team efficacy).

Societies expect businesses to fulfill their mission within a framework of legal requirements (Carroll, 1979). Employees perceive themselves as legally and fairly treated when they believe that they have received, or will receive what they are entitled to or deserve according to contracts or to the law (e.g., Frazier et al. 2010; Tyler 1989). Organizational engagement with illegal behavior and law-breaking would damage team members’ self-efficacy and self-esteem whereas a law-abiding workplace facilitates work teams’ efficacy and their self-esteem. There is a link between an employee’s perception of the legal system and their self-efficacy, leading to seek help from others (Jeannin 2009). Complementary studies from both cross-sectional field studies (Wiesenfeld et al. 2000) and controlled experiments (Koper et al. 1993) indicate that fair perceptions which are based on law and company regulations can be related to organization-based self-esteem and context-specific self-esteem (McAllister and Bigley 2002). For example, in a case of a serious discrimination in a workplace, those falling victim to the transgression become discouraged, and as a result their self-efficacy and their self-esteem decrease. Consequently, we present the hypotheses below:

H5

Legal citizenship is positively related to team efficacy.

H6

Legal citizenship is positively related to team self-esteem.

Ethical CSR represents behaviors and activities that are not necessarily codified into law, but nevertheless are expected by society’s members and by employees (Carroll 1979). An important way for linking ethical CSR and self-evaluation in terms of self-worth as team members is through inferring from how fairly or ethically their organization treats them (Li and Cropanzano 2009; McAllister and Bigley 2002; Rosenberg 1979). This basic idea of “reflected appraisal” is fundamental to sociological theories of the self (Gecas 1982), and it has been extended to describe more fully how members’ perceptions of being treated by the firm influence their understandings of who they are in their work unit or team (e.g., Brockner 1988; McAllister and Bigley 2002; Tyler et al. 1999). That is, increased perceived existence of ethics by employees provides a basis for a positive self image—leading to an increased self-esteem and self-efficacy. An ethically oriented business strategy is generating profits through upgrading employees’ self-efficacy in order to create competitive competence, not by harming competitors, but by building up resources through one’s own well doing (Barney 1996; Park 1998). Therefore, it implies that ethical CSR is positively related to self-efficacy of team members.

Employees’ perceptions about their firm’s ethics and social responsiveness play a significant role in motivating employees to obtain strong team self-esteem. Employees’ ethical experiences of their working life play a vital role in personal and psychological health, in particular in the creation of self-esteem (Collier and Esteban 2007). By identifying with a successful ethical organization, members may enhance their self-esteem also by comparing their organization to those of lesser ethical quality (Peterson 2004; Smith et al. 2001). When organizations are honest in their relationships with employees (De los Salmones et al. 2005), the latter are likely to perceive being trusted and valued by the organization. Collectively, the hypotheses about ethical citizenship can be summarized as below:

H7

Ethical citizenship is positively related to team efficacy.

H8

Ethical citizenship is positively related to team-based self-esteem.

If, as expected, the three components or dimensions of CSR are positively influencing team efficacy and team self-esteem, and if both team efficacy and team self-esteem influence future team performance, then, inevitably, team efficacy and team self-esteem mediate the relationship between CSR and team performance at the. Positive attitudes and perceptions are anticipated to influence actual performance at the team-level (Lin 2010b). As Rynes et al. (2005) and others have suggested, testing mediation mechanisms is critical and essential if we are to understand the duality of formation processes and practical implementations of new theories. Thus, we propose the following hypothesis to encompass the mediation relationships we anticipate:

H9

Team efficacy and team self-esteem mediate the relationship between CSR and performance.

Methods

Subjects and Procedures

The research hypotheses described above were empirically tested using a survey of professionals in work teams from IT firms in Taiwan. Teamwork is an essential element in IT firms (Lin 2010b). We invited IT professionals in Taiwan, employed in 20 large IT firms within the high-tech sector, as defined by Baruch (1997), to provide an appropriate representative sample of organizational teams. These were existing working team, and the members were well acquainted with each other. We surveyed five members from each team, including managers and their subordinates. In cases where a manager was supervising more than one team, we only surveyed one of his or her teams to avoid any confusion to the manager. Since members in our sample teams work in close geographical proximity, they often naturally team up to work together during their long working hours (i.e., at least 8 h a day).

Of the 1,000 questionnaires distributed to the members of 200 teams, 801 usable questionnaires from 172 teams were returned for a questionnaire response rate of 80.1%, of them 473 were male (59.10%) and 687 have a bachelor’s degree or higher (85.77%). Our sample contained a total of 744 participants have been working in their current team for a year or more (92.88%), and it also contained 701 employees (87.52%) with Internet experience of a year or more to do their job. These characteristics (e.g., education, job experience, and internet experience) show that the sample firms were appropriate representatives for the population of knowledge workers in IT industries. A satisfactory response rate of our survey is mainly due to the strong support of our sample firms in which their managers of personnel departments helped distribute the questionnaires to both team leaders and members who express their voluntariness. Besides, this study surveyed respondents without obtaining their names to reduce their suspicion or hesitation in factually filling out our survey questionnaires. Specifically, respondents were assured of complete anonymity in the cover letter (e.g., Baruch and Holtom 2008) confirming that neither their personal names nor the names of their teams would be disclosed.

Measures

The constructs in this study are measured using 5-point Likert scales drawn and modified from previous literature (Janssen and Van Yperen 2004; Lin 2010a; Mosley et al. 2008; Pierce et al. 1989; Turker 2009). We first conducted a pilot study with a sample of 57 working professionals and analyzed the data using an exploratory factor analysis with the promax oblique. Then, we re-worded items with low factor loadings in our pilot test. Last, we employed the back-translation procedure suggested by Brislin (1970) and by Reynolds et al. (1993), using focus groups to insure a match between the original wordings and their translation. The pilot study’s data were excluded from the subsequent actual survey.

Team performance was measured using four items modified from Janssen and Van Yperen (2004). Two sample items included “I always complete the duties specified in my job description,” and “I meet all the formal performance requirements of the job.” Team efficacy was measured using five items modified from Mosley et al. (2008). Two sample items included “I am confident in meeting the quality demands of the job,” and “I am confident in correcting the mistakes in my work.” Team self-esteem was measured using six items modified from Pierce et al. (1989). Two sample items were “I count around here,” and “I am taken seriously.”

Corporate social responsibility was measured by four dimensions: economic, legal, ethical and discretionary citizenship (note that discretionary citizenship was used as a control variable). These four dimensions were measured using twenty items (five items for each dimension) modified from Lin (2010a) and Turker (2009). For example, two sample items for measuring perceived economic citizenship included “My firm supports employees who want to acquire additional education (or skills),” and “My firm has flexible company policies that enable employees to better balance work and personal life.” Two sample items for measuring perceived legal citizenship included “The managers of my firm comply with the law,” and “My firm follows the law to prevent discrimination in workplaces.” Two sample items for measuring perceived ethical citizenship included “My firm has a comprehensive code of conduct in ethics,” and “Fairness towards co-workers and business partners is an integral part of the employee evaluation process in my firm.” Two sample items for measuring perceived discretionary citizenship included “My firm gives adequate contributions to charities,” and “My firm sponsors partnerships with local schools or institutions.”

Data Analysis

To test the mediation effects of team efficacy and team self-esteem between CSR and team performance, we conducted an analysis with two steps. While the survey data was analyzed at the first step using confirmatory factor analysis (CFA) on all data collected to assess scale reliability and validity (i.e., measurement modeling), the second step applied path analysis (PA) based on the team-level data in which individual responses were aggregated. We employed PA rather than structural equation modeling (SEM) because SEM requires more than 200 samples to obtain stable test results (Hatcher 1994). Previous literature indicates that under some unstable sample conditions it may be even necessary to have a sample of 400–800 (Fabrigar et al. 2010). Empirical test results from each step of our analysis are presented next.

Confirmatory Factor Analysis for Measurement

Confirmatory factor analysis was performed on all items corresponding to the seven constructs measured in Likert-type scales. The goodness-of-fit of the CFA model was evaluated using a variety of fit metrics, as shown in Table 1. The root mean square error of approximation (RMSEA) is smaller than 0.08 and the root mean square residual (RMR) is smaller than 0.05. Whereas the goodness-of-fit index (GFI) is slightly smaller than 0.90, the other indices including the comparative fit index (CFI), the non-normed fit index (NNFI), and the normed fit index (NFI), all exceed 0.90. These figures suggest that the hypothesized CFA model in this study fits well within the empirical data.

Table 1 Standardized loadings and reliabilities
Full size table

Three primary criteria (Fornell and Larcker 1981) were examined to confirm the convergent validity of the empirical data in this study. First, as evident from the t statistics listed in Table 1, all factor loadings were statistically significant at p < 0.001, which was the first requirement to assure convergent validity of the construct. Second, the average variance extracted (AVE) for all constructs exceeded 0.50, revealing that the overall hypothesized items capture sufficient variance in the underlying construct, more than that which is attributable to measurement error. Third, the reliabilities for each construct exceeded 0.70, satisfying the general requirement of reliability for research instruments. Overall, the empirical data collected by this study meet all three criteria required to support convergent validity.

Discriminate validity was assessed by Chi-square difference tests. The advantage of using this technique is the simultaneous pair-wise comparisons for the constructs based on the Bonferroni method. Controlling for the experiment-wise error rate by setting the overall significance level to 0.01, the Bonferroni method indicated that the critical value of the Chi-square difference should be 12.21. Our Chi-square difference statistics for all pairs of constructs in this study exceeded this critical value of 12.21 (see Table 2), thereby supporting discriminate validity for our data sample. Overall, the empirical results of this study suggest that measurement instruments used for measuring the research constructs are statistically adequate.

Table 2 Chi-square difference tests for examining discriminate validity
Full size table

Team-Level PA for Construct Relationships

After aggregation of individual responses to team-level measures had been justified (see Appendix A), Team-level data was analyzed via PA for testing our hypotheses. To reduce unexpected biases, we included various team-level control variables, one of which is the perceived discretionary citizenship, as well as average age, pay satisfaction, organizational trust, average career experience, and average tenure. These variables may be considered influential for performance in organizations (Dokko et al. 2009; Motowidlo 1982; Salamon and Robinson 2008), and were thus controlled in our statistical analyses. Figure 2 presents the test results of the PA, revealing that six out of eight paths are significant.

Fig. 2
figure 2

Path analytical results of the research model. **p < 0.01

Full size image

Moreover, based on our model in Fig. 2, we conducted post-hoc analyses by adding three model paths from our antecedents to the outcome (i.e., team performance) to confirm whether CSR has direct effects on team performance or not. The post-hoc test results confirm that none of three antecedents has direct and significant effects on team performance, further supporting our hypothesized mediating role of team efficacy and team self-esteem. Furthermore, we have conducted post-hoc analyses by excluding our control variables, and found that the empirical results remain unchanged. That is, relationships remain significant, with or without our suggested control variables.

Results

Table 3 lists the correlation matrix for our constructs, based on the actual survey data.

Table 3 Mean, SD, correlation, and reliability results of the main study variables
Full size table

As presented in Fig. 2, six out of our eight model paths were significant at the p < 0.01 significance levels, and these empirical results indicate that only hypotheses H6 and H7 are unsupported, while the other hypotheses are fully supported in this study (i.e., H1–H5 and H8 are supported). The meaning of unsupported hypotheses is intriguing. The insignificant effect of perceived legal citizenship on team efficacy and the insignificant effect of perceived ethical citizenship on team efficacy suggest that team efficacy and team self-esteem are driven by different dimensions of CSR even if they are both constructs of self-evaluations. Nevertheless, the empirical results for the unsupported hypotheses may warrant further study so that the insights behind the insignificant model paths can be accurately interpreted.

Discussion

Our study confirms positive and indirect influences in three dimensions of CSR on team performance, complementing previous research that considers CSR empirically as a purely single construct (Brammer et al. 2007). In addition, we established a critical bridge between CSR and team performance by exploring the mediation impact of team efficacy and team self-esteem. Previous studies link CSR mostly and directly with its outcomes such as profits or financial performance (Becker-Olsen et al. 2006; Baruch and Ramalho 2006).

Theoretical Implications

Our study complements previous qualitative research of Wood (1991) by embedding CSR into an I–P–O theory and by empirically testing CSR within the boundaries of this theory. This approach has rarely been tried in previous study. In fact, the input-process-output models are considered a dominant theoretical perspective on team performance (Guzzo and Shea 1992). For example, previous research found support for the input- process-output model when member satisfaction and team self-ratings of production were used as criterion measures (Gladstein 1984; Stewart and Barrick 2000). Given the importance of I–P–O system, our study is one of a few studies which provide theoretical contributions by applying it to contexts of CSR. More importantly, our multiple operationalization for various dimensions of CSR is a value-added advantage because it helps determine whether a “true” relationship exists in the high-tech industry context with our two focal mediators (team efficacy and self-esteem).

The significant influence of team efficacy and team self-esteem (i.e., mediators) on team performance suggests that both mediators should be taken as checking points for monitoring how in actual fact CSR affects team performance. These findings are relevant for strategic thinking of CSR, as it will enable management to learn about what kind of corporate citizenship should be strategically strengthened in order to boost team efficacy and team self-esteem, leading to an improvement of team performance. More specifically, each organization has explicit and limited resources, and therefore any organization must make learned choices as to how many different resources should be allocates to CSR activities (Lin 2010a). According to our findings, investing a considerable amount of resources for improving ethical citizenship should lead to team self-esteem and team efficacy.

We suggest that CSR is a most powerful factor that affects team performance through mediating paths, indicating that the prime social responsibility of a corporation is to its employees (see Conduit and Mavondo 2001 and McWilliams and Siegel 2010, for example). The key role of CSR is theoretically and practically justifiable since team performance is unlikely to improve if the team members lack confidence, respect and trust (Frazier et al. 2010).

The significant influence of perceived legal citizenship on team performance through team efficacy suggests that legitimacy brings about righteous and self-confidence among team members, consequently facilitating team performance. Managers should promote business codes and regulations related to law, and clarify any confusion to avoid employees misunderstanding over legal citizenship issues. Besides, management has to avoid illegal opportunistic behavior that hazards business legitimacy (Lin 2010a), as such conduct actually results in a guilty conscience of team members, a lack of their confidence and eventually weakens their team performance.

Finally, the significant influence of perceived ethical citizenship on team performance through team self-esteem suggests that ethics adds to people’s appraisal and beliefs in their own worth and hence encourages their team performance. Managers should resolve to apply positive businesses practice beyond satisfying legal issues, as a long-term perspective. For example, firms should refrain from driving up prices dramatically, taking advantage of temporary market shortages for specific goods. Practicing positive ethical management provides moral support for employees, enabling them to focus on their job. Team members perceiving that they belong to an ethical organization and being proud of their team membership would hold positive beliefs about their own worth too and will generate high quality outputs.

Managerial Implications

Managers should realize that team performance is indirectly linked to CSR. Without periodically examining the two mediators (team efficacy and team self-esteem), managers might blindly pursue trendy issues of CSR in the business world, ignoring the role of these basic mediators. In such a case, managers will be unlikely to make good use of their resources in CSR contexts for improving team performance.

Our findings indicate that team performance is indirectly related to various kinds of corporate citizenship including economic, legal, ethical, and discretional citizenship, suggesting “Do as you would be done by.” The initiatives by managers to improve these three different kinds of citizenship can win high ‘return for investment’ from team workers through their enhanced team efficacy, team self-esteem, and consequently team performance. First, in terms of economic citizenship, managers should frame their organizational climate for facilitating team workers’ intrinsic motivation, such as arranging senior mentors to support personal growth, caring for life outside work, and creating inspiring teaming environment. Second, in terms of legal citizenship, managers should enforce practices to prevent any violation of the law, such as monitoring new legal developments (e.g., updating work hours and minimum wages), planning courses in business law and regulation for team workers (e.g., to avoid patent stealing), etc. Last, in terms of ethical citizenship, managers should insure good ethical conduct within the organization, such as establishing hotlines for complaints about the company’s wrongdoing (e.g., toxic waste), periodically examining organizational justice for different stakeholders, etc.

Note that team performance cannot be arbitrarily obtained by an immediate decree of management, but rather it can be achieved after employees thoroughly observe their firm’s actions in different social perspectives (e.g., ethical citizenship). The view of multiple influencers (the three dimensions of CSR) is quite different from the one prevailing in traditional literature, which tends to focus on team structure and processes, affecting team performance, but neglecting to recognize the necessity of corporate social responsibilities. The given definitions of the three dimensions of CSR are closely interrelated with the different concepts and values of employees (Turker 2009). By understanding the dimensions in depth, management can learn to tailor a variety of teamwork policies or programs to individuals’ needs in order to increase their team efficacy and team self-esteem.

Limitations and Future Research

We do acknowledge three limitations regarding the interpretations of the results. The first limitation is the cross-sectional nature of the data collection. That is, the causality between CSR and team performance can only be confirmed in future research with longitudinal data. The second limitation is generalizability, due to the highly specific nature of the subject sample across IT firms in a single country setting. The inferences drawn from such a sample in Taiwan may not be fully generalizable to employees from different industries or from other countries in quite a different national culture. The third limitation is the possibility of common method bias, given that the constructs of this study were measured perceptually using Likert scales from the same subjects. To test for this bias, a Harman’s single factor test (Podsakoff and Organ 1986) was performed herein. The results of this test have revealed that no single factor accounted for the majority of the variances, suggesting that the common method bias is an unlikely hurdle in our data sample. Nevertheless, future research can improve the shortcoming of this study by collecting data from different sources to avoid the risk of the common method bias. The fourth limitation is that due to the research scope that focuses on CSR across teams we did not address institutional, organizational, or personal variables. For example, a variety of factors such as organizational commitment, organizational trust, emotional attachment, emotions at the workplace, team identification, etc., were not investigated in this study, even though they may have an influence CSR or team performance. Future scholars may try to add to current knowledge by including more critical variables, surveying more samples across various industries, and observing research subjects across different time points, so that the genuine influences of CSR on team performance can be transparently examined from a longitudinal aspect. Lastly, our findings point out a new direction for future research, the search for additional potential mediators (e.g., team identification, psychological contract, etc.) to better understand the linkages between CSR and team performance. In addition to team performance, some outcomes such as team knowledge sharing and team knowledge creation may be tested in future studies.