Antecedents and consequences of collective empathy in software development project teams

Information & Management 52 (2015) 247–259

Contents lists available at ScienceDirect

Information & Management journal homepage: www.elsevier.com/locate/im

Antecedents and consequences of collective empathy in software development project teams Ali E. Akgu¨n *, Halit Keskin, A. Yavuz Cebecioglu, Derya Dogan Science and Technology Studies, Gebze Institute of Technology, Turkey

A R T I C L E I N F O

A B S T R A C T

Article history: Received 8 February 2014 Received in revised form 18 September 2014 Accepted 16 November 2014 Available online 23 November 2014

The term empathy has attracted many researchers from a variety of disciplines; however, a team’s collective empathy, which is composed of cognitive, affective, and behavioral dimensions, has rarely been addressed in the literature. In this study, we empirically investigated the relationship between the collective empathy of a team and the effectiveness of its project process. Additionally, we tested the role of team intimacy-related factors, such as interpersonal trust, within-team communication, and team member familiarity, in collective empathy, as well as the moderating role of group norms on the collective empathy-process effectiveness link. By studying 122 software development projects, we found that cognitive-based trust, formal within-team communication, and team member familiarity influence the collective empathy of project teams. We also found that collective empathy affects team learning and product speed-to-market and results in lower project development costs. Furthermore, we determined that the existence of group norms moderates the relationships among collective empathy, speed-to-market, and lower development costs. The managerial and theoretical implications of the study have also been provided. ß 2014 Elsevier B.V. All rights reserved.

Keywords: Collective empathy Team learning Speed-to-market Software development Project teams

1. Introduction Software development teams, composed of people from different functional areas who have varying technical skills and personalities, are critical for the success of firms’ software development and implementation projects [44]; such projects involve specific activities that start and end at identifiable points in time and produce quantitative and qualitative software deliverables [77]. Researchers have indicated that software development teams are knowledge-intensive social bodies in which team members interact, behave, and organize and then share their information/ knowledge to develop better and faster new software products [82]. Additionally, researchers have asserted that software development teams have their own emotions, and each team member’s understandingof the other team members’ emotions, similar feelings or relevant feelings, and response (i.e., interpersonal empathy) are vital for enhancing software development project performance [15,83]. Thus, interpersonal empathy among team members evokes people’s altruistic motivations [39], increases their

* Corresponding author at: Gebze Institute of Technology, Isletme Fakultesi, Istanbul Cad. No. 101, 41400 Gebze-Kocaeli, Turkey. Tel.: +90 2626051457. E-mail addresses: [email protected], [email protected] (A.E. Akgu¨n). http://dx.doi.org/10.1016/j.im.2014.11.004 0378-7206/ß 2014 Elsevier B.V. All rights reserved.

concern for the welfare of the team/group as a whole [17], and helps them to better resolve conflicts within the team [27]. Neverthless, whereas most studies have discussed or investigated the concept of empathy at the individual level, such as interpersonal empathy in work groups or teams [17], few studies have suggested empathy at the team level as a collective phenemenon, that is, collective empathy, in software development teams [4]. Additionally, although the term collective empathy is partially or implicitly mentioned in the studies of emotionally intelligent teams [39,30], the emotional capability of organizations [4,48], group emotions [14,62], group/organizational compassion [50], and corporate philanthropy decisions [73], it has not been conceptualized or operationalized for an empiricaltest in the software development context. In addition to the lack of operationalization of collective empathy, the antecedents and consequences of collective empathy in software development teams have not been investigated from a managerial perspective in the literature thus far. Furthermore, the moderating variables that shape the relation between collective empathy and project-related outcomes (i.e., consequences) have received less attention in the literature. Indeed, when a project team has too much collective empathy among its members, team members may develop deep emotions (e.g., sympathy) or a group-think phenomenon. Conversely, with a

248

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

lower level of collective empathy, team members may display egocentricity and narcissism, thereby destroying the potential for insightful thinking during the new software development project. Thus, the empirical investigation of whether the relation between collective empathy and project-related outcomes should be regulated with group-related processes is warranted. To address the above mentioned issues, we argue that team collective empathy can be conceptualized as an appreciation and understanding of what team members are experiencing emotionally, as well as an emotional reaction to other team members’ feelings during the software development project [4,30,50,73,53]. Specifically, collective empathy is a shared state of empathy that includes more than one person and indicates the extent to which team members collectively empathize within the teamduring the software development project. Regarding the operationalization of collective empathy, based on individual-level studies [65,31] and group-level studies [4,50,53,86], we suggest that collective empathy is a multidimensional construct composed of (a) cognitive (i.e., collective perspective taking or the extent to which team members attempt to understand each other by imagining the others’ perspective), (b) affective (i.e., collective empathic concern or the extent to which team members feel concern for a person or group of people), and (c) behavioral (i.e., outward display of empathy or affective responsiveness) dimensions. These dimensions are reflective measures, which are observed variables that serve as manifest indicators of the collective empathy of a project team. For antecedents of collective empathy, we investigated team intimacy-related factors. The rationale is that, whereas previous studies identified a variety of factors, such as social connections [90], role clarity [72], and information sharing [33], that influence the development of empathy in interpersonal relations, there is a common argument, influenced by attachment theory [19], in the literature that intimacy (i.e., feelings of closeness, connectedness in relationships) plays a critical role in empathy formation [51]. For example, in their studies on attachment theories, Mikulincer and Shaver [70] discovered that what encourages or muddles empathy is the sense of security and closeness that people feel within themselves. de Vignemont and Singer [23] also suggested that attachment theories provide support for empathy. Nevertheless, intimacy-related factors, such as interpersonal trust (i.e., reliance on the integrity, ability, or character of team members) [49], within-team communication (i.e., the exchange of information in a formal and informal manneramong team members) [61], and team member familiarity (i.e., team members’ past interactions) [25], as contributors to the development of collective empathy at the project team level, are not explored in

the literature; this warrants empirical study, as recommended by Rosh et al. [87]. Regarding the consequences of empathy, we investigated project process effectiveness as implicitly recommended by Reus and Liu [83], Nicholson and Sahay [75], and Akgu¨n et al. [4]. Specifically, whereas past studies demonstrated that interpersonal empathy influences group cohesiveness [86], conflict reduction, group motivation [17,27], and ethical decision making [73,68], the role of collective empathy in project performance in general; in addition, process effectiveness, in particular, has not been empirically investigated in software development projects. Here, based on the project management literature, the process effectiveness variables in which project managers have the most interest are team learning (i.e., gathering and implementing new knowledge, solving software product-related problems), development cost, and speed-to-market (i.e., developing and implementing software products quickly) [3]. Finally, for a moderating variable, we selected one of the group process variables, the existence of group norms, as recommended by Kelly and Barsade [53] and Bagozzi et al. [11]. Indeed, researchers have suggested that empathy is shaped or regulated through the establishment and reinforcement of group norms [31], as shown by the forms of feeling rules and display rules [39,48,50,53]. Nevertheless, we know less regarding how the existence of group norms moderates the relation between collective empathy and process effectiveness in software development project teams. Therefore, as shown in Fig. 1, this study investigated (a) the role of team intimacy-related variables (e.g., interpersonal trust, within-team communication, team member familiarity) on collective empathy formation, (b) the impact of collective empathy on software development project process effectiveness (e.g., team learning, speed-to-market, lower development cost), and (c) the moderating role of group norms on the relation between collective empathy and software development project process effectiveness. 2. Collective empathy in project teams The concept of collective empathy, influenced by intergroup emotions theory [14,62], which describes how individual empathic emotion converges to become collective, and affective events theory [99], which illustrates how the needs of others arouse empathy in individuals, is a relatively new research area in the management and group behavior literature. At the group level of studies, for instance, Kelly and Barsade [53], in their mood and emotions studies, described the collective empathy when group members share another’s feelings by placing themselves

Fig. 1. Research model.

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

psychologically in that person’s circumstance. Roberge [86] argued that collective empathy is the mutual agreement of people within a group that feel for one another; the researcher then theoretically noted collective empathy as a moderator in the team diversity and performance link. At the organizational level of studies, Huy [48] noted collective empathy as part of the dynamics of the experiencing variable of the organizational emotional capability construct, which are people’s ability to understand others’ feelings and re-experience those feelings in the organization. Muller et al. [73] argued that collective empathy exists in the corporate philanthropy decision-making process and noted that collective empathy indicates the collectively shared desire to help others in need. We observe that the common theme in these past group-level and organizational-level studies as well as intergroup emotions and affective events theories is that collective empathy, as an inherently social emotion, occurs through interpersonal interactions and enhanced group or organizational identities, such that, when an individual identifies with a group, that ingroup becomes part of the self [62], thus acquiring social and emotional significance. Additionally, past studies and theories have noted that collective empathy emerges when people’s needs or experiences are more likely to be perceived as relevant or similar, such as work-related objectives aligned with their particular needs or working as colleagues or group mates. Those studies and theories further show that collective empathy is distinct from individual-level empathy [92]. Through sharing processes, individuals can experience collective empathy even when they are not personally involved in or exposed to the empathy-arousing event [84]. In addition, collective empathy is continually reactivated and, thus, sustained at higher intensity levels than any one individual’s empathy alone [84]. In addition to the definitions and features of collective empathy, from an operationalization perspective, we also observe that most past studies and theories have perceived collective empathy as a unitary factor [4,48,73], except for Roberge [86], who explicitly argued the affective and cognitive dimensions of empathy within groups. Although, based on the above studies and theories, we observe that the literature on collective empathy is emerging or growing, an empirical investigation of collective empathy in a special type of work group (i.e., software development project team) remains missing in the literature. Therefore, to enhance the theory of collective empathy in software development teams, we argue that collective empathy exists when all team members (e.g., programmers, system developers, testers) perceive or imagine their teammates’ affects, partially feel what others are feeling, and then demonstrate prosocial behaviors within their team during the project [91]. Here, we specifically suggest that collective empathy has three features/dimensions in software development project teams (cognitive, affective, and behavioral empathy); this is consistent with Bandura’s [12] conceptualization of empathy as involving ‘‘social perspective taking, imaginative self-involvement, and emotional responsiveness’’ (p. 314). Cognitive empathy refers to the ability of team members to understand the feelings of others within the team. This aspect of collective empathy resonates with the notion of taking the perspective of the other, understanding others’ worlds [4], or what Goleman [38] noted as ‘‘we know how you see things.’’ For example, when teammates have personal problems and the software project schedule is very crowded, they will normally be more anxious than the other team members. If they feel comfortable describing their situation to their teammates, and if their teammates understand and appreciate their problem and take their perspective, the teammates will feel that they are in an empathic environment.

249

Affective empathy is the affective reaction team members have to the affective states of their teammates. Team members vicariously feel the feelings of others within the team that they have noticed and understood. This dimension of empathy is related to ‘‘we feel with you’’ [38] or ‘‘we are feeling for you’’ [86]. For instance, when team members have difficulty with a new job definition and try to learn details regarding the new tasks, they may feel worried about not being effective. If other team members are truly interested in the problem and make the troubled team members feel that they share the concern (e.g., team members openly express their emotions), the troubled team members feel that they are on an empathic team. Behavioral empathy indicates the ability of team members to respond to the feelings of others within the team [50]. Here, behavioral empathy is engagement in helping behavior, usually expressed as the communicative reaction of people to others’ feelings. For example, when team members have difficulty with task completion, other team members may allow them to finish their tasks over a longer period of time or tell emotion-laden stories regarding how to solve the problems [7]. In this situation, the troubled team members feel that they are in an empathic team environment. Based on the above features/dimensions, we argue that collective empathy provides benefits to software development project teams. Specifically, collective empathy prevents team dissolution by facilitating the development of bonds among team members [93], as well as creating and affirming a sense of groupness [7]. Indeed, with collective empathy, team members are aware that the personality of each team member differs from that of the others and this variety of people and culture influences relationships in the team. In addition, team members understand the strengths and limitations of others and understand the reasons for others’ behaviors (i.e., know what motivates or demotivates them). Additionally, collective empathy helps team members determine what the software development process really needs with respect to new information or ideas, different technical expertise and skills, and extra time and then helps to pursue endeavors to fulfill these needs by developing intergroup interactions and cohesion [39]. Here, emotional information, which refers to information that conveys team members’ states of mind, such as anger, distress, and surprise (i.e., talking about feelings, giving voice to feelings) [95] resulting from collective empathy, enables team members to take in and understand the full range of issues and problems in the software development process, as well as the actions necessary to coordinate and move forward to meet project goals. Having established the characteristics of collective empathy in software development teams, we now develop the arguments regarding the role of antecedents in collective empathy, how collective empathy affects project process effectiveness, and how the existence of group norms moderates the collective empathyprocess effectiveness link. 3. Hypotheses development 3.1. Antecedents of team collective empathy We argue that interpersonal trust, which includes cognitivebased trust (i.e., beliefs regarding others’ competence and reliability) and affective-based trust (i.e., beliefs regarding reciprocated care and concern) [64], fosters the development of collective empathy by providing a psychologically safe environment for team members [83]. This means that when team members face user and implementation-related uncertainties and lack information regarding other people’s intentions and feelings, they are in a psychological state of vulnerability, and thus, they need to trust each other’s

250

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

intentions and feelings. To the extent that such a safe climate is provided, team members will experience fewer self-doubts and less wasted effort to understand, feel, and respond to others’ emotions during their interactions within the team, as Lee and Dubinsky [58] noted. In addition, interpersonal trust reduces the corporate and interaction-related fears of team members regarding others within the team. For example, because established trust among team members reduces their fear of being betrayed, being deceived, or being easily replaced (i.e., the fear of losing their unique value during the project), they are not afraid to share their emotions with others and explicitly express their emotions to each other within the team. Additionally, interpersonal trust among team members enhances collective empathy by leveraging the information/ knowledge exchange among them. The existence of trust among team members is evidence of their honesty in communication and their willingness to share information with each other. Thus, people understand others’ feelings, states of mind, and motives. McKay et al. [67] also noted that members who trust each other know each other’s minds to some extent and can often anticipate each other’s needs and feelings. Therefore, we hypothesized that: H1. A positive relation exists between interpersonal trust and the development of collective empathy in software development teams. Within-team communication, which includes formal communication (e.g., communication through writing, scheduled meetings, and other non-interactive methods and impersonal communication channels) and informal communication (e.g., personal and interactive communication, such as talks in the hallway, quick phone calls, and short e-mails) [46], also enhance team collective empathy by leveraging the social interactions among people. For example, [34] noted that social talks are discourse strategies that assist in establishing and maintaining good relationships with co-workers. Through communication, people have the opportunity to create and share their emotions and others’ norms, values, and culture. Thus, members of the team mutually understand each other’s internal states. Additionally, within-team communication enhances collective empathy by creating an atmosphere of openness and transparency of emotions within the team. Here, team members create their own reasons for the emotions of others with less speculation and rumor. In addition, team members have a sense of ownership in the shared emotions that have been created because they feel they have helped develop those shared emotions. Therefore: H2. A positive relation exists between within-team communication and the development of collective empathy in software development teams. Next, we argue that team member familiarity influences collective empathy by leveraging the emotional bonds or ties among team members. For example, when team members know each other from previous projects or social interactions, they feel comfortable working together on new projects. They also feel attached to each other, and thus they tend to express goodwill toward each other and are more attentive to others’ moods and feelings, as Bartel and Saavedra [16] noted. Team member familiarity also affects the development of collective empathy by making the behaviors of people more predictable during project activities. For example, by anticipating the responses of members to particular types of emotions based on previous experiences together, such as joking, other members can act without concern for negative reactions and provide others greater latitude in their actions. Therefore: H3. A positive relation exists between team member familiarity and collective empathy in software development teams.

3.2. Consequences of team collective empathy We argue that collective empathy positively influences team learning (e.g., covering software-related problems) by creating a more interactive and pleasant atmosphere during the project. Here, team members suspend their judgments and understand their functional differences to foster more enlightened, secure, and reliable relationships [27] and leverage their forgiveness, which reduces destructive behaviors such as rumination, avoidance, and revenge during the project [66]. Therefore, team members feel free to note software and project-related errors and problems without being viewed as incompetent and then to make changes and solve problems in better and more efficient ways with a sense of confidence [29]. Collective empathy also enhances team learning by leveraging team members’ helping behaviors toward each other during the project [9]. When team members focus on the perspective of other team members, being sensitive to what they want, need, and feel, and then accept their personal emotions and needs, they improve one another’s learning efforts in projectrelated activities [17]. In addition, team members donate their time [73] and knowledge for the benefit of others, which develops the transmission of complex tacit knowledge and fosters knowledge integration among them [83]. In addition to team learning, we also argue that collective empathy decreases the project’s development time of the project (i.e., speed-to-market) by enhancing the listening capability of team members. When team members listen to each other during the project, they become more aware of subtle cues from others, are able to quickly understand and identify others’ emotions, and can quickly respond to others’ emotions to ensure that messages are effectively transmitted throughout the project team. In addition, team members enhance their negotiation skills [63], which provides them confidence in conducting their projectrelated jobs quickly. Accordingly, the project team performs and completes project-related activities faster. Collective empathy also increases speed-to-market by solving the rational ego/selfrelated barriers among team members (i.e., between self and others) and then finding common ground for solution building on project-related problems. In particular, when team members attend to the perspective of other team members and discuss differences of opinion openly rather than privately, they can detect problems and solve them more quickly [29]. Furthermore, we suggest that collective empathy promotes lower development costs by leveraging the connectedness among team members during the project. Here, team members perceive the project team as a ‘‘whole’’ by recognizing their interconnectedness and interdependencies rather than considering the project team merely as a collection of individuals [73]. Additionally, team members examine the project’s larger picture rather than focusing on separate task components during the project, and they consider project and project-related decisions in the context of consequences and implications for mitigating the project’s development cost [5]. Collective empathy also reduces development costs by allowing team members to regulate the negative emotions they experience during conflict or confrontation [27]. Here, team members justify their decisions by considering others’ viewpoints, and they consider other realities and alternative meanings in situations during the project to avoid unnecessary project activities. Therefore: H4. A positive relation exists between collective empathy and (a) team learning, (b) speed-to-market, and (c) lower development cost. 3.3. Moderating effect of group norms We argue that the existence of group norms moderates the relation between collective empathy and project process

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

effectiveness. In particular, the existence of group norms influences the empathy-effectiveness link by regulating the emotions of people during the project. Here, the successful regulation of emotions allows team members to refocus their own and others’ attention on important project-related problems [83] rather than spending their valuable time to understand and then respond to each other’s emotions [30]. The existence of group norms also affects the empathyeffectiveness link by providing the advantage of free displays of emotions to team members [30,53]. For example, the suppression of conflict-related emotions can lead to anxiety and selfrighteousness in the team, but when emotions are appropriately expressed, team members can affirm interdependency and then direct their focus and energy to the project and software productrelated problems [94]. In addition, the existence of group norms raises the predictability of team members’ behaviors. As shared group norms generate observable work-related behaviors and attitudes in team members [37], team members can accurately interpret others’ nonverbal emotional expressions and reactions and then assess whether others are addressing work-related issues or solving task/project-related problems effectively [27]. Therefore, we hypothesized that: H5. The existence of group norms positively moderates the relation between collective empathy and (a) team learning, (b) speedto-market, and (c) lower development cost.

4. Research methods 4.1. Measures To test the above hypotheses, we developed or adopted multiitem scales from prior studies for the measurement of variables. We used 5-point Likert scales ranging from ‘‘strongly disagree’’ (1) to ‘‘strongly agree’’ (5) to measure our variables. However, as control variables, we assessed project team size and duration questions with a ratio scale. The appendix includes the measures used. A summary of the measures follows. For the cognitive empathy variable, we modified the question items from individual-level studies (e.g., [20,52]) by asking to what extent team members understand others’ feelings, thinking, attitudes, and perspectives within the project team. For the affective empathy variable, we developed new question items based on Roberge [86], including items inquiring into the extent to which team members become emotionally involved with others’ feelings and the extent to which team members’ attitudes are affected by the feelings of others within the project team. With regard to behavioral empathy, we extended the question items of Akgu¨n et al. [4] by inquiring into the extent to which team members react in response to others’ feelings and to which team members demonstrate verbal and non-verbal communication reactions in response to others’ feelings within the project team. With respect to process effectiveness variables, we adopted question items from the project management literature. We adapted the team learning variable from Akgu¨n et al. [2] by asking four questions covering the extent to which the team did an outstanding job correcting the software product’s problematic areas with which users/customers were dissatisfied and the degree of customers’/users’ perception that this software product had fewer problems than what was considered normal in the company. For speed-to-market, which is the team’s ability to develop the new software product rapidly, we used question items from Kessler and Chakrabarti [54]. The speed-to-market assessment was measured in relation to preset schedules, company standards, and similar competitive projects. Regarding the lower implementation

251

cost, we adapted question items from Griffin and Page [40], including questions covering the extent to which the project finished at or below cost estimates for planning, design, and implementation. For team member familiarity, we used question items from DeChurch and Marks [25] by asking to what extent team members knew the others (on average) and interacted with the others (on average) when the project team was formed. For cognitive- and affective-based trust variables, we adapted question items from Kanawattanachai and Yoo [49]. We asked two questions to assess the formal and informal team communication variables, adapted from Lynn and Akgu¨n [61]. For group norms, we modified past studies (e.g., Gelfand et al. [37]) by asking four questions regarding the existence of social norms (such as customs, etiquette, and work ethic) by which project team members should abide. After developing the new question items in English, the question items were first translated into Turkish by one person and then retranslated into English by a second person using the parallel translation method. The two translators then jointly reconciled all differences. A draft questionnaire was developed and then evaluated and revised in discussions with academics from Turkey who have knowledge of organizational behavior and software development as expert judges. The suitability of the Turkish version of the questionnaires was then pre-tested by five part-time graduate students who are full-time employees working in industry. In addition, four project managers, randomly selected from a diverse cross-section of firms located in Istanbul, evaluated the content and meaningfulness of the items. Respondents did not demonstrate any difficulty in understanding the items or scales. After confirming the questionnaire items, the questionnaires were distributed and then collected by the authors, applying the ‘‘personally administered questionnaire’’ method (i.e., the authors of this study personally administered the questionnaires to the respondents). 4.2. Sampling The initial sample consisted of 100 firms located in Istanbul that have an affiliation with European firms. First, the firms’ managers were contacted by telephone and the study’s objective was explained to them. Of the 100 firms contacted, 64 agreed to participate in the study. We asked at least two respondents who are the most knowledgeable regarding the projects to complete our surveys to avoid single-source bias in each software development project. After selecting the respondents, we informed them that all responses would remain anonymous and would not be linked to them individually, their companies, or software products. This step was performed to ensure anonymity, thereby increasing informants’ motivation to cooperate without fear of potential reprisals. Additionally, we assured respondents that there were no right or wrong answers and that they should answer questions as honestly and forthrightly as possible. Furthermore, we developed a cover story to make it appear that the measurement of the predictor variable was not connected or related to the measures of the criterion variable. These procedures reduced people’s evaluation apprehension and made them less likely to edit their responses to be more socially desirable, lenient, or consistent with how they thought the researchers wanted them to respond [81]. Of the 64 firms that agreed to participate, 29 completed our questionnaires; 244 surveys from 122 projects were returned. Thus, usable data for our analysis include 122 software development projects with an average of two respondents from each. In the sample, the projects were related to financial services (44%), information and communication technologies (29%), and business services (27%). The respondents held the following positions: engineer/programmer (48%), information services (IS) specialist/

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

252

analyst (19%), product/project manager (12%), tester/quality control (9%), department manager (7%), and senior engineer/ technical leader (5%). The duration of the projects was less than 3 months (32%), 4–6 months (24%), 7–9 months (19%), 10–18 months (18%), and longer than 18 months (7%).

5. Analysis and results 5.1. Measure validity and reliability After data collection, we assessed the reliability and validity of measures by employing a purification process [6,36]. Because the collective empathy construct is new, we first conducted an exploratory factor analysis including 17 measured items of three variables, using a principle component with a varimax rotation and an eigen value of 1 as the cutoff point. We found that the Kaiser– Meyer–Olkin (KMO) measure of sampling adequacy was 0.89, and the Bartlett test of sphericity was significant at p < .01, indicating the suitability of these data for factor analytic procedures. Table 1 shows the items and their factor loadings after exploratory factor analysis, eigen value, and percentage of variance explained. After performing the exploratory factor analysis, which is useful for scale construction, we conducted a subsequent confirmatory analysis to assess the resulting scales of collective empathy. The initial results of the confirmatory factor analysis (CFA) revealed that the initial model did not fit adequately (x2ð116Þ ¼ 354:04, CFI = .87, RMSEA = .10). After elimination of the problematic items that had a low factor loading or a cross-load to the other variables in a step-by-step procedure, as demonstrated in the Appendix, the results indicated that the models adequately fit the data. Additionally, the fit indexes were x2ð62Þ ¼ 164:22, CFI = .93, RMSEA = .09. Next, we performed a series of two-factor model tests [10]. In total, we evaluated six models using AMOS 4.0. We found that the chi-squared changes (Dx2) in each model, constrained and unconstrained, were significant, Dx2 > 3.84, suggesting that collective empathy variables demonstrate discriminant validity, as shown in Table 2. Furthermore, we conducted a second-order CFA of a model depicting the collective empathy with 13 observable items. This model also yielded acceptable fit indexes (x2ð64Þ ¼ 168:45; CFI = .92, RMSEA = .07). In addition, all first-order and second-order factor Table 1 Discriminant validity of construct measures factor rotation. Constructs

Items

Cognitive empathy (F1)

CE1 CE2 CE3 CE4 CE5 CE6 CE7 CE8 AE4

.82 .76 .78 .77 .79 .80 .60 .71 .67

.11 .16 .13 .15 .09 .17 .18 .07 .27

.04 .13 .27 .19 .11 .23 .25 .30 .07

Behavioral empathy (F2)

BE1 BE2 BE3 BE4

.22 .17 .05 .45

.81 .82 .77 .61

.17 .29 .25 .09

Affective empathy (F3)

AE1 AE2 AE3 AE5

.49 .18 .26 .04

.09 .25 .21 .45

.62 .74 .76 .66

7.66

2.27

45.11

13.35

Eigenvalue % of variance explained

F1

F2

F3

1.078 6.33

Table 2 Discriminate analysis of the construct measures. Constructs

Unconstrained (x2/d.f)

Constrained (x2/d.f)

Dx2

Cognitive empathy (F1) vs behavioral empathy (F2) Cognitive empathy (F1) vs affective empathy (F3) Behavioral empathy (F2) vs affective empathy (F3)

189.85/64

262.33/65

72.48

121.78/34

188.42/35

66.64

38.16/13

105.83/14

30.9

All Dx2 are significant at P < .05 level.

loadings provided evidence that collective empathy is a multidimensional construct composed of cognitive, affective, and behavioral dimensions. In addition to the collective empathy construct, we also evaluated the reliability and validity of our antecedents and consequences variables using CFA [36]. To assess unidimensionality, we divided measures into two subsets of theoretically related variables: (1) six antecedent variables (i.e., cognitive-based and affective-based trust, informal and formal communication, team member familiarity, and group norms) and (2) three project outcome measures (team learning, speed-to-market, and development cost) using AMOS, as Moorman and Miner [71] recommended. After eliminating problematic items that had low factor loadings or cross-loads with other variables in a step-by-step procedure, the results indicated that the two models fit adequately: six antecedent variables (x2ð75Þ ¼ 149:89, CFI = .95, RMSEA = .07) and three outcome variables (x2ð32Þ ¼ 73:96, CFI = .96, RMSEA = .08). Finally, we subjected all the measures to CFA by including all factors in one CFA model. During the CFA analysis, we used subscales or parcels (a method aggregating or taking the mean of several items that purportedly measure the same construct as indicators of a latent variable) for the CFA instead of individual items, as recommended by Schmit and Ryan [89]. These researchers noted that goodness-of-fit measures are affected when the number of items used to identify a small number of factors is relatively large. Consistent with this approach, we created two sub-scores or parcels for each scale, each consisting of a randomly divided subset of the items in the scale. The CFA produced a good fit with a comparative fit index (CFI) of .95 (also, x2ð186Þ ¼ 303:27, RMSEA = .05). Table 3 shows the correlations among all variables. The relatively moderate correlations provide further evidence of discriminant validity. Additionally, all reliability estimates, including coefficient alphas, average variance extracted (AVE) for each variable, and AMOS-based composite reliabilities, are well beyond the threshold levels suggested by Fornell and Larcker [36]. Furthermore, as Fornell and Larcker [36] also suggested, the squared root of AVE for each construct was greater than the latent factor correlations between pairs of constructs, suggesting discriminant validity. The results strongly indicate that the measures are unidimensional and have adequate reliability and discriminant validity. 5.2. Hypothesis testing To test our hypotheses, we used partial least squares (PLS) as our analysis methodology because PLS avoids many of the restrictive assumptions underlying maximum likelihood techniques and ensures against improper solutions and factor indeterminacy [22]. Additionally, PLS is insensitive to sample size

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

253

Table 3 Correlations and descriptive statistics. Variables Cognitive empathy Affective empathy Behavioral empathy Team learning Speed-to-market Less development cost Cognitive-based trust Affective-based trust Informal communication Formal communication Team familiarity Group norms Team size (logarithmic) Project duration (logarithmic) Mean S.dev. Average var. ext. (AVE) Composite reliability Cronbach’s a Inter-rater agreement (rwg)

1 2 3 4 5 6 7 8 9 10 11 12 13 14

1

2

3

4

5

6

7

8

9

10

11

12

13

14

(.77) .49*** .35*** .57*** .45*** .31*** .64*** .56*** .30*** .33*** .51*** .39*** .07 .07

(.70) .51*** .35*** .19** .14 .31*** .37*** .31*** .19** .18** .40*** .13 .07

(.80) .19** .13 .05 .22** .38*** .31*** .17* .19** .29*** .17* .20**

(.78) .52*** .34*** .47*** .47*** .24*** .05 .43*** .27*** .13 .03

(.85) .69*** .41*** .39*** .14 .22** .46*** .29** .22** .23**

(.85) .29*** .31*** .09 .17* .32*** .08 .13 .31***

(.77) .65*** .28*** .16* .18** .29*** .15* .20**

(.73) .34*** .27*** .48*** .31*** .05 .01

(.84) .24*** .29*** .18** .02 .06

(.82) .27*** .19** .04 .13

(.90) .28*** .21** .25**

(.75) .03 .01

NA .42***

NA

3.53 .59 .59 .90 .89 .87

3.39 .56 .49 .80 .79 .75

3.47 .55 .63 .84 .83 .79

3.62 .64 .61 .86 .85 .81

3.34 .86 .71 .88 .88 .86

3.24 .83 .72 .89 .88 .85

3.83 .65 .59 .85 .85 .82

3.63 .60 .53 .69 .71 .70

3.48 .84 .71 .83 .82 .80

3.39 .86 .66 .79 .83 .79

2.42 .71 .80 .90 .89 .88

3.26 .75 .56 .72 .72 .71

.81 .35 NA NA NA NA

.84 .41 NA NA NA NA

Note: Numbers on diagonals indicate square root of AVE. No correlation is greater than the corresponding square root of AVE. * p < .1. ** p < .05. *** p < .01.

considerations and handles both very small and very large samples with more ease than does structural equation modeling (SEM) [43]. Furthermore, PLS handles both reflective and formative constructs [43]. Before conducting any analysis, because the unit of analysis is the ‘‘project team,’’ we aggregated the team scores of each question item. All inter-rater agreement (rwg) values ranged from .72 to .91, well above the .60 benchmark [47], indicating a satisfactory level of inter-rater agreement for each aggregate measure in a project team (see Table 3). We used bootstrapping samples to estimate the standard errors and to test the statistical significance of the structural paths by using SmartPLS 2.0 [85]. This procedure entailed generating 500 sub-samples of cases randomly selected, with replacement, from the original data. Path coefficients were then generated for each randomly selected subsample. T-statistics were calculated

for all coefficients based on their stability across the subsamples, indicating which links were statistically significant. As shown in Table 4, the results illustrate that antecedent- and consequence-related hypotheses are largely confirmed. With regard to antecedents, we found that cognitive-based trust (b = .35, p < .01) is positively associated with collective empathy, whereas affectivebased trust (b = .18, p > .1) is not related to collective empathy, partially supporting H1. Regarding H2, we found that formal communication (b = .15, p < .1) is positively associated with collective empathy. However, we did not find any statistically significant relation between informal communication (b = .14, p > .1) and collective empathy, partially supporting H2. Furthermore, we found that team member familiarity (b = .16, p < .1) is positively related to collective empathy, supporting H3. With regard to consequences, the results show that collective empathy is positively related to team learning (b = .53, p < .01),

Table 4 Results of PLS analysis. Hypotheses

Relationship

Path coefficient (b)

Results

H1

Cognitive-based trust ! collective empathy Affective-based trust ! collective empathy

.35*** .18

Partially supported

H2

Informal com. ! collective empathy Formal com. ! collective empathy

.14 .15*

Partially supported

H3

Team familiarity ! collective empathy Collective empathy ! team learning

.16* .53***

Supported

H4

Collective empathy ! speed-to-market Collective empathy ! less dev. cost Project duration ! collective empathy Team size ! collective empathy

.38*** .25** .12 .11

Supported

Fit measures

Endogenous construct

Final model

R2

Collective empathy Team learning Speed-to-market Less dev. cost

.48 .28 .15 .06

* **

p < .1. p < .05. p < .01.

***

254

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

speed-to-market (b = .38, p < .01), and lower development cost (b = .25, p < .05), supporting H4. The results in Table 4 show that interpersonal trust, communication, team member familiarity, and control variables1 explain 48% of the variance (R2 = .48) in collective empathy and collective empathy explains 28% of the variance (R2 = .28) in team learning, 15% of variance (R2 = .15) in speed-to-market, and 7% of variance (R2 = .07) in lower development cost. Regarding the moderating hypothesis, we employed a hierarchical approach to test our hypotheses, in which we first estimated a model with the main effects only and then added the existence of group norms and interaction effects. We mean-centered the variables involved in the interaction terms to guard against multicollinearity [1]. As Table 5 illustrates, we found a significant statistical association between the interaction effect of group norms and collective empathy (collective empathy * group norms), speed-to-market (b = .25, p < .05), and lower development cost (b = .26, p < .05). However, we did not find any moderator role of norms on the relation between collective empathy and team learning (b = .08, p > .1). To further assess the association of collective empathy and the moderators’ influence on speed-tomarket and lower development cost, the R2 values of the dependent variables were compared. Accordingly, we computed the effect size (f2) to compare the R2 values between the main and interaction effect models as a gauge of whether the interactions had a small (0.02), medium (0.15), or large effect (0.35) on speedto-market and lower development cost. We found that the existence of group norms has a positive moderating effect on the relation between collective empathy and speed-to-market ( f2 = .15) and collective empathy and lower development cost (f2 = .07), partially supporting H5. 6. Discussion and implications This study has demonstrated the interrelations among team intimacy-related factors, collective empathy, and process effectiveness variables in software development project teams. Specifically, this study first empirically showed that interpersonal trust in general, and cognitive-based trust in particular, is positively related to the development of collective empathy in software development teams, which elevates the previous studies on interpersonal trust in software development teams. Whereas prior studies found that trust affects virtual team performance [79], knowledge-sharing effectiveness [76], conflict resolution [100], and team integration [26], we specifically demonstrated that 1 In this study, we could not find any significant effect of team size (b = .11, p > .1) and project duration (b = .11, p > .1) on collective empathy development. The reason might be that our team size and duration variables involve a variety of project team sizes and project durations during the analysis. In this respect, we divided our sample into small versus large project team size and short versus long project duration time for four post hoc analyses. We assessed the small project team whose members number less than eight (n < 8) and large project team whose members number more than eight (n  8), as Carmel and Bird [21] suggested. We re-analyzed our model, and our results revealed that the team size variable in the larger team sample, involving 46 teams, had a significant effect on collective empathy (b = .15, p < .1), whereas the team size variable in the smaller sample, involving 76 teams, had no significant effect on collective empathy (b = .02, p > .1). With regard to project duration, we assessed the short-term project that takes less than six months (t < 6) and the long-term project that takes more than six months (t  6), as suggested by Meyer and Curley [69]. We re-analyzed the model and our results demonstrated that duration of the project, whether shorter or longer, had no significant effect on the development of collective empathy. Accordingly, these findings show that when team size is larger, the project team needs more collective empathy development. On the other hand, in a smaller team, collective empathy is either implicitly embedded or emerges during the project. Regarding project duration, the development of collective empathy is less bounded to the time frame of the project. Here, as seen in the group think phenomenon, sympathy rather than empathy among team members may be developed over time, which warrants future empirical research.

Table 5 Results of moderations effects. Relationship Main effect

Interaction effect

Fit measures R2

* **

Collective empathy ! team learning Collective empathy ! speed-to-market Collective empathy ! less dev. cost Collective empathy * norms ! team learning Collective empathy * norms ! speed-to-market Collective empathy * norms ! less dev. cost Team learning Speed-to-market Less dev. cost

Model 1 .53

***

Model 2 .51***

.39***

.27**

.26**

.22*

–

.08

–

.25**

–

.26**

.29 .15 .07

.30 .27 .13

p < .1. p < .05. p < .01.

***

when team members have confidence in the ability or competence of others, they understand the feelings of others, become emotionally involved with others’ feelings, and react/respond to others’ feelings during the project. Interestingly, in this study, we could not find a statistical relation between affective-based trust and the development of collective empathy. This maybe related to the precursor role of cognitive-based trust on affect-based trust formation, as Van Kleef et al. [97] noted, or the cycling process of trust building, as Vangen and Huxham [98] suggested. Here, our post hoc analysis also revealed that cognitive-based trust influences collective empathy (b = .48, p < .01) and that collective empathy strongly influences affect-based trust (b = .57, p < .001). Additionally, our post hoc test on the mediating role of collective empathy between cognitivebased trust and affective-based trust, using Baron and Kenny’s [13] procedure, demonstrated that collective empathy partially mediates the relation between cognitive-based trust and affectivebased trust. Accordingly, our non-significant finding on the relation between affective-based trust and collective empathy empirically leverages Lewicki and Bunker’s [59] proposal, suggesting that cognitive-based trust gives rise to affective-based trust in which displays of empathy are central. Thus, team members form expectations regarding the intended project-related outcomes and the way other team members will contribute to achieving them based on cognitive-based trust during the project. When a projectrelated outcome meets the expectations of team members, their trusting attitudes toward each other are reinforced, which helps them better understand each other’s feelings (e.g., collective empathy formation). The team’s collective empathy then triggers emotional openness without significant concern for the emotional vulnerability of team members or the perceived strength of the affective relationships among team members (e.g., affective-based trust formation). This study also demonstrated that formal within-team communication influences the development of collective empathy, which empirically supports Sakka et al. [88]. Here, when team members conduct frequent formal communications through team meetings and memos with fellow project team members, they can understand the feelings of others, are affected by the feelings of others, and respond to the feelings of others. Thus, formal communication leverages team collective empathy by providing credible sources of information and reducing the negative effects of rumors [96]. Conversely, we did not find a statistical association between informal

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

communication and the development of collective empathy. The reason maybe attributed to the interrelations among formal and informal communication and collective empathy. For example, the literature has implicitly noted the impact of formal communication on empathy and the effect of that empathy on informal communication [55]. Indeed, informal communication, which compensates for the weakness of communication formation, is spontaneous and voluntary and is based on social relationships among people. The level of informal communication is influenced by how people understand, share, and respond to each other’s feelings; this is collective empathy. Our post hoc analysis also showed that formal communication influences collective empathy (b = .18, p < .05) and collective empathy strongly affects informal communication (b = .34, p < .01). Additionally, we found that collective empathy fully mediates the relation between formal communication and informal communication by using Baron and Kenny’s [13] procedure, empirically leveraging Kodama’s [55] arguments. Next, this study empirically illustrated the impact of team member familiarity on the development of collective empathy in software development project teams; this extended past studies [73]. For instance, previous studies demonstrated that familiarity among people influences effective information sharing and problem solving [42] and coordination, cohesion, and trust formation among team members [32]. Here, we specifically showed that, when people know each other from previous projects, they are empathically attuned, emotionally responsive, authentically present, and open to change. Second, this study showed the consequences of collective empathy in software development project teams. Specifically, we empirically demonstrated that collective empathy influences the learning behaviors of software development project teams, which leverages the literature on learning in software development teams. For example, previous studies examined the effect of knowledge management [75], team psychological climate [29], team size [8], and group interactions [28] on team learning in software development projects, ignoring the emotional aspects of project teams. Here, we have observed that when people understand, share, and respond to each other’s feelings, they are more adept at discovering and correcting software product-related problems and issues. This study also showed that collective empathy influences the project team’s ability to develop and implement software products faster (i.e., speed-to-market), leveraging the study of Akgu¨n et al. [4], who investigated collective empathy as a unidimensional construct. In addition, this finding showed that the ability of a project team to remain quick does not just depend on the use of advanced technology with which information can be located, processed, and disseminated, but depends on the manner in which collective empathy and emotions are managed and organized within the project team.2 This study further showed that collective empathy affects project development costs. Whereas past studies identified a variety of project- and team-related factors influencing development cost (see Lagerstro¨m, et al. [57], for a review), we demonstrated that understanding, sharing, and responding to the feelings of team members constitute a critical cost reduction factor in projects. In 2 Managing emotions can often be difficult as team members may have habitual ways of responding to events. However, studies on ‘‘emotional labor’’ [45], ‘‘emotional regulation’’ [35], ‘‘philanthropic emotion management’’ [18,56], and bounded emotionality [74] indicate that emotion can be managed. Those studies argued that emotions can be managed by social and cultural influences demonstrated within the emotional display rules of an organization and work setting, on-the-job learning approaches where individuals receive feedback on the consequences of the particular emotional management behaviors they employ, and enhanced social exchange and interaction among management, employees, and customers depending on the feeling rules of situations.

255

particular, this finding highlights the role of soft skills in project teams (e.g., collective empathy) on project development costs. When team members increase their ability to adapt and optimize the advantages of others’ feelings, they reduce rework, redundancies, and quality problems, which ultimately reduces the software development and implementation costs [78]. Third, this study empirically demonstrated the moderator role of group norms on the relation between collective empathy and project process effectiveness. This finding improves our understanding of ‘‘project control’’ through regulating collective empathy of the project team to evaluate its software development time and reduce its development cost [41]. Here, we observe that the existence of norms preempts harmful emotions (e.g., avoiding conflicts in social contacts during the project) and moderates the response to or the appraisal of an incident after it has occurred (e.g., discontinuing meaningless or detrimental relationships among people during the project) to ensure the adjustment of team members to changing situations. Additionally, this finding highlights that the existence of group norms turns collective empathy into a ‘‘project resource’’ for performance improvements. The existence of norms allows people to avoid investing in feelings in the relationships that have no value for the development process (e.g., jealousy) and frees up people’s positive feelings (e.g., joy, hope) to cultivate relationships that enhance process effectiveness. Notably, in this study, we did not find a moderator role for norms in collective empathy and team learning; this indicates that whether there is a group norm, collective empathy always influences the team’s ability to learn, such as discovering and correcting project and product-related problems. From this research, management can understand the soft side of project teams to increase project performance. Thus, management should focus on collective emotions in general, and collective empathy in particular, rather than simply following software engineering principles, such as establishing more formalized production methods and using more formalized frameworks and models, to solve software product and project-related problems to develop software products faster with lower costs. Management should also consider that building interpersonal trust is important for the development of collective empathy in software project teams. Accordingly, management should clarify the roles, duties, and responsibilities of team members, align their interests, and provide a psychologically safe environment for them during the projects. Managers should also show benevolent concern for and interest in team members, demonstrate that they value and care regarding team members’ emotions, and leverage their pride and confidence in being a team member. Next, management should enhance formal communication among team members to enhance the collective empathy of the team. Thus, management should promote within-team communication by using emails, letters, and telephone conversations, interactive software, project documents and posts, and project websites. Management should also help team members’ socialization process to leverage members’ desire for human contact during the project. Here, management can allow team members to speak freely regarding their problems, attitudes, feelings, job, or whatever they prefer. Furthermore, management can create social networks or transactive memory systems (i.e., group memory systems that detail the expertise possessed by group members in addition to an awareness of who knows what within the group) [60] to improve communication among people. Management should further consider the prior familiarity of team members to leverage collective empathy. Here, management can consider, to some extent, the team members’ nomination or selection of persons with whom to work. Management can also develop psychological profiles of potential members in an attempt to reduce problems resulting from interpersonal conflict.

256

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

Finally, management should set group norms to regulate the collective empathy of teams during projects. Thus, management should facilitate storytelling in ways that emphasize the existence of norms. Additionally, management should reinforce or limit certain behavioral patterns and emotions by communicating widely and vividly. In addition, management should provide team members opportunities to negotiate a consensus on key values that are used to govern their feelings and then create norms through day-to-day interactions. Note that these norms should address the team’s tasks as well as its social needs. 7. Limitations and future research There are some methodological limitations to this study. Specifically, this study is prone to common method bias because the same respondents answered questions for the dependent variables and independent variables in a cross-sectional manner. We checked this potential problem with the Harman one-factor test [80]. The results of an unrotated principal component analysis indicated that common method variance is not a problem because several factors with an eigenvalue greater than 1 were identified, which explained 73.43% of the total variance, and because no factor accounts for nearly all the variance (i.e., highest single variance extracted = 29.92%). Using a cross-sectional design with questionnaires was also a limitation of this study. However, we should note that as a cross-sectional field study, this study provides evidence of associations. In addition to the nature of the data, the generalizability of sampling is another limitation. The study was conducted in a specific national context, Turkish firms and people (team members) from a Near Eastern culture. It is important to note that readers should be cautious when generalizing the results to different cultural contexts. We believe that team collective empathy triggers opportunities for future research. For example, in this study, we argued the positive aspects of collective empathy. However, too much or too little empathy in the team may decrease critical thinking, increase failure to recognize risk situations, and lead to compassion fatigue, burn out, and emotional exhaustion [24]. Thus, under certain conditions, such as level of task complexity, project duration, team size, and intensity of environmental turbulence and uncertainty, whether collective empathy is detrimental or useful for teamwork can be investigated. In addition, in this study, we argued that collective empathy is a multidimensional construct. However, arguments in the literature suggest that empathy can be viewed as a process [86]. Thus, the interrelations among cognitive, affective, and behavioral empathy can be investigated in a path model analysis. Additionally, how cognitive empathy influences affective empathy, and how affective empathy affects behavioral empathy, can be examined in a longitudinal study. In this study, we used process effectiveness variables solely as the consequences of collective empathy. The relation between collective empathy and collective hope, relational resilience, team competency, and potency development merits investigation. The antecedents of collective empathy can also be enhanced by future research. For example, how team culture, use of information technology (IT), psychological factors (e.g., motivation, attention), and team design factors (e.g., team composition, task allocation, task design), group identification, inter-team communication, organizational stories, and common language influence collective empathy can be investigated. In addition, in this study, we asked to what extent team members were familiar with each other but did not consider whether they had positive or negative emotional experiences in previous projects on which they worked together. The emotional experience experienced by a group becomes part of the group’s particular emotional history, which influences the emotional expression in future group interactions [53]. If group

members have emotional history, involving positive emotions may develop a higher level of collective empathy. Conversely, negative interactions and emotions in the emotional history of the group may lead the group to begin at a lower collective empathy level. Thus, for future studies, the level of emotional history of team members should be included in research models as an antecedent or a moderator variable. In this study, we specifically investigated software development project teams. However, researchers can investigate collective empathy in other types of project teams, such as new product development, new process implementation, new service development, and ephemeral or crisis management. 8. Conclusion Collective empathy is an emergent group property that provides the capability to team members to identify and work with the emotions felt/expressed as a result of group interactions. However, how collective empathy can be enhanced and its effect on project performance is missing from and should be added to the literature. In this study, we tested the role of collective empathy on project process effectiveness and the impact of intimacy-related variables on the development of collective empathy. Our results confirm that collective empathy has a significant effect on the performance of software development projects. Additionally, our results demonstrate that cognitive-based trust, formal communication, and the past familiarity of team members influence the development of collective empathy. Furthermore, our results show that the existence of group norms leverages the role of collective empathy on project process effectiveness. This research just scratches the surface of this important but understudied subject. Future researchers will find inquiry into collective empathy to be rich and fruitful.

Appendix A. Measures * Denotes items dropped after the confirmatory factor analysis. Collective empathy Cognitive empathy CE1: Our team members are able to see things from each other’s points of view. CE3: Our team members understand the feelings of others within our team. CE4: Our team members are able to read the subtle social cues and signals given by others within our team to determine what emotions are being expressed and understand the perspective of the other individuals. CE5: Our team members can predict how someone within our team will feel and what they will do. CE6: Our team members understand what others within our team are thinking. CE7: On our team, we make an effort to understand one another’s attitudes and views. CE2: It is common for our team members to put themselves temporarily in the shoes of other members within our team.* CE8: Our team members accurately sense and read people’s moods, feelings, and nonverbal cues within our team.* Affective empathy AE1: Our team members tend to get emotionally involved with others’ feelings within our team. AE2: The feelings and moods of our team members are influenced by the feelings of the other team members around them. AE3: Members of our team experience the same or other appropriate emotions in response to others’ feelings within our team.

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

AE5: The attitudes of team members are affected by the feelings of others within our team. AE4: Members of our team demonstrate care and concern for others within our team. * Behavioral empathy BE1: In a coordinated manner, our team members demonstrate verbal and non-verbal communication reactions in response to the feelings of others within our team. BE2: Our team members react in response to the feelings of others within our team. BE3: Our team members have behavioral answers to the feelings of others within our team. BE4: Our team members express or show in some communicative way a quality of felt awareness of others’ experiences within our team.* Team learning  Overall, the market perceived that this software product had fewer problems than what is considered normal in the industry.  Post-launch, this software product had far fewer technical problems than our nearest competitor’s product or our own previous products.  Overall, the team did an outstanding job uncovering software product problem areas with which users/customers were dissatisfied.  Overall, the team did an outstanding job correcting software product problem areas with which users/customers were dissatisfied. Speed-to-market  This project was completed in less time than what is considered normal and customary for our industry.  This project was launched on or ahead of the original schedule developed at initial project go-ahead.  Top management was pleased with the time it took us to achieve full commercialization/implementation. Lower development cost  This project came in at or below cost estimates for planning and design.  This project came in at or below cost estimates for implementation. Team member familiarity  We knew the other members of our team (on average) at the time our project team was formed.  We had interactions with the other members of our team (on average) at the time our project team was formed. Formal team communication  Team members conducted frequent formal communications through team meetings with fellow project team members.  Team members conducted frequent formal communications through memos with fellow project team members. Informal team communication  Team members conducted frequent informal communications at the water cooler/coffee maker with fellow project team members.  Team members conducted frequent informal communications at lunch or after work with fellow project team members.

257

Cognitive-based trust  Most of our teammates approach their job with professionalism and dedication.  We see no reason to doubt our teammates’ competence and preparation for the job.  We can rely on our teammates not to make our job more difficult by careless work.  Most of our teammates can be relied upon to do as they are expected to do. Affect-based trust  We can speak freely to our team about difficulties we are having at work and know that our team will want to listen.  If we share our problems with our team, we know team members will respond constructively and caringly. Group norms  There are many social norms (such as customs, etiquette, and working ethics) with which team members are supposed to abide.  In our team there are very clear expectations for how people should act in most situations.  Our team members agree on what behaviors are appropriate vs inappropriate in most situations in our team.*  In our team, if someone acts in an inappropriate way, others will strongly disapprove.*  People in our team almost always comply with social norms.*

References [1] L.S. Aiken, S.G. West, R.R. Reno, Multiple Regression: Testing and Interpreting Interactions, Sage Publications, Newbury Park, 1991. [2] A.E. Akgu¨n, G.S. Lynn, C. Yılmaz, Learning process in new product development teams and effects on product success: a socio-cognitive perspective, Ind. Mark. Manag. 35, 2006, pp. 210–224. [3] A.E. Akgu¨n, H. Keskin, J. Byrne, S.Z. Imamoglu, Antecedents and consequences of team potency in software development projects, Inf. Manag. 44 (7), 2007, pp. 646–656. [4] A.E. Akgu¨n, H. Keskin, J.C. Byrne, A. Gu¨nsel, Antecedents and results of emotional capability in software development project teams, J. Prod. Innov. Manag. 28 (6), 2011, pp. 957–973. [5] D.G. Ancona, H. Bresman, K. Kaeufer, The comparative advantage of X-teams, MIT Sloan Manag. Rev. 43 (3), 2002, pp. 33–39. [6] J.C. Anderson, D.W. Gerbing, Structural equation modeling in practice: a review and recommended two-step approach, Psychol. Bull. 103, 1988, pp. 411–423. [7] B.E. Ashforth, R.H. Humphrey, Emotion in the workplace: a reappraisal, Hum. Relat. 48, 1995, pp. 97–125. [8] Y.A. Au, D. Carpenter, X. Chen, J.G. Clark, Virtual organizational learning in open source software development projects, Inf. Manag. 46, 2009, pp. 9–15. [9] R.P. Bagozzi, D.J. Moore, Public service advertisements: emotions and empathy guide, J. Mark. 58, 1994, pp. 56–71. [10] R. Bagozzi, Y. Yi, L.W. Phillips, Assessing construct validity in organizational research, Adm. Sci. Q. 36, 1991, pp. 421–458. [11] R.P. Bagozzi, W. Verbeke, J.C. Gavino, Culture moderates the self-regulation of shame and its effects on performance: the case of salespersons in the Netherlands and the Philippines, J. Appl. Psychol. 88 (2), 2003, pp. 219–233. [12] A. Bandura, Social Foundations of Thought and Action: A Social Cognitive Theory, Prentice-Hall, Englewood Cliffs, NJ, 1986. [13] R. Baron, D.A. Kenny, The moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations, J. Pers. Soc. Psychol. 51, 1986, pp. 1173–1182. [14] S.G. Barsade, The ripple effect: emotional contagion and its influence on group behavior, Adm. Sci. Q. 47 (4), 2002, pp. 644–675. [15] S.G. Barsade, A.J. Ward, J.D.F. Turner, J.A. Sonnenfeld, To your heart’s content: a model of affective diversity in top management teams, Adm. Sci. Q. 45 (4), 2000, pp. 802–837. [16] C.A. Bartel, R. Saavedra, The collective construction of work group moods, Adm. Sci. Q. 45 (2), 2000, pp. 197–234. [17] C.D. Batson, N.Y. Ahmad, Using empathy to improve intergroup attitudes and relations, Soc. Issues Policy Rev. 3, 2009, pp. 141–177.

258

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259

[18] S. Bolton, Emotion Management in the Workplace, Palgrave, Hampshire, 2005. [19] .J. Bowlby, Attachment and loss, Attachment, 1st ed., Basic Books, New York, 1982 [20] R.E. Boyatzis, How and why individuals are able to develop emotional intelligence, in: C. Cherniss, D. Goleman (Eds.), The Emotionally Intelligent Workplace: How to Select For, Measure, and Improve Emotional Intelligence in Individuals, Groups, and Organizations, Jossey-Bass, San Francisco, 2001, pp. 234–253. [21] E. Carmel, B.J. Bird, Small is beautiful: a study of packaged software development teams, J. High Technol. Manag. Res. 1, 1997, pp. 129–148. [22] W.W. Chin, The partial least squares approach for structural equation modeling, in: G.A. Marcoulides (Ed.), Modern Methods for Business Research, Lawrence Erlbaum Associates, New Jersey, 1998, pp. 295–336. [23] F. de Vignemont, T. Singer, The empathic brain: how, when and why? Trends Cogn. Sci. 10 (10), 2006, pp. 435–441. [24] J. Decety, M. Svetlova, Putting together phylogenetic and ontogenetic perspectives on empathy, Dev. Cognit. Neurosci. 2, 2012, pp. 1–24. [25] L.A. DeChurch, M.A. Marks, Maximizing the benefits of task conflict: the role of conflict management, Int. J. Conf. Manag. 12, 2001, pp. 4–22. [26] J. Dhaliwal, C.G. Onita, R. Poston, X.P. Zhang, Alignment within the software development unit: assessing structural and relational dimensions between developers and testers, J. Strateg. Inf. Syst. 20 (4), 2011, pp. 323–342. [27] V.U. Druskat, S.B. Wolff, in: C. Cherniss, D. Goleman (Eds.), Group Emotional Intelligence and Its Influence on Group Effectiveness, The Emotionally Intelligent Workplace, Jossey-Bass, San Francisco, 2001, pp. 132–155. [28] E.W. Duggan, C.S. Thachenkary, Integrating nominal group technique and joint application development for improved systems requirements determination, Inf. Manag. 41, 2004, pp. 399–411. [29] A.C. Edmondson, Psychological safety and learning behavior in work teams, Adm. Sci. Q. 44 (2), 1999, pp. 350–383. [30] H.A. Elfenbein, J.T. Polzer, N. Ambady, Team emotion recognition accuracy and team performance, in: C.E.J. Ha¨rtel, N.M. Ashkanasy, W.J. Zerbe (Eds.), Research on Emotion in Organizations: Functionality, Intentionality and Morality, Lawrence Erlbaum Associates, Inc., New Jersey, 2007, pp. 87–119. [31] E.M. Engelen, B. Ro¨ttger-Ro¨ssler, Current disciplinary and interdisciplinary debates on empathy, Emot. Rev. 4, 2012, pp. 3–8. [32] J.A. Espinosa, S.A. Slaughter, R.E. Kraut, J.D. Herbsleb, Familiarity, complexity, and team performance in geographically distributed software development, Org. Sci. 18 (4), 2007, pp. 613–630. [33] H. Fallon, The joint implementation process: a workshop technique for coordinating the implementation of information systems, J. Syst. Manag. 46 (6), 1995, pp. 18–23. [34] M.J. Fay, Informal communication of co-workers: a thematic analysis of messages, Qual. Res. Org. Manag. 6 (3), 2011, pp. 212–229. [35] S. Fineman, A. Sturdy, The emotions of control: a qualitative study of environmental regulation, Hum. Relat. 52 (5), 1999, pp. 631–663. [36] C. Fornell, D.F. Larcker, Evaluating structural equation models with unobservable variables and measurement error, J. Mark. Res. 18, 1981, pp. 39–51. [37] M.J. Gelfand, M. Erez, Z. Aycan, Crosscultural organizational behavior, Ann. Rev. Psychol. 58, 2007, pp. 479–514. [38] D. Goleman, in: C. Cherniss, D. Goleman (Eds.), Emotional Intelligence: Issues in Paradigm Building, The Emotionally Intelligent Workplace, Jossey-Bass, San Francisco, 2001, pp. 13–26. [39] D. Goleman, R. Boyatzis, A. McKee, The emotional reality of teams, Glob. Bus. Org. Excell. 21, 2002, pp. 55–65. [40] A. Griffin, A.L. Page, PDMA success measurement project: recommended measures for product development success and failure, J. Prod. Innovat. Manag. 13, 1996, pp. 478–496. [41] J.J. Gross, R.W. Levenson, Hiding feelings: the acute effects of inhibiting negative and positive emotion, J. Abnorm. Psychol. 106, 1997, pp. 95–103. [42] D.H. Gruenfeld, E.A. Mannix, K.Y. Williams, M.A. Neale, Group composition and decision making how member familiarity and information distribution affect process and performance, Organ. Behav. Hum. Dec. 67 (1), 1996, pp. 1–15. [43] J.H. Hair, W.C. Black, B.J. Babin, R.E. Anderson, R.L. Tatham, Multivariate Data Analysis, 6th ed., Pearson, New Jersey, 2006. [44] J. He, B.S. Butler, W.R. King, Team cognition: development and evolution in software project teams, J. Manag. Inf. Syst. 24 (2), 2007, pp. 261–292. [45] A. Hochschild, The Managed Heart, University of California, Berkeley, 1983. [46] M. Hoegl, H.G. Gemuenden, Teamwork quality and the success of innovative projects: a theoretical concept and empirical evidence, Org. Sci. 12 (4), 2001, pp. 435–449. [47] R.F. Hurley, G.T.M. Hult, Innovation, market orientation, and organizational learning: an integration and empirical examination, J. Mark. 62, 1998, pp. 42–54. [48] Q.H. Huy, Emotional capability, emotional intelligence, and radical change, Acad. Manag. Rev. 24, 1999, pp. 325–345. [49] P. Kanawattanachai, Y. Yoo, Dynamic nature of trust in virtual teams, Strateg. Inf. Syst. 11, 2002, pp. 187–213. [50] J.M. Kanov, S. Maitlis, M.C. Worline, J.E. Dutton, P.J. Frost, J.M. Lilius, Compassion in organizational life, Am. Behav. Sci. 7 (6), 2004, pp. 808–827. [51] M. Keil, H.K. Lee, T. Deng, Understanding the most critical skills for managing IT projects: a Delphi study of IT project managers, Inf. Manag. 50 (7), 2013, pp. 398–414. [52] J.B. Kellett, R.H. Humphrey, R.G. Sleeth, Empathy and complex task performance: two routes to leadership, Leadersh. Q. 13 (5), 2002, pp. 523–544. [53] J.R. Kelly, S.G. Barsade, Mood and emotions in small groups and work teams, Organ. Behav. Hum. Dec. 86 (1), 2001, pp. 99–130.

[54] E.H. Kessler, A.K. Chakrabarti, Innovation speed: a conceptual model of context, antecedents, and outcomes, Acad. Manag. Rev. 21, 1996, pp. 1143–1191. [55] M. Kodama, Strategic business applications and new virtual knowledge-based business through community-based information network, Inf. Manag. Comput. Secur. 7 (4), 1999, pp. 186–199. [56] A. Koskina, D. Keithley, Emotion in a call centre SME: a case study of positive emotion management, Eur. Manag. J. 28 (3), 2010, pp. 208–219. [57] R. Lagerstro¨m, L.M. Wu¨rtemberg, H. Holm, O. Luczak, Identifying factors affecting software development cost and productivity, Softw. Qual. J. 20, 2012, pp. 395–417. [58] S. Lee, A.J. Dubinsky, Influence of salesperson characteristics and customer emotion on retail dyadic relationships, Int. Rev. Retail Distrib. Consum. Res. 13, 2003, pp. 23–36. [59] R.J. Lewicki, B.B. Bunker, Developing and maintaining trust in work relationships, in: R. Kramer, T. Tyler (Eds.), Trust in Organizations: Frontiers of Theory and Research, Sage Publication, Thousand Oaks, 1996, pp. 114–139. [60] K. Lewis, Knowledge and performance in knowledge-worker teams: a longitudinal study of transactive memory systems, Manag. Sci. 50 (11), 2004, pp. 1519–1533. [61] G.S. Lynn, A.E. Akgu¨n, Launch your new products/services better, faster, Res. Technol. Manag. 46 (3), 2003, pp. 21–26. [62] D.M. Mackie, T. Devos, E.R. Smith, Intergroup emotions: explaining offensive action tendencies in an intergroup context, J. Pers. Soc. Psychol. 79 (4), 2000, pp. 602–616. [63] B. Martinovski, D. Traum, S. Marsella, Rejection of empathy in negotiation, Group Decis. Negot. 16, 2007, pp. 61–76. [64] D.J. McAllister, Affect- and cognition-based trust as foundations for interpersonal cooperation in organizations, Acad. Manag. Rev. 38, 1995, pp. 24–59. [65] D.A. McBane, Empathy and the salesperson: a multidimensional perspective, Psychol. Mark. 12, 1995, pp. 349–370. [66] M.E. McCullough, K.C. Rachal, S.J. Sandage, E.L. Worthington, S. Wade-Brown Jr., T. Hight, Interpersonal forgiving in close relationships II: theoretical elaboration and measurement, J. Pers. Soc. Psychol. 75, 1998, pp. 1586–1603. [67] P.F. Mckay, D.R. Avery, M.A. Morris, Mean racial–ethnic differences in employee sales performance: the moderating role of diversity climate, Pers. Psychol. 61 (2), 2008, pp. 349–374. [68] J. Mencl, D.R. May, The effects of proximity and empathy on ethical decisionmaking: an exploratory investigation, J. Bus. Ethics 85, 2009, pp. 201–226. [69] M.H. Meyer, K.F. Curley, The impact of knowledge and technology complexity on information systems development, Expert Syst. Appl. 8 (1), 1995, pp. 111–134. [70] M. Mikulincer, P.R. Shaver, Attachment theory and emotions in close relationships, Pers. Relatsh. 12 (2), 2005, pp. 149–168. [71] C. Moorman, A. Miner, The role of organizational memory in new product performance and creativity, J. Mark. Res. 34, 1997, pp. 91–106. [72] A. Mukherjee, N. Malhotra, Does role clarity explain employee-perceived service quality? A study of antecedents and consequences in call centers Int. J. Serv. Ind. Manag. 17 (5), 2006, pp. 444–473. [73] A.R. Muller, M.D. Pfarrer, L.M. Little, A theory of collective empathy in corporate philanthropy decisions, Acad. Manag. Rev. 39 (1), 2014, pp. 1–21. [74] D. Mumby, L. Putnam, The politics of emotion, Acad. Manag. Rev. 17, 1992, pp. 465–486. [75] B. Nicholson, S. Sahay, Embedded knowledge and offshore software development, Inf. Org. 14 (4), 2004, pp. 329–365. [76] S. Nidhra, M. Yanamadala, W. Afzal, R. Torkar, Knowledge transfer challenges and mitigation strategies in global software development – a systematic literature review and industrial validation, Int. J. Inform. Manag. 33 (2), 2013, pp. 333–355. [77] L. Ozdamar, E. Alanya, Uncertainty modelling in software development projects (with case study), Ann. Oper. Res. 102 (1), 2001, pp. 157–178. [78] R. Patnayakuni, C. Ruppel, A. Rai, Managing the complementarity of knowledge integration and process formalization for systems development performance, J. Assoc. Inf. Syst. 7 (8), 2006, pp. 545–567. [79] P. Pinjani, P. Palvia, Trust and knowledge sharing in diverse global virtual teams, Inf. Manag. 50 (4), 2013, pp. 144–153. [80] P.M. Podsakoff, D.W. Organ, Self-reports in organizational research: problems and prospects, J. Manag. 12 (4), 1986, pp. 531–544. [81] P.M. Podsakoff, S.B. MacKenzie, J.Y. Lee, N.P. Podsakoff, Common method bias in behavioral research: a critical review of the literature and recommended remedies, J. Appl. Psychol. 88, 2003, pp. 879–903. [82] J.D. Procaccino, J.M. Verner, S.J. Lorenzet, Defining and contributing to software development success, Commun. ACM 49 (8), 2006, pp. 79–84. [83] T.H. Reus, Y. Liu, Rhyme and reason. Emotional capability and the performance of knowledge-intensive work groups, Hum. Perform. 17 (2), 2004, pp. 245–266. [84] B. Rime´, The social sharing of emotion as an interface between individual and collective processes in the construction of emotional climates, J. Soc. Issues 63 (2), 2007, pp. 307–322. [85] C.M. Ringle, S. Wende, A. Will, SmartPLS 2.0 (M3) Beta, 2005 Hamburg, http:// www.smartpls.de. [86] M. Roberge, A multi-level conceptualization of empathy to explain how diversity increases group performance, Int. J. Bus. Manag. 8 (3), 2013, pp. 122–135. [87] L.R. Rosh, R. Offermann, V. Diest, Too close for comfort? Distinguishing between team intimacy and team cohesion Hum. Resour. Manag. Rev. 22 (2), 2012, pp. 116–127. [88] O. Sakka, H. Barki, L. Coˆte´, Interactive and diagnostic uses of management control systems in IS projects: antecedents and their impact on performance, Inf. Manag. 50 (6), 2013, pp. 265–274. [89] M.J. Schmit, A.M. Ryan, The big five in personnel selection: factor structure in applicant and non-applicant populations, J. Appl. Psychol. 78, 1993, pp. 966–974.

A.E. Akgu¨n et al. / Information & Management 52 (2015) 247–259 [90] E. Seppala, T. Rossomando, J.R. Doty, Social connection and compassion: important predictors of health and well-being, Soc. Res. 80 (2), 2013, pp. 411–430. [91] T. Singer, C. Lamm, The social neuroscience of empathy, Ann. N. Y. Acad. Sci. 1156, 2009, pp. 81–96. [92] E.R. Smith, C.R. Seger, D.M. Mackie, Can emotions be truly group level? Evidence regarding four conceptual criteria J. Pers. Soc. Psychol. 93 (3), 2007, pp. 431–446. [93] J.R. Spoor, J.R. Kelly, The evolutionary significance of affect in groups: communication and group bonding, Group Process. Interg. Relat. 7, 2004, pp. 398– 412. [94] D. Tjosvold, F. Su, Managing anger and annoyance in organizations in china: the role of constructive controversy, Group Org. Manag. 32 (3), 2007, pp. 260–289. [95] E.M. Trauth, L.M. Jessup, Understanding computer-mediated discussions: positivist and interpretive analyses of group support system use, MIS Q. 24 (1), 2000, pp. 43–79. [96] J.R. Turner, R. Mu¨ller, Communication and co-operation on projects between the project owner as principal and the project manager as agent, Eur. Manag. J. 2 (3), 2004, pp. 327–336. [97] G.A. Van Kleef, C.K. De Dreu, D. Pietroni, A.S. Manstead, Power and emotion in negotiation: power moderates the interpersonal effects of anger and happiness on concession making, Eur. J. Soc. Psychol. 36, 2006, pp. 557–581. [98] S. Vangen, C. Huxham, Nurturing collaborative relations, building trust in inter-organizational collaboration, J. Appl. Behav. Sci. 39 (1), 2003, pp. 5–31. [99] H.M. Weiss, R. Cropanzano, Affective events theory: a theoretical discussion of the structure, causes and consequences of affective experiences at work, Res. Org. Behav. 18, 1996, pp. 1–74. [100] X. Zhang, T.F. Stafford, J.S. Dhaliwal, M.L. Gillenson, G. Moeller, Sources of conflict between developers and testers in software development, Inf. Manag. 51 (1), 2014, pp. 13–26. ¨ n is a professor of science and technology Ali E. Akgu studies in the School of Business Administration at Gebze Institute of Technology, Turkey. He received his PhD in technology management from Stevens Institute of Technology and his MS in Engineering Management from Drexel University. His research areas are new product/technology development, organizational learning, and cognitive and social psychology in innovation management.

259

Halit Keskin is a professor of science and technology studies in the School of Business Administration at Gebze Institute of Technology, Turkey. He received his PhD in management and organization from Gebze Institute of Technology. His research interests include technology and innovation management, knowledge management, and human resource management in high-tech firms.

A. Yavuz Cebecioglu is a PhD student at the Department of Business Administration at Gebze Institute of Technology, Turkey. His research focuses on software development and management.

Derya Dogan is a PhD student at the Department of Business Administration at Gebze Institute of Technology, Turkey. Her research focuses on social networks and new product/technology development.