Team Climate, Team Cognition, Team Intuition, and Software Quality ...

Group Decis Negot (2014) 23:1145–1176 DOI 10.1007/s10726-013-9367-1

Team Climate, Team Cognition, Team Intuition, and Software Quality: The Moderating Role of Project Complexity Atif Açıkgöz · Ay¸se Günsel · Nizamettin Bayyurt · Cemil Kuzey

Published online: 1 August 2013 © Springer Science+Business Media Dordrecht 2013

Abstract Teams represent a prevailing approach to getting work done in today’s hypercompetitive business environment. Although there is a widely held assumption that team-related capabilities determine the success of new product development projects, empirical research on team capabilities is scant. Based on the resourcebased view of the firm, organizational learning theory, and situated learning theory, this study investigates the interrelationships among team climate, two informationprocessing capabilities (i.e., team cognition and team intuition), and software quality. As well, this study explores the moderating effect of project complexity between the information-processing capabilities and the quality of the software. In studying the data from 139 software development projects using the partial least squares structural equation modeling methods, we found that team climate has a direct influence on team cognition. Moreover, the findings showed that team cognition was positively related to the quality of the software product in general; in particular, this relationship was found to be far more significant when project complexity was used as a moderator.

This work is supported by the Scientific Research Fund of Fatih University under the project number P54081202_B. A. Açıkgöz (B)· N. Bayyurt · C. Kuzey Department of Management, Fatih University, 34500 Büyükçekmece, Istanbul, Turkey e-mail: [email protected] N. Bayyurt e-mail: [email protected] C. Kuzey e-mail: [email protected] A. Günsel Kocaeli University, 41380 Umuttepe, Kocaeli, Turkey e-mail: [email protected]

123

1146

A. Açıkgöz et al.

This finding indicates that the software development team’s ability to process information logically in order to interpret situations effectively allows the team to launch superior software products when unexpected and undesirable events make a project complicated and challenging to perform. In particular, managers should encourage teams to benefit from new ideas and make collective efforts for reaching goals. Managers should also enable teams to specialize in their tasks and improve their collective information-processing capabilities. Keywords Team climate · Team cognition · Team intuition · Software quality · Project complexity · Situated learning theory

1 Introduction The first decade of the twenty-first century stands witness to a dramatic change of the business landscape at an astonishing rate (Günsel and Açıkgöz 2013). In the hypercompetitive business environment, existing technologies easily became obsolete due to the diversity of products as well as rapid changes in customer preferences (D’aveni 1994; Günsel and Açıkgöz 2013; Zander and Kogut 1995). According to Drach-Zahavy (2004), teams have become the common units for managing hypercompetition, especially in knowledge-intensive industries. The underlying reason is that teams can more easily manage stress, more rapidly adapt to new situations, make better decisions, and be more productive than individuals (Cooke et al. 2004). In particular, high-technology firms have begun to prefer using teams in order to maximize the utilization of know-how, minimize the influence of workload, take full advantage of sophisticated technologies, and reach higher levels of organizational learning (He et al. 2007). In this regard, new product development (NPD) teams (e.g., software development teams) are extremely important for producing today’s high-technology products (Cooke et al. 2004; Drach-Zahavy 2004; He et al. 2007). Due to the vast amount and diversity of the available information, rational individuals with their limited individual abilities are not capable of processing information adequately in order to develop high quality products. Instead, team-level capabilities are necessary. In this study, we used both team cognition and team intuition as information-processing capabilities. In the organizational learning literature, teamlevel studies abound, providing evidence of the importance of team cognition and team intuition (e.g., Akgün et al. 2008; Dayan and Benedetto 2011; Dayan and Elbanna 2011; Hodgkinson et al. 2008; Klimoski and Mohammed 1994; Mohammed et al. 2010; Sadler-Smith and Shefy 2004). This stream of literature argues that teams routinely process information based on both their cognitions and intuitions (Dane and Pratt 2007; Sayegh et al. 2004). At the team level, cognition refers to a team’s shared and organized understanding of its relevant environment (Klimoski and Mohammed 1994). A high level of cognition allows the team to learn more quickly, amass larger amounts of information, reorganize information for efficient utilization, and appeal to their own logical knowledge when choosing suitable actions (He et al. 2007; Resick et al. 2010). The main benefit of using team cognition is to create organized mental images or representations of the

123

The Moderating Role of Project Complexity

1147

crucial elements within the environment (Klimoski and Mohammed 1994; Mohammed et al. 2010). In addition to team cognition, team intuition has also started to attract attention from various scholars (e.g., Dayan and Benedetto 2011; Dayan and Elbanna 2011), perhaps because of their dissatisfaction with rationality and its limits. In this sense, scholars are now arguing that many information-processing actions take place automatically outside of team consciousness, in the realm of intuition (Sadler-Smith and Shefy 2004). More specifically, NPD teams may rely more on intuition, because they engage more upon less formal team activities than other types of team activity (Dayan and Elbanna 2011). For instance, NPD teams operate in complex projects and need to respond to unpredictable challenges in the light of the hypercompetitive business environment. Organizations potentially contain different contexts which have the potential to play a central role in shaping teams’ perceptions, cognitions, and behaviors (Ashforth 1985; Maruping and Magni 2012; Mohammed et al. 2010). In this study, we chose to use team climate—that of the team members’ collective perceptions of task-related behaviors, practices, and procedures—because it may facilitate or restrain the teams’ information-processing potentials (Maruping and Magni 2012; Zellmer-Bruhn and Gibson 2006). For example, team members integrate their idiosyncratic knowledge through sharing and mutual experiences in a supportive climate (Bertels et al. 2011). According to this, Antoni and Hertel (2005) claimed that team climate explained more than half of the variance of NPD teams’ innovative behaviors. Based on these aspects, it is worthwhile to analyze the role of team climate on both team cognition and team intuition. The resource-based view (RBV) of the firm and organizational learning theory (OLT) depict organizations as repositories of resources and capabilities that form a basis for sustainable competitive advantage (Barney 1991; Huang and Li 2012; Zander and Kogut 1995). In the RBV literature, significant attention is given to explicit and tacit knowledge (Nonaka 1994; Dosi et al. 2008). Explicit knowledge can be readily explained and codified, whereas tacit knowledge is difficult to codify and articulate (Nonaka 1994). RBV also stresses the importance of the development of team-level capabilities which are valuable, rare, inimitable, and non-substitutable (Barney 1991). According to Grant (1996), capabilities are developed through the exchange of both explicit and tacit knowledge within special work groups (e.g., NPD teams). In this regard, RBV tries to provide an explanation of competitive heterogeneity based on the premise that close competitors differ in their resources and capabilities (Barney 1991; Helfat and Peteraf 2003; Hodgkinson et al. 2008; Mohammed et al. 2010). There is a growing body of literature reflecting the importance of organizational learning (Senge 1990; Hurley and Hult 1998; Škerlavaj et al. 2007; García-Morales et al. 2012)—the process of improving organizational functions and activities (e.g., new product development) as a result of knowledge creation through information processing (Edmondson and Nembhard 2009; Huang and Li 2012; Uhlenbruck et al. 2003). The literature widely emphasizes that organizational learning depends on both individual and team learning (Huang and Li 2012). Team members integrate their distinctive knowledge and experiences in order to shape a team’s knowledge base and routines that are institutionalized by the organization (Yang and Chen 2007). According to this perspective, team learning is an interface between individual and

123

1148


organizational learning (Yang and Chen 2005). In this sense, team learning, as a way of cooperatively processing information, enables NPD teams to modify their behaviors by reflecting new knowledge and insights (Blazevic and Lievens 2004; Nonaka 1994; Sinkula et al. 1997). Situated learning theory (SLT) (Lave and Wenger 1991) suggests that the learning process depends upon the context in which the actors find themselves (Tyre and Hippel 1997). This theory requires teams to be situated within a specific context, because learning is a social process that is affected by social conditions and cultural values (Bertels et al. 2011). The concept of “legitimate peripheral participation”, which strongly influences what teams learn through the internalization of knowledge and skills, is closely related to context (Bechky 2003; Bertels et al. 2011; Lave and Wenger 1991). According to the theory, knowledge is not absolute, but rather can only be defined with respect to a specific context (Tyre and Hippel 1997). SLT does not consider the team memory as a container waiting to be filled up; instead, it appraises memory-inaction, which simultaneously generates knowledge and learning in interaction with the physical and psychological environment (Fox 1997). This study empirically investigates the effects of team climate on both team cognition and team intuition in software development projects. To date, we are unaware of any study of the effects of team climate on team cognition and team intuition in the relevant literature. More precisely, the study is guided by the following research questions: (i) how does team climate influence team cognition and team intuition, (ii) how does team cognition and team intuition affect software quality, and (iii) to what extent are the effects of team cognition and team intuition on software quality moderated by project complexity? By providing answers to these questions, the study contributes to our theoretical and practical understanding of team cognition and team intuition in the context of software development projects. Theoretically, this study contributes to RBV and OLT in general, SLT in particular. This study contributes to RBV by examining explicit and tacit knowledge through the development and utilization of the team-level capabilities in software development projects. This study also makes a contribution to OLT by investigating team learning through the information-processing capabilities at team level in software development projects. Lastly, the study makes a mounting contribution to SLT by analyzing team climate as an antecedent of both team cognition and team intuition. From a managerial point of view, teams are plunged into numerous activities during the NPD projects and they need to use their cognitions and intuitions to make timely and effective decisions in respond to emerging conditions. In this context, the project leaders/managers stress the importance of team climate as a workplace atmosphere to facilitate or hinder the usage of team cognition and team intuition. Team climate enhances professionals’ understanding on the roles of information-processing capabilities (e.g., team cognition and team intuition) and then project outputs, such as software quality. The project leaders/managers try to find out some kind of practices, norms, techniques, and procedures in each stage of the NPD process that enable the team to take the benefit of their cognition and intuition, and ultimately allow the team to develop new products of quality. The remainder of the article is organized as follows. Section 2 provides a review of the relevant literature in order to establish a clear theoretical ground. Section 3 describes the specific hypotheses for the research model. Section 4 presents the empir-

123


1149

ical results in order to tests these assumptions. In Sect. 5, a discussion of the necessary implications related to managerial and theoretical aspects are provided, and Sect. 6 is the conclusion. 2 Theoretical Backgrounds The following literature review is intended to provide a theoretical base for a holistic model of team climate, team cognition, team intuition, and software quality. In order to do so, this study focuses on two information-processing capabilities—team cognition and team intuition respectively. The review begins with an overview of the concept of team cognition as a team level capability; furthermore the functions of team cognition within NPD projects are discussed. This is followed by a review of team intuition from the NPD perspective, providing more details about the definition of borders as well as the interconnectedness of team cognition and team intuition. Thus, through this theoretical review, these two complementary capabilities are deeply examined in order to: (i) reveal the nature of information processing in relation to the NPD practices, and (ii) understand how team cognition and team intuition should be employed within the NPD context. 2.1 Team Cognition Through norms and close interaction, teams can generate the mental activities of remembering, storing, utilizing, and sharing information much like an individual (Klimoski and Mohammed 1994; Park et al. 2008). In this sense Cooke et al. (2004) claim that teams think and perform rationally. Rationality is based on a belief in a specific cause-and-effect relationship in the physical and social world. Team cognition emerges from the interplay of the individual rationality of each team member and NPD process behaviors (Cooke et al. 2004). Important terms in this context are defined as follows. Team cognition is the capability to perform a learning process related to the rational acquisition, processing, and dissemination of information for the purpose of creating team-level intellectualness (Akgün et al. 2008; Klimoski and Mohammed 1994; Park et al. 2008). Information acquisition refers to information gathering from internal and external environment to develop mental models (Resick et al. 2010). Information processing is the use of existing and acquired information to generate creative ideas during a project (Moorman 1995). Finally, information dissemination is the sharing of information during a project by formal and informal means (Huber 1991). Operationally, team cognition allows NPD teams to generate explanations of system functioning, predictions of estimated situations, and identification of new opportunities (Mathieu et al. 2000; West 2007). The comprehensive aim of team cognition is to generate distinctive knowledge and in turn, learning through the information-processing capability for accomplishing goals (He et al. 2007). Systematic information processing is at the core of team cognition which starts with the acquisition of new information from various resources (e.g., customers, competitors, suppliers, databases, and the Internet) (Mohammed et al. 2010; West 2007). After this, new information is shared

123

1150


among the team members leading the mutual expectations of future events (Mathieu et al. 2000). As a result, NPD teams quickly build a shared vision or organized mental images of how they will perform their tasks (Klimoski and Mohammed 1994; Mohammed et al. 2010). In other words, team cognition is related to what is happening, why it is happening, and what is likely to happen next (Mohammed et al. 2010). According to Park et al. (2008), team cognition refers to the desire to enlarge teams’ knowledge structures. Enlarged knowledge structures enable NPD teams to form accurate explanations and expectations for a specific task in order to coordinate their actions and adapt their behaviors to the demands of a turbulent environment (Cannon-Bowers and Salas 2001; Levesque et al. 2001; Mohammed et al. 2010). It also allows NPD teams to be more effective in solving problems, to make better decisions, and to face fewer crises during a project (Park et al. 2008). It is assumed by some scholars (e.g., Klimoski and Mohammed 1994; Mohammed et al. 2010) that team cognition enhances the quality of teamwork and team effectiveness.

2.2 Team Intuition Most previous studies on the information-processing capability of teams have been conducted a bit further on the basis of team cognition (Akgün et al. 2007, 2008; Lynn et al. 1999). However, detecting meaningful cause-and-effect relationships and generating mental models of the relevant environment require intuition as well as cognition, because of the rapid changes in customers’ preferences, short product life cycle, and quick depreciation of know-how in the hypercompetitive marketplace (Dayan and Benedetto 2011). In other words, team cognition has some limitations because of unavailable, incomplete, or overwhelming information in complex, uncertain and time-pressured conditions. For example, NPD teams may not be able to take the benefit of their cognition efficiently due to asymmetric information so that using intuition would be a better alternative for eliminating this disadvantage. In this regard, some scholars (e.g., Dane and Pratt 2007; Dayan and Elbanna 2011; Hodgkinson et al. 2009; Sadler-Smith and Shefy 2004) claim that an effective information-processing capability is not only cognitive but also intuitive at the team level. According to them, the significance of team intuition as an information-processing capability should not be underestimated. The roots of the term “intuition” may be traced to the Latin word intueor or intueri, which can be translated as “looking, regarding or knowing from within” (SadlerSmith and Shefy 2004; Hodgkinson et al. 2008). Intuition may refer to perception without recourse to rational methods, inexplicable comprehension, and the automatic processing of accumulated information without conscious awareness (Dane and Pratt 2007; Hodgkinson et al. 2008; Sadler-Smith and Shefy 2004; Sauter 1999). Intuitive action is characterized by comprehensive thinking without special effort (Betsch and Glöckner 2010). Intuition is hard to verbalize but easy to recognize (Sadler-Smith and Shefy 2004) because it is a subjective response that emerges through non-conscious holistic associations founded on largely tacit ways of knowing (Hodgkinson et al. 2009). Attempts at verbalizing intuition can restrict access to these more tacit ways of

123


1151

knowing; instead of verbalizing, intuition may manifest itself as an image or narrative (Sadler-Smith and Shefy 2004). Intuition as a team-level capability is defined by Sauter (1999) as “holistic thinking, immediate insight, seeing the answer without knowing how it was reached” (p. 110). Based on Sadler-Smith and Shefy’s (2004) comprehensive definition, we consider team intuition as a capability to process information automatically for attaining direct knowledge or understanding without the obvious intrusion of rational thought. Teams use this capability to generate a holistic image in checking on rational analysis. For example, most successful NPD teams frequently call upon their intuitions to deepen their reflections from information such as charts, graphs, records, and spreadsheets (Dayan and Elbanna 2011). In this sense, intuitive capability enables NPD teams to comprehend problems at once and recognize the suitable behavior patterns to follow in response to those problems long before they are able to formulate their reasoning as to why those behavior patterns are suitable (Hodgkinson et al. 2009). When team intuition is used in this way, it can be experienced as a hunch, gut feeling, or sixth sense (Dayan and Benedetto 2011; Hodgkinson et al. 2008). Dayan and Benedetto (2011) have enumerated the reasons for the use of intuition as the most suitable team-level capability in turbulent conditions as follows: (i) availability of little previous precedent considering emerging trends, (ii) complexity of situations, and (iii) asymmetric information. More clearly, existing precedents in terms of structure, norms, and culture embedded in organizations may no longer to be effective to enable NPD teams to use only cognitive information-processing capability (Moorman and Miner 1997). NPD teams often need to use the intuition spontaneously in turbulent conditions, because the volume and complexity of existing information in such conditions has the potential to be overwhelming (Dane and Pratt 2007; SadlerSmith and Shefy 2004). NPD teams feel a greater need to exhibit intuitive behavior if the conditions they are working on are not analyzable due to incomplete or poor data (Dayan and Elbanna 2011). Following the above arguments, it can be claimed that intuitive synthesis allows NPD teams to evaluate complex situations and deal with asymmetric information (Patton 2003). In contrast to team cognition, which only allows teams to process information stepby-step, team intuition allows them to process multiple pieces of information in parallel (Betsch and Glöckner 2010). NPD teams need to use parallel information processing, particularly in highly complex conditions where there may be an abundance of relevant information (Dayan and Benedetto 2011). Similarly, intuitive capability enables NPD teams to encode limited information in an extensive mode (e.g., tacit knowledge) irrespective of its origin (Betsch and Glöckner 2010). This knowledge may take the form of narratives and can be stored as rules for how to achieve the team’s aims in specific situations (Sadler-Smith and Shefy 2004). The main benefit of team intuition is to accelerate information processing in order to generate tacit knowledge, solve problems creatively, make better decisions, and improve project outcomes (Dayan and Benedetto 2011). Team cognition and team intuition are not unbridgeable, but it is challenging to weave the two together (i.e., combine intuition with cognition) in order to process information efficiently within the context of dual-process theories (Evans 2003; Hodgkinson et al. 2008, 2009; Sadler-Smith and Shefy 2004). In this direction, it is increasingly

123

1152


claimed that team cognition and team intuition complete with each other. According to Sadler-Smith and Shefy (2004), for example, teams may attain a more balanced view by considering both cognition and intuition as complementary enhancing components of the information-processing capability. Accordingly, team intuition is neither the opposite of team cognition nor a random process of forecasting; rather, team intuition corresponds to thoughts, conclusions and choices generated considerably through automatic mental processes (Hodgkinson et al. 2009) in relatively turbulent conditions while they operate in a rule-based, analytic, and controlled manner in the long run (Evans 2003). In a way, team cognition and team intuition are the two sides of the same coin, completing each other. Based upon the reasons mentioned above, we have included team intuition in addition to team cognition into our research model.

3 Hypotheses Development 3.1 Team Climate and Team Cognition Any attempt to research NPD teams’ capabilities without considering the role of organizational context such as team climate will fail (Aladwani 2002; Maruping and Magni 2012; Mohammed et al. 2010). Team climate can be defined as the combination of norms, attitudes, and expectations that team members perceive in order to function in a particular context (Pirola-Merlo 2010). In the organizational behavior literature, team climate has been categorized into dimensions because of its complicated and multifaceted nature (e.g., Amabile and Gryskiewicz 1989; Ekvall et al. 2001; West 1990). In this study, we used the González-Romá et al.’s (2009) classification because it is applied in settings such as capability building and team development. In addition, it presents the person-team fit, has been translated into different languages (e.g., Dutch, Italian, Norwegian, Spanish, and Swedish) and has been validated by several studies (e.g., González-Romá et al. 2009). It takes its roots from West’s (1990) study by emphasizing the following four dimensions: (i) support from the organization (which we will call “organizational support”), (ii) innovation (which we will call “innovation orientation”), (iii) goal achievement (which we will call “goal orientation”), and (iv) enabling formalization (which we will call “informal structure”). Organizational support shows the extent to which team members are supported by the whole organization. Innovation orientation shows the extent to which new ideas about work are implemented by the team effectively. Goal orientation shows the extent to which team members make an effort efficiently for reaching goals. Informal structure shows the extent to which team norms and procedures are designed to enable team members to specialize in their tasks and improve their capabilities. Team climate plays a central role in determining members’ behaviors, attitudes, and actions through the improvement of positive psychological atmosphere (Basaglia et al. 2010). It also enables NPD teams to generate shared meaning in situations and to design original methods for reaching mutual outcomes (González-Romá et al. 2009). Team climate also plays a pivotal role in the creation of a supportive psychological atmosphere to enable NPD teams to develop and use information-processing capabilities. We therefore expect that team climate positively affects team cognition. In

123


1153

other words, we consider team climate a critical factor influencing the generation and development of team cognition in software development projects. Following the prior literature, we used the elements of team climate, which include organizational support, innovation orientation, goal orientation and informal structure, to test the effect of team climate on team cognition. Accordingly, we hypothesize as follows: Hypothesis 1 Team climate, which is measured in terms of 1a) organizational support, 1b) innovation orientation, 1c) goal orientation, and 1d) informal structure, will be positively related to team cognition in software development projects. 3.2 Team Climate and Team Intuition The driving power of environmental uncertainty makes it necessary for NPD teams to use their intuition because the role of team cognition is limited when there is asymmetric information in such situations. Similarly, NPD teams sometimes manifest spontaneous reactions to unexpected situations. In these situations, they generally utilize their innate knowledge and capabilities based upon their emotions, imagination, and memories. For these reasons, we also expect team climate to be positively related to team intuition. Following the prior literature, we used the elements of team climate, which include organizational support, innovation orientation, goal orientation, and informal structure, to test the effect of team climate on team intuition. Accordingly, we hypothesize as follows: Hypothesis 2 Team climate, which is measured in terms of 2a) organizational support, 2b) innovation orientation, 2c) goal orientation, and 2d) informal structure, will be positively related to team intuition in software development projects. 3.3 Software Quality as Project Performance The software development process is related to customer requirements, time pressure, and quality issues (Akgün et al. 2007). In this direction, it is crucial to understand what is necessary to produce high quality software products successfully—within a given time frame and budget—in order to meet customer requirements (Akgün et al. 2007; Günsel and Açıkgöz 2013). In general, the term “quality” refers to how well any product satisfies its customers’ requirements (Nidumolu 1995). The ISO/IEC: FCD 9126–1.2. (2001) provide important categorization about quality factors for software products. These factors include functionality (the capability to provide functions which meet stated and implied needs), reliability (the capability to maintain its level of performance under stated conditions), usability (the capability to be understood, learned, and used by users), efficiency (the capability to provide appropriate performance relative to the amount of resources used), maintainability (the capability to be modified through corrections, improvements, and adaptations), and portability (the capability to be transferred from one platform to another). In parallel to this categorization, Nidumolu (1995) reclassified software quality into three factors as follows: (i) operational efficiency (functionality and efficiency), (ii) flexibility (maintainability and portability), and (iii) responsiveness (reliability and usability).

123

1154


Team cognition represents a logical information-processing capability that takes its roots from the systematic acquisition, processing, and dissemination of information. On the other hand, team intuition provides direct access to the teams’ tacit knowledge and experiences. When the teams have access to and use their tacit knowledge and experiences to cope with unexpected events within the software development process, they can go beyond the uncertainties about customer expectations, business needs, and environmental changes. In this context, we argue that both team cognition and team intuition can positively contribute to developing superior software products in order to meet stated and implied needs, adapt to instable business and environmental needs, and generate user-friendly products. Accordingly, we hypothesize as follows: Hypothesis 3 Team cognition will be positively related to software quality in terms of 3a) operational efficiency, 3b) flexibility, and 3c) responsiveness. Hypothesis 4 Team intuition will be positively related to software quality in terms of 4a) operational efficiency, 4b) flexibility, and 4c) responsiveness. 3.4 Moderator Effect: Project Complexity According to the Standish Group (2001), 16–28 % of information technology projects (e.g., software development projects) can be considered successful (i.e., completed within a given time frame and budget with all features and functions as originally specified), 23–40 % can be considered unsuccessful (i.e., canceled before completion or never implemented), and 33–53 % can be considered challenged (i.e., completed and operational, but over-budget, past deadline and/or with fewer features and functions than initially specified). The reason for these high rates of unsuccessful projects may be that developing any software product is a highly complex process because of the interaction of a large number of parts in non-traditional ways (Dawidson et al. 2004; Vlaar et al. 2008). According to Geraldi (2009), the term “complexity” implies something unwanted that make the software development project more complicated and challenging to perform. In this sense, Williams (1999) emphasizes that the core source of complexity is the development process of the project (Xia and Lee 2005). In parallel to these perspectives, Lynn and Akgün (1998) developed a scale to measure project complexity, as a process-related variable, including development, commercialization, and communication processes of the projects. In this direction, we use project complexity as a moderator because we believe that project complexity moderates the relationships between the given information-processing capabilities (i.e., team cognition and team intuition) and software quality. Accordingly, we hypothesize as follows: Hypothesis 5 Project complexity positively moderates the relationship between team cognition and software quality in terms of 5a) operational efficiency, 5b) flexibility, and 5c) responsiveness. Hypothesis 6 Project complexity positively moderates the relationship between team intuition and software quality in terms of 6a) operational efficiency, 6b) flexibility, and 6c) responsiveness.

123


1155

4 Research Design The research design entailed a large-scale cross-sectional survey. Based on the literature review, a field study was conducted with five managers from the Turkish software development industry; discussed the design implications and directions in order to lay the groundwork for our research. To delve deeply into the dynamics of the industry, interviews were conducted with the managers. We gained much useful information from these interviews. As a result of the interviews, using the new knowledge as a basis for change and implementing these changes to current team practices including changing behaviors and shared mental models of the team members were emerged as a major challenge. For example, one of the managers stated that “The change process is the most difficult—to change the way people work. There are a lot of good ideas, tools and methods, but it is too frustrating to start using them.” Based on these interviews which emphasized the difficulties of change, a list of constructs was identified; consequently the survey questions were configured. In order to illustrate a change in team practices by transferring new knowledge and ideas into team work in our questionnaire, items like “During the project, the team had the ability to transfer customer needs to product design specifications,” “The norms and procedures in my work team help us to find the best way to do things,” or “New ideas are put into practice to improve the work and its results” were employed. The list of the constructs and corresponding measurements were then submitted to a panel of academic experts in the technology and innovation management areas. On the basis of feedback from this panel of experts, a draft questionnaire which was consistent with the proposed constructs was prepared. In order to clarify the research questions, the items were first translated into Turkish by one expert and then translated from Turkish into English by a second expert to ensure that the questions were correctly rendered from English to Turkish. The two translators then jointly reconciled the differences. The suitability of the Turkish version of the questionnaire was then given a preliminary test by five engineers who were involved in at least one software development project. Based on data from the preliminary test, an exploratory factor analysis (EFA) was conducted during which the measurements were refined by checking factor loadings, item-to-total correlation, and Cronbach’s alpha (Nunnally 1978). After the preliminary process, the questionnaire was distributed and collected by the authors, using the “personally administered questionnaire” method.

4.1 Measures Various studies have employed reflective models to measure emotional intelligence, emotional capability, environmental turbulence, team unlearning, and team intuition (Akgün et al. 2007; Günsel and Açıkgöz 2013; Dayan and Benedetto 2011). The latent constructs were assessed using multi-item measures from prior studies on a five-point Likert-type scale ranging from “strongly disagree” (1) to “strongly agree” (5); no single-item constructs were used in the research model. In this study, first order reflective model as opposed to formative model was chosen. Determining the right model (either reflective or formative) is not always easy (Bollen 1989). However, Jarvis

123

1156


et al. (2003) showed the distinctions between types of measurement models in which they indicated that reflective models have the following characteristics: the direction of causality is from construct to measure, measures are expected to be correlated, dropping an indicator from the model does not affect the construct, and measurement error is taken into account at the item level rather than the construct level. As a result of these criteria, since all the indicators are expected to be highly correlated with the latent variable score in this research model, as well as constructs cause measures, it was therefore appropriate to employ reflective measures in our research model. A short explanation of each measure is introduced here. In order to measure the team climate of software development projects, four dimensions adapted from GonzálezRomá et al. (2009) were instituted: organizational support, innovation orientation, goal orientation, and informal structure; each dimensions included four items. To measure team cognition, five questions adapted from Akgün et al. (2002) were employed. Also, in order to measure team intuition, seven questions adapted from Dayan and Elbanna’s (2011) team-level intuitive capability scale were used. As well, three product quality dimensions adapted from Nidumolu (1995) were instituted as follows: operational efficiency, flexibility, and responsiveness. Accordingly, four questions for operational efficiency, three questions for flexibility, and three questions for responsiveness were employed respectively. Project complexity refers to the degree to which the development process is complicated and difficult. The extent of the complexity and difficulty is closely related to the similarity of the development, commercialization, and communication processes used during the project compared to the process the team had been using traditionally. For these reasons, the three-item scale, which was intentionally structured in this vein, developed by Lynn and Akgün (1998) was modified. Therefore, the participants were asked whether the development, commercialization, and communication processes within their project were in congruence with their traditional ones.

4.2 Sampling The initial sample was obtained from 99 firms, which were either directly operating in the software development industry or had a software development department, based on records supplied by the Istanbul Chamber of Commerce. Initially, the managers of these firms were contacted by telephone so that we could explain the aim of our study. Specifically, the respondents were informed that they should be the software engineers/developers who were the most knowledgeable about their software development projects. Of the 99 firms contacted, 71 agreed to participate in this study out of which 42 firms completed our questionnaire (for a response rate of 59 %). A total of 143 questionnaires were returned (several firms participated in our study with more than one respondent). Due to missing data, four questionnaires could not be included. As a result, the final sample for analysis consisted of 139 new software development projects. In terms of organizational sampling distribution, out of 42 participant firms, 5 software development projects were from 9 IT departments of the firms while 4 projects from 9 IT department, 3 projects from 11 IT departments, 2 projects from 12 IT departments, and 1 project from 1 IT department of the participant firms. In

123


1157

addition, 63 % of the projects were from information and communication technology organizations, 24 % of them were from business service organizations, and 13 % of the projects were related to financial service organizations. According to Phillips and Bagozzi (1986), informants tend to provide more valid information or data on issues directly related to their work roles, therefore we expected the software engineers/developers to perceive questions more accurately than nonengineers/developers due to their experience, involvement, and responsibilities. To increase the respondents’ motivation to cooperate by removing any fear of retaliation, each respondent was assured that his or her response would remain anonymous (Huber and Power 1985). The respondents were also informed that there were no predetermined right or wrong answers in order to encourage them to respond the questions as honestly and directly as possible. These reassurances reduced the participants’ resistance; as a result it allowed them less likely to edit their answers so as to make them socially desirable, permissive, or consistent with how they thought the researchers wanted them to respond (Podsakoff et al. 2003). Only one team member from each team was requested to participate in our survey. Also, these team members were asked to evaluate one unique project. Accordingly, this research study used a convenience sample of 139 participants from different teams consisting of both males (85 %) and females (15 %). The ages of the participants were as follows: 24 % under 26 years of age, 33 % between 26–28, 21 % between 29–31, 14 % between 32–34, and 8 % over 35. In addition, the work experience of participants was as follows: 58 % 0–5 years, 27 % 6–10 years, and 15 % more than 10 years. The team sizes were as follows: 39 % 3–5 developers, 28 % 6–9 developers, 20 % 10–15 developers, 6 % 16–19 developers, and 7 % more than 20 developers. 4.2.1 Measure Validity and Reliability After collection, the data were subjected to a purification process in order to evaluate their reliability, discriminant validity, convergent validity, and unidimensionality (Fornell and Larcker 1981; Segars 1997). EFA was conducted into 39 measured items; the constructs comprises ten variables. A principal component with a varimax rotation was employed; an eigenvalue of 1 as the cut-off point were selected. Due to the low levels of factor loadings, five items were dropped from the analysis—one from organizational support, one from goal orientation, two from team intuition, and one from operational efficiency. An examination of these items revealed that dropping them would not compromise the content validity of organizational support, goal orientation, team intuition, and operational efficiency. A single factor was extracted for each multiple-item scale in this analysis. The Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy was .840 which was higher than the proposed threshold value of 2 = 2,728.890). .7, also the Bartlett test of sphericity was significant at p < .000 (χ(528) These results indicate the appropriateness of the data for EFA procedure. The items (including the dropped items) and their factor loadings after EFA, eigenvalue, and percentage of variance explained appear in Table 1. Additionally, the extent of common method bias with Harman’s one-factor test was measured. The test includes entering all constructs into an unrotated principal components factor analysis and examining the resultant variance (Harman 1960). The

123

1158


Table 1 The result of exploratory factor analysis LV

Manifest variables

OS

IO

In my work team... Team members feel supported by the organization You can tell that the company is interested in the members of the team The human resources management is carried out keeping the team members in mind *The team manager contributes to creating a friendly and cordial work climate In my work team . . .

GO

New ideas and methods are often tried out New ideas are put into practice to improve the work and its results The development of new methods, products or services is often proposed Team members take advantage of their knowledge and skills to develop new ways of working, new services or new products In my work team . . .

IS

TC

SL

VE (%)

UV (%)

2.53

7.67

5.68

3.06

9.28

9.88

2.08

6.31

2.53

3.32

10.07

31.21

2.28

6.90

3.82

.76 .78

.73

–

.79 .82 .82

.53

Team members try hard to reach the team goals.

.70

Team members aspire to achieving greater performance.

.81

Everyone contributes enthusiastically to reaching the goals.

.64

*High, difficult goals are viewed as a challenge.

–

The norms and procedures in my work team . . . Help our team to function better.

.78

Help us to find the best way to do things.

.77

Facilitate relationships between team members.

.79

Help us to understand the relationship between each person’s work and that of his/her co-workers *During the project, the team had the ability to transfer customer needs to product design specifications During the project, the team had the ability to generate different market and technology scenarios During the project, the team had the ability to gather information from different functions within our organization During the project, the team had the capability to gather information from outside the organization (for example, customers, vendors, suppliers, libraries, consultants, and so on)

.82

123

E

– .64

.79

.79


1159

Table 1 continued LV

Manifest Variables

SL

E

VE (%)

UV (%)

TI

On many occasions, the members of our team did not have enough information and had to make some decisions based on a “gut feeling.” To what extent did the team members in this project depend on a “gut feeling”? Did team members trust their hunches when confronted by an important decision during this project? Did team members put more emphasis on feelings than data when making decisions during this project? Did team members’ intuition turn out to have been right all along? *To what extent did participants in this project rely basically on personal judgment? *Did team members put a lot of faith in their initial feelings about other people and situations? The development process used in this project was different to the process the team traditionally uses The commercialization process used in this project was different to the process the team traditionally uses The processes used on this project to communicate across functional disciplines, with suppliers and customers were different to the communication process the team traditionally uses The software is reliable.

.80

2.59

7.85

6.84

2.50

7.57

5.15

2.42

7.33

4.64

There is a quick response time by the product.

.65

The client is satisfied with the overall operational efficiency of the software *The cost of software operations is efficient.

.68

2.14

6.49

3.04

2.16

6.54

3.24

PC

OE

FL

RE

The software adapts to changes in business with cost efficiency The software adapts to changes in business requirements.

.79 .80 .71 – – .83 .83 .82

.67

– .79 .77

The final product achieves overall long-term flexibility of the software The software is easy to use.

.42

The software customizes outputs to various client needs

.80

The software is responsive overall to client needs.

.67

.56

LV Latent variable, SL standardized loading, E eigenvalue, VE variance explained, UV unrotated variance, OS organizational support, IO innovation orientation, GO goal orientation, IS informal structure, TC team cognition, TI team intuition, PC project complexity, OE operational efficiency, FL flexibility, RE responsiveness * Denotes the dropped item

threat of common method bias is high if a single factor accounts for more than 50 % of the variance (Harman 1960; Pee et al. 2010; Podsakoff and Organ 1986). Our results demonstrated that none of the factors significantly dominated the variance (see the last column of Table 1); hence, it is clear that common method bias was unlikely. Since EFA alone does not provide an explicit test of unidimensionality (Segars 1997), a confirmatory factor analysis (CFA) was performed as well. In order to

123

1160


assess the discriminant validity, two-factor models (as recommended by Bagozzi et al. 1991) were estimated, in which individual factor correlations, one at a time, were restricted to unity. The fit of the restricted models was compared to that of the original model. In total, 90 models were evaluated using AMOS (Segars 1997; Akgün et al. 2007). As shown in Table 2, the Chi-square change (χ 2 ) in each model, constrained and unconstrained, was significant (χ 2 > 3.84), which suggests that the constructs demonstrate discriminant validity (Anderson and Gerbing 1988). The measures were also subjected to one model CFA. As shown in Table 3, the 2 = resulting measurement model was found to fit the data reasonably well: χ(450) 711.662, comparative fit index (CFI) = .89, incremental fit index (IFI) = .90, Tucker– Lewis Index (TLI) = .88, χ 2 /d. f. = 1.58, and root mean square error of approximation (RMSEA) = .07. In addition, all items loaded significantly on their respective constructs (with the lowest t value being 2.50), providing support for convergent validity. Table 4 shows the correlations among all 10 variables. The relatively low-tomoderate correlations provide further evidence of discriminant validity. Also, all the reliability estimates, including the coefficient alphas, the average variance extracted (AVE) for each construct, and the AMOS-based composite reliability are well beyond the threshold levels suggested by Nunnally (1978) and Fornell and Larcker (1981). Further, as suggested by Fornell and Larcker (1981), the squared root of AVE for each construct was greater than the latent factor correlations between the pairs of constructs which suggests discriminant validity. All in all, the obtained results conclude that the measures were unidimensional, having adequate reliability and discriminant validity.

4.2.2 Hypothesis Testing Partial Least Squares (PLS) approach (Smart PLS 2.0; Ringle et al. 2005 ) and the bootstrapping re-sampling method (Chin 1998) were used in this study to estimate both the main and the interaction effects in our proposed model (see Fig. 1). 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 sub-sample. T -statistics were calculated for all coefficients based on their stability across the sub-samples in order to determine which links were statistically significant. The path coefficients and their associated t-values demonstrated the direction and impact of each hypothesized relationship. As suggested by Chin et al. (2003), a hierarchical approach to test our hypotheses was employed: first a model with main effects (and covariates) only was assessed and then the interaction effects were added. Table 5 shows our hypotheses, including paths, betas, significance levels, and results. With regard to antecedents, the findings illustrated that three sub-dimensions of team climate—innovation orientation (β = .251, p < .01), goal orientation (β = .338, p < .01), and informal structure (β = .205, p < .05)—were positively associated with the cognitive capability of the software development teams. However, the results showed that there was not any significant statistical association between the organizational support and team cognition as well as

123

The Moderating Role of Project Complexity Table 2 Discriminant analysis of the construct measures

1161

Constructs

Unconstrained (χ 2 /d. f.)

Constrained (χ 2 /d. f.)

χ 2

22.2/13

48.6/14

26.4 29.4

OS

↔

IO

OS

↔

GO

13.4/8

42.8/9

OS

↔

IS

19/13

39.6/14

20.6

OS

↔

TC

14.6/8

41.6/9

27

OS

↔

TI

6.2/13

64.4/14

58.2

OS

↔

PC

9.4/8

49.9/9

40.5

OS

↔

OE

7.6/8

39.5/9

31.9

OS

↔

FL

24.3/8

48.5/9

24.2

OS

↔

RE

4.9/8

40.5/9

35.6

IO

↔

GO

21.3/13

66.2/14

44.9

IO

↔

IS

48/19

84.3/20

36.3

IO

↔

TC

20.2/13

55.5/14

35.3

IO

↔

TI

30.7/19

101.6/20

70.9

IO

↔

PC

29.7/13

72.4/14

42.7

IO

↔

OE

22.7/13

71.9/14

49.2

IO

↔

FL

40.6/13

75.3/14

34.7

IO

↔

RE

15.9/13

57.8/14

41.9

GO

↔

IS

23.8/13

64/14

40.2

GO

↔

TC

11/8

43.2/9

32.2

GO

↔

TI

24.1/13

83.7/14

59.6

GO

↔

PC

16.2/8

62.8/9

46.6

GO

↔

OE

8.6/8

60.4/9

51.8

GO

↔

FL

12.1/8

56.5/9

44.4

GO

↔

RE

16.6/8

59.8/9

43.2

IS

↔

TC

12.2/13

74.8/20

62.6

IS

↔

TI

30.7/19

102.1/20

71.4

IS

↔

PC

26.9/13

80.2/14

53.3

IS

↔

OE

23.4/13

73.8/14

50.4

IS

↔

FL

16.2/13

63.7/14

47.5

IS

↔

RE

24.9/13

72.7/14

47.8

TC

↔

TI

18/13

80.6/14

62.6

TC

↔

PC

16.8/8

50.8/9

34

TC

↔

OE

3.8/8

64.3/9

60.5

TC

↔

FL

24.4/8

68.1/9

43.7

TC

↔

RE

11.3/8

57.1/9

45.8

TI

↔

PC

29/13

66.6/14

37.6

TI

↔

OE

32.9/13

124.9/14

92

TI

↔

FL

21/13

108/14

87

123

1162 Table 2 continued

OS Organizational support, IO innovation orientation, GO goal orientation, IS informal structure, TC team cognition, TI team intuition, PC project complexity, OE operational efficiency, FL flexibility, RE responsiveness


Constructs

Unconstrained (χ 2 /d. f.)

Constrained (χ 2 /d. f.)

χ 2

TI

↔

RE

13/13

93.2/14

80.2

PC

↔

OE

13.3/8

98.5/9

85.2

PC

↔

FL

10.8/8

80.3/9

69.5

PC

↔

RE

13.2/8

81.1/9

67.9

OE

↔

FL

6.9/8

43/9

36.1

OE

↔

RE

9.5/8

55.1/9

45.6

FL

↔

RE

15.8/8

52.7/9

36.9

between the team climate and intuitive capability of the software development teams. Accordingly, H1 was partially supported, while H2 was not supported. Concerning the project outcomes, the research results demonstrated that team cognition was positively associated with three sub-dimensions of software quality— operational efficiency (β = .320, p < .01), flexibility (β = .390, p < .01), and responsiveness (β = .416, p < .01). In contrast, the results provided no empirical evidence supporting the relationships between the intuitive capability of the software development teams and software quality. According to these results, H3 was fully supported while H4 was rejected. To address the hypotheses pertaining to the moderating effects of project complexity (H5–H6), a two-step construction procedure (Chin et al. 2003) was employed. The PLS approach allows for explicit estimation of the standardized latent variable scores after saving the obtained results (Tenenhaus et al. 2005). The relevant interaction terms using the product-indicator approach were constructed and then these were included in the model. This method enabled us to test for a relatively large number of interaction effects while simultaneously correcting for measurement error (Chin et al. 2003). The results demonstrated a positive interaction effect between team cognition and flexibility (β = .315, p < .01) as well as between team cognition and responsiveness (β = .357, p < .01), while at the same time providing no evidence in support of an interaction effect between team cognition and operational efficiency. Surprisingly, we could not find a statistical relationship between the interaction effect of team intuition and software quality. As a result, H5 was partially supported while H6 was not supported. 4.2.3 Structural Model The PLS structural model was validated by the R 2 of the endogenous latent variable (Chin 1998), the effect size f 2 (Cohen 1988), and the Goodness-of-Fit (GoF) index (Tenenhaus et al. 2005). According to Chin (1998), threshold R 2 values of .26, .13, and .02 for endogenous latent variables are considered to be large, medium, and small respectively. Following this, effect sizes were estimated using a method identified by Cohen (1988) and recommended by Chin et al. (2003) in PLS path models. According to this method, effect size is attained by comparing the explained amount of vari-

123

The Moderating Role of Project Complexity Table 3 Measurement models and confirmatory factor analysis

Construct

Parametera

Standardized Coefficient

t valueb

OS

λ O S1

.81

Scaling

λ O S2

.96

11.61

λ O S3

.62

7.68

λ I O1

.82

Scaling

λ I O2

.87

11.33

λ I O3

.82

10.60

λ I O4

.61

7.32

λG O1

.75

Scaling

λG O2

.63

5.96

λG O3

.82

5.93

λ I S1

.87

Scaling

λ I S2

.77

10.81

λ I S3

.82

11.86

λ I S4

.80

11.48

λT C1

.79

Scaling

λT C2

.79

8.92

λT C3

.75

8.49

λT I 1

.71

Scaling

λT I 2

.72

6.99

λT I 3

.76

7.18

λT I 4

.59

5.95

λ PC1

.77

Scaling

λ PC2

.78

8.25

λ PC3

.78

8.25

λ O E1

.75

Scaling

λ O E2

.83

9.67

λ O E3

.85

9.83

λ F L1

.71

Scaling

λ F L2

.80

8.45

λ F L3

.76

8.06

λ R E1

.67

Scaling

λ R E2

.74

7.41

λ R E3

.85

8.15

IO

GO

IS

TC

TI

2 χ(450) = 711.662, CFI = .89, IFI = .90, TLI = .88, RMSEA = .07. OS Organizational support, IO innovation orientation, GO goal orientation, IS informal structure, TC team cognition, TI team intuition, PC project complexity, OE operational efficiency, FL flexibility, RE responsiveness a parameters indicate paths λ from measurement items to first-order constructs b Scaling denotes value of λ indicator set to 1 to enable latent factor identification

1163

PC

OE

FL

RE

ance when a predictor is either included or not included in the model. In addition, threshold f 2 values of .02, .15, and .35 represented small, medium, and large effects respectively. Furthermore, GoF was employed to globally evaluate the overall fit of the model, seeking a concordance between the performance of the measurement and the structural model, as well as consistent with the geometric mean of the average communality and the average R 2 of endogenous latent variables. GoF ranges between 0 and 1: a higher value represents better path model estimation. According to the categorization

123

123

3.50

3.90

4.12

3.63

3.63

2.71

3.00

3.98

4.04

4.17

1

2

3

4

5

6

7

8

9

10

.49** .90 .74 .83

GO

IS

TC

TI

PC

OE

FL

RE

CR

AVE

α

.77

.81

.84

.85

.96

.73

.73

.73

.86

IO

.47**

.39**

.18*

−.04

.37**

.52**

.39**

.45**

OS

1

.79

Variables

1.02

Standard Deviation

.84

.48**

.48**

.36**

.85

.70

.90

.48**

.50**

.50**

.21*

−.09

2

.78

.69

.88

.42**

.39**

.47**

.23**

.06

.53**

.44**

.83

3

.49**

.87

.89

.75

.92

.50**

.36**

.45**

.22**

−.04

4

.82

.73

.89

.35**

.38**

.43**

.34**

.01

.85

5

.17

.75

.79

.56

.83

−.12

−.17*

−.06

6

.82

.50

.72

.02

.06

.13

.71

7

.85

.77

.91

.63**

.61**

.88

8

.80

.71

.88

.68**

.84

9

.79

.71

.88

.84

10

OS Organizational support, IO innovation orientation, GO goal orientation, IS informal structure, TC team cognition, TI team intuition, PC project complexity, OE operational efficiency, FL flexibility, RE responsiveness, CR composite reliability, AVE average variance extracted, α = Cronbach’s Alpha * p < .05; ** p < .01. Diagonals show the square root of AVEs

Mean

No

Table 4 Correlations and descriptive statistics

1164 A. Açıkgöz et al.


1165

Team Cognition

H3

H1 Organizational Support Innovation Orientation

H5 Operational Efficiency Team Climate

Moderator: Project Complexity

Software Quality

Goal Orientation

Flexibility

Responsiveness

H6 Informal Structure

H2

H4 Team Intuition

Fig. 1 Proposed model

of the effect sizes for R 2 by Cohen (1988), using .5 as a cut-off value for communality (Fornell and Larcker 1981), threshold values for GoF criteria were categorized as R 2 : .1 demonstrates small effect sizes; R 2 : .25 shows medium effect sizes; and R 2 : .36 represents large effect sizes. Table 6 shows the results for our structural model. The R 2 values of the endogenous constructs were used to assess model fit (Chin 1998; Tenenhaus et al. 2005). According to the obtained results, the effect sizes of constructs were large for the value of the team cognition (R2 = .41), flexibility (R2 = .30), and responsiveness (R2 = .35); medium for the value of the operational efficiency (R2 = .25); and small for the value of the team intuition (R2 = .04). Because of the interaction effect of project complexity, the R 2 for the value of operational efficiency in the final model was .25 while the main effects model yielded only .15. Using an incremental F-test (Chin et al. 2003), the results asserted that the R 2 i − R 2 e of .10 was statistically significant at α = .05 (F2.133 = 8.87; p = .0002). Following this test, f 2 was estimated to assess the effect size of the interaction terms in the final model, so the results suggested a small-to-medium effect size ( f 2 = .4). The same statistical analysis procedures were performed for flexibility and responsiveness as well. Accordingly, though the R 2 of flexibility in the main effects model yielded .19, the R 2 of it in the final model was .30. The results showed that the R 2 i − R 2 e of .11 was statistically significant at α = .01(F2.133 = 10.45; p < .0001) which states a medium-to-large effect size of flexibility ( f 2 = .16). Similarly, even though the R 2 of responsiveness in the main effect model was .20, the value of R2 of it in the final model was .35. In terms of incremental F-test, the R 2 i − R 2 e of .15 was statistically significant at α = .01(F2.133 = 15.35; p < .0001) which indicates a medium-to-large effect size ( f 2 = .23).

123

1166


Table 5 The results Relationships

Path Coefficient (β)

Sub-hypotheses

Sub-results

Hypotheses

Results

−.031

H1a

H1

Partially supported

H2

Not supported

H3

Fully Supported

H4

Not supported

H5

Partially supported

H6

Not supported

OS

→

TC

IO

→

TC

.251**

H1b

Not Supported Supported

GO

→

TC

.338**

H1c

Supported

IS

→

TC

.205*

H1d

Supported

OS

→

TI

.012

H2a

IO

→

TI

−.205

H2b

GO

→

TI

.110

H2c

IS

→

TI

−.014

H2d

TC

→

OE

Not supported Not supported Not supported Not supported Supported

TC

→

FL

.390**

H3b

Supported

TC

→

RE

.416**

H3c

Supported

TI

→

OE

−.142

H4a

TI

→

FL

−.166

H4b

TI

→

RE

−.077

H4c

TC*PC

→

OE

.282

H5a

TC*PC

→

FL

.315**

H5b

Not supported Not Supported Not Supported Not supported Supported

TC*PC

→

RE

.357**

H5c

Supported

TI*PC

→

OE

.080

H6a

TI*PC

→

FL

−.077

H6b

TI*PC

→

RE

.074

H6c

Not supported Not supported Not supported

.320**

H3a

OS Organizational support, IO innovation orientation, GO goal orientation, IS informal structure, TC team cognition, TI team intuition, PC project complexity, OE operational efficiency, FL flexibility, RE responsiveness * p < 0.05; ** p < 0.01

According to another fit measure, the result of GoF was .43 for the final model while it was .37 for the main effects model. The obtained GoF results indicate a good fit (see Table 6). 5 Discussion and Implications In this study, we attempted to contribute to the RBV and OLT in general, and SLT in particular, by presenting a model to understand interrelationships among team climate, team cognition, team intuition, and software quality.

123


1167

Table 6 Structural model Fit measures

Endogenous constructs

Main effect model

Final model

R2

TC TI OE FL RE

.41 .04 .15 .19 .20 .37

.41 .04 .25 .30 .35 .43

GoF

Probability (α)

f2

= .0002