Master Thesis for Master BA Change Management

(1)

An Empirical Analysis of the impact of leadership on adaptive system use

How can leadership contribute to adaptive system usage of IT in teams? By Stephanie Kuipers S2163071 Eikenlaan 109 9741 EK Groningen a.s.kuipers@student.rug.nl University of Groningen Faculty of Economics and Business

January 2017 Supervisor(s): Dr. I. Maris-de Bresser MA. E. Smailhodzic

(2)

Abstract

The main purpose of this research is to better understand adaptive system use by further examining how transactional and transformational leadership styles can contribute to adaptive system use and its triggers to decrease underutilization of features in IT. This study is the first to combine transactional and transformational leadership in a research model to address adaptive system use. An empirical study across several industries is conducted, in which 117 responses were collected among team members that use IT on a daily basis. The study confirms the positive effect of the initial triggers of Sun (2012) and surprisingly shows that transactional leadership has a more extensive effect on adaptive system use than transformational leadership. Further research is necessary to confirm the effects of transactional and transformational leadership on adaptive system use. The practical implications suggest that managers should keep transactional leadership methods in mind whilst trying to encourage employees in adaptive system use. Moreover, this research provides interesting insights into the understanding of adaptive system use and its interaction with leadership.

Keyword: Post-adoptive behaviors, transactional leadership, transformational leadership, adaptive system use, personal innovativeness in IT, triggers, features in use.

(3)

Table of Content

Abstract 1

Introduction 3

Theoretical Background 6

Adaptive System Use 6

Leadership styles 12 Personal Innovativeness in IT 17 Conceptual model 18 Methodology 19 Data Collection 19 Sample Description 20 Measurements 21

Analysis and Results 22

Analysis 22

Results 29

Discussion 30

Theoretical implications 32

Practical implications 32

Limitations and future research 32

References 33

Appendix 42

(4)

Introduction

In the last decade, IT has transformed from a mere resource (Porter and Miller, 1985) into one of organization's major backbones (Preittigun, Chantatub and Vatanasakdakul, 2012) causing organizations to be innovative in order to remain competitive (Lertpachin, Wingwon and Noithonglek, 2013). IT has become essential for organizations to remain relevant in today’s complex and rapidly changing business environment (Carr, 2003; Porter, 1996; Zammuto, Griffith, Majchrzak, Dougherty and Faraj, 2007; Jasperson, Carter and Zmud, 2005), but at the same time making organizations more dependent on IT (Wang, Butler, Hsieh and Hsu, 2008). Despite that, organizations invest significant amounts of their budgets into IT (Stratopoulos and Lim, 2010), but unfortunately the investments do not always live up to their expectations (Jasperson et al., 2005). The investments are often costly and rather unsuccessful (Legris, Ingham and Collerette, 2003) especially as organizations are unable to translate the investments into the expected increase of firm performance (Soh and Markus, 1995; Tippins and Sohi, 2003; Bala and Venkatesh, 2016). One of the reasons for this disappointing contribution is resulting from actual underutilization of implemented IT (Jasperson et al., 2005). Nevertheless IT-related innovations are still expanding more than ever (Burnes, 2014; Cawsey, Deszca and Ingols, 2016) which sparks interest in the research field of IT. Consequently, research started to pay attention to post-adoptive system behaviors, focusing on the dynamic processes of individual IT use behaviors appearing through interaction among users, technologies and institutional properties (Nan, 2011).

With the increasing interest in the literature on the processes of adaptation to IT, several concepts have been developed that discuss the contingencies to make adaptation effective. For example Bala and Venkatesh (2016) introduced a model about technology adaptation behaviors. Their model states that the employees’ experiential engagement and psychological engagements during the implementation determine their technology adaptation behaviors (Bala and Venkatesh, 2016). Furthermore, both Markus (1983) and Lapointe and Rivard (2005) present models on forms and sources of resistance for implementing technological change. Although resistance to change is argued to be a customary and logical response to change (Markus, 1983; Buelens et al., 2011) resistance does not represent recipients adapting to the new situation but in contrast trying to hold on to the equilibrium of the past. This study chose to focus on adaptive system use, because it is the most extensive concept concerning adaptation behaviors. Adaptive system use (ASU) is the accumulation of a certain set of behaviors which all deal with the revision of someone’s use of IT features (Sun, 2012). The feature level refers to the focus on the specific usage of the functionalities within IT. In comparison to the other models mentioned above, adaptive system use seems to be the only model that (besides focusing on the feature level) considers why behavioral changes occur. The components of adaptive system use present several manners of use and adaptation on a feature level, which can be initiated by the triggers: novel situations, discrepancies and deliberate initiative (Sun, 2012). Therefore, the aim of this study is to better understand adaptive system use.

(5)

been studied in the context of IT adoption, it has not been investigated in relation to adaptive use, in which the academic interests in IS literature is increasing, making it interesting to investigate within this context.

Transactional leadership is well known for its influence on motivation and job performance (Bass, 1999). By establishing a frame of reference of desirable behavior through rewarding desirable behavior and punishing non-desirable behaviors leaders can stimulate innovative behavior within a controlled environment (Zhou, 2003). However, researchers have not yet established if this is also applicable to the follower’s technology adaptation and use and how to do so, leaving an interesting research gap within the context of the interplay between leadership and technology adaptation. Until now leadership has only been specified in the theory of adaptive system use by mandatory use in form of deliberate initiatives (Sun, 2012), but rewarding desirable behavior within the concept of transactional leadership is distinctive from mandatory use as followers still have the possibility of choosing not to perform extra effort. Furthermore, there is existing research addressing the influence of leadership behavior like transactional leadership on innovative behavior (Zhou, 2003), but it does not consider the feature level that is an essential part of adaptive system use, nor does it pay attention to how it can contribute to the effectiveness of subordinates on a team level. This leaves a research gaps that needs to be answered to fully understand the interplay between this leadership style and adaptive system use.

Transformational leadership is in research often presented as the better half of transactional leadership, because it illustrates ways to motivate others and trigger their internal motivation (Steward, 2006). The possible correlation between transformational leadership and innovative behavior makes it an interesting component to investigate in relation with adaptive system use (Janssen, 2003). Innovative behavior is a collective noun for the whole process of working towards realization of improving group work by introducing new ideas to improve performance (Kanter, 1988; Scott and Bruce, 1994; West, 1989; West and Farr, 1989; Woodman, Sawyet and Griffin, 1993). Adaptive system use can be argued to be a specific distinction of innovative behavior, which focuses on the feature level of IT. It seems necessary to review the established theories of transactional and transformational leadership through the perspective of IT, because most researchers on these leadership styles did not consider the increased importance of IT for the continuity of organizations (Podsakoff, MacKenzie and Bommer, 1996; Bass, 1985). Some researchers have indicated leadership as a possible antecedent for adaptation behavior (Beaudry and Pinsonneault, 2005; Bruque, Moyano and Eisenberg, 2008), but none of them elaborates on how to do so nor have they elaborated on which of these leadership styles are effective.

(6)

model has not yet been researched and has the potential to uncover some interesting results. Besides, this would be a manageable model considering the scope of the research project.

RQ1: How do transactional and transformational leadership enable teams towards adaptive system use?

This study adds an additional layer to its conceptual model by including personal innovativeness in IT as a moderator of the relationship between leadership and adaptive system use. Using PIIT as a moderator is a logical step, because the willingness of an individual to try any new technology is a necessary starting point before supervisors consider spending time and resources on stimulating subordinates to perform adaptive system use behaviors. This reasoning is corresponding with the argument of Louis and Sutton (1991) on the influence of individual factors. They argue that the existence of triggers does not guarantee active thinking and behavior. In other words, it is a condition before individuals may act upon certain behavioral triggers, that they need the ability and willingness to notice the presence of the trigger (Burke, Stagl, Sales, Pierce and Kendall, 2006; Langer, 1986).

By doing so, this study is one of the first to consider both personal characteristics of employees (in the form of personal innovativeness in IT) and the influence of leadership styles in explaining innovative behavior (in the form of adaptive system use) within one research model (Janssen, 2000: 2003). The vigilant attitude of a person, capability and easiness of someone to be innovative should help individuals in situations of discrepancies of novelty. Therefore this study argues that personal innovativeness in IT has a substantial effect on the ability to change behavior. The second research question concerns the moderating effect of PIIT between leadership and adaptive system use.

RQ2: How does personal innovativeness in IT moderate the effect of leadership on adaptive system use?

(7)

Theoretical Background

This chapter will elaborate on the academic concepts that serve as the foundation for this study. First, a basic understanding of the dependent variable adaptive system use will be established together with the corresponding concepts. Subsequently, specific attention will be paid to the triggers responsive for adaptive system use as the research gap of this study is specifically focusing on them. Next, the leadership styles of Bass (1985: 1990: 1999) will be elaborated and their possible connection with adaptive system use, followed by a closer look at the bifurcation of transformational leadership and transformational IT leadership. Finally, personal innovativeness in IT is presented. The study of Sun (2012) functions as the basis for this research.

Adaptive System Use

The main purpose of this research is to better understand adaptive system use by further examine how leadership styles can contribute to adaptive system use and its triggers to decrease underutilization of features in IT. Adaptive system use (ASU) is a post-adoptive active system use at the feature level (Sun, 2012). ASU is the accumulation of a certain set of behaviors which all deal with the revision of someone’s use of IT features. Sun (2012) splits these behaviors up in two dimensions: revising the content of feature in use and revising the spirit of features in use, which in other words represents the differentiation between “content” what features are used and the “spirit” which means how these features are used.

Table 1. Dimensions and Sub-dimensions of Adaptive System Use (Sun, 2012) Revising the Content of Features in Use

A user’s revising regarding what features are included in his/her Features in Use: what features are used. Trying New Features Add new features to one’s Features in Use and so

expanding the scope of the Features in Use. Feature Substitution Replacing features in the Features in Use with other

features with similar functions. Revising the Spirit of Features in Use

A user’s revisions regarding how features in his/her Features in Use are used.

Feature Combining Using features in Features in Use together for the first time.

(8)

In accordance with this, Yamauchi and Swanson (2010) introduced the concept of familiarity pockets, defining it as “work routines and components accumulated through situated interactive use of the system” (Yamauchi and Swanson, 2010, p.200). In other words, each individual has its own personal collection of interactive features (familiarity pocket), which is established by his/her interaction with the IT. The individual essence of familiarity pockets allows it to make several distinctions. Firstly, Yamauchi and Swanson (2010) argue that a user’s familiarity pocket incorporates the routines and the constructive components of these routines including all the actions and sequences the user bases his/her working routine on (Yamauchi and Swanson, 2010; Pentland, Liu, Feldman and Becker, 2009). Furthermore, the interactive nature of a user’s interactions with IT connects this individual not only to the system but also to other users whom are also interacting with the IT, allowing the facilitation of feedback and adaptive learning as these interactions slowly determine sequences and routines (Yamauchi and Swanson, 2010). Finally, users will move in and out of their familiarity pocket, which serves as the basis for improvisation (Yamauchi and Swanson, 2010). This shows that the development of the familiarity pocket can be interpreted as the process of a narrative network: a set of specific patterns that has been or could be generated by combining and recombining the elements of the system (Feldman and Pentland, 2003). The study considers the team level specifically as a result of the interactive nature of adaptive system use, arguing teams are a somewhat secluded environment that interacts and is dependent on each other as a living organism (Buelens et al., 2011). This means that team members learn and adapt their behavior as a result of their interaction with each other (Buelens et al., 2011). In light of this, this study argues that the group interaction is a vital part of the process of adaptive system use.

The presentation of the familiarity pockets is relevant, because it show the pitfalls of the theory of adaptive system use. In other words, the familiarity pockets (Yamauchi and Swanson, 2010) are the starting point for this study to discuss the pitfalls of adaptive system use. The pitfalls are relevant, because they demonstrate why the original model of ASU (Sun, 2012) needs to be further expanded.

The narrative network perspective Yamauchi and Swanson (2010) use for their introduction of familiarity pockets is in contrast with Sun’s (2012) notion that features in use are solid motion, hence a contradiction in the theory or at least limitations of applicability of adaptive system use in complex environments. First of all, the benefits of adaptive system use are uncertain. Some researchers argue that due to system complexity and malleability which supplies the user with a bandwidth of how to use the system, it is not guaranteed that new features or new manners to use them will be beneficial in comparison to the old situation. Some research suggests that upon initial implementation, firm performance often drops, rather than improves as employees grapple with the transition (Markus and Tanis, 2000; Ross, 1998). Therefore, some researchers argue that improving organizational performance can only be achieved via higher levels of system use (Cooper and Zmud, 1990).

(9)

resistance to the IT (Lapointe and Rivard, 2005). Lastly, in line with the theory of Louis and Sutton (1991) on the circle of cognitive processing, the last noted risk of adaptive system use is that it can be rather time consuming as it takes time for new behaviors to convert into the new routines (Louis and Sutton, 1991; Bargh, 1989). This being stated, I continue by introducing the components of adaptive system use, which are subdivided into the dimensions of adaptive system use mentioned earlier.

Revising the content of features in use. The first dimension of adaptive system is concerned with what features are used (Sun, 2012). It includes trying new features (Barki, Titah and Boffo, 2007; Jasperson et al., 2005) and feature substitution (Parthasarathy and Bhattacherjee, 1998) according to Sun’s theory (2012).

According to Sun (2012) trying new features is observed user adaptation behavior as by doing so an individual is permanently expanding his/her features in use and can be construed as a practice which extends an individual's knowledge of features. Users are continuously discovering and adopting new features (even in the post-adoptive phase of the information system) and when they gain experience with the system this will expand even further (Hiltz and Turoff, 1981; Sun, 2012).

Meanwhile, feature substitution can be done both physically and psychologically (Sun, 2012) and is not permanent or exclusive. Sun (2012) argues that users may go back to old features when these are considered to be useful for the task at hand or if external factors compel them to. For example the substituted features are not compatible with other features in the FIU. Nevertheless, as the definition implies it does expand the FIU of the user. Therefore, these components of adaptive system use are to be considered parts of IT usage. IT usage is defined as a “set of behaviors individuals undertake concerning a specific task technology-individual-context” (Barki et al., 2007, p.174). As already described above, these behaviors deal with actions users perform in pursuit of a task for which IT is needed (Goodhue, 1995; Goodhue and Thompson, 1995).

Revising the spirit of features in use. The other dimension of adaptive system use is concerned with the manners in which features are used within the collection of an individuals’ FIU (Sun, 2012). Research suggest that users may use existing features in ways that are not intended by the developers of the feature, but also in manners that allow an individual or group of users to complete a task under the condition that it matches the emergent conceptualization (Harrison and Datta, 2007; Sun, 2012). Accordingly, these components are demonstrating innovative behavior with features in use. The dimension of revising the spirit of features in use contains feature combining (Boudreau and Robey, 2005; Desouza, Awazu and Ramaprasad, 2007; Rice and Roger, 1980) and feature repurposing (Ahuja and Thatcher, 2005; Desouza et al., 2007; Jasperson et al., 2005; Saga and Zmud, 1994; Singletary, Akbulut and Houston, 2002).

(10)

Finally, feature repurposing is defined as using features in the FIU in new ways, which may not have been intended by the developers (Sun, 2012; Ahuja and Thatcher, 2005; Desouza et al., 2007; Jasperson et al., 2005; Saga and Zmud, 1994; Singletary et al., 2002). Feature repurposing is only limitedly possible, as Sun (2012) argues that some features are not revisable due to their recognition of value or their simplicity. Nevertheless, this study agrees on the notion that trying such innovations can evolve and may even leads to optimization of task performance (Ahuja and Thatcher, 2005).

The concept of features in use as well as the components of adaptive system use will not be separately included in the conceptual model of this study in order to keep the model simplistic and understandable. These concepts are included as part of the overall meaning of adaptive system use. Furthermore, it is necessary to introduce these concepts to give readers a basic understanding of adaptive system use and besides that the components of adaptive system use will play a vital role in the analysis section of this study.

Triggers of Adaptive System Use

Accompanied with his introduction of the concept of adaptive system use, Sun (2012) discusses three triggers to be a precondition for adaptive system use, which were first introduced by Louis and Sutton (1991). Their research about switching cognitive gears presented three types of triggers encouraging people to reconsider their way of thinking: 1) novel situations, 2) discrepancies and 3) deliberate initiatives (Table 2). These triggers are essential to fully understand adaptive system use as these conditions can cause active thinking, which is required for users to enable active system use (Sun, 2012; Jasperson et al., 2005; Louis and Sutton, 1991). This study adopts the theoretical perspective stating that contributing to active thinking and possible changing subordinates’ attitude towards features is a precondition for changing behavior does not guarantee change as behavior is greater than attitude alone (Cawsey et al., 2016). Table 2. Types of Triggers (Sun, 2012)

Trigger Definition Examples in System Use

Novel Situations Situations where a person encounters things that are unfamiliar, previously unknown, unique or that appear to be out of ordinary.

- A new task.

- An observation of an unfamiliar feature being used by someone else. - An organization system changes. Discrepancies Situations where an unexpected failure, a

disruption or a significant difference exists between expectations and the reality.

- An unexpected failure of a feature. - The outcome of using a system is different from the expectation. Deliberative initiatives The initiatives one takes in response to a

request for an increased level of attention, when asked to think or while being explicitly questioned.

- A user is asked by his or her supervisor to use system features which are yet unfamiliar to him or

her.

(11)

Novel situations occur when someone is experiencing unfamiliar things (Sun, 2012). This trigger consists of three elements: new tasks (Sun, 2012; Ahuja and Thatcher, 2005; Jasperson et al., 2005), Other’s use (Sun, 2012; Boundreay and Robey, 2005; Compeau and Higgins, 1995) and change in system environment (Sun, 2012; Benamati, Lederer and Singh, 1997) (Table 2). It represents situations where a person encounters things that are unfamiliar, previously unknown, unique or that appear to be out of ordinary (Ahuja and Thatcher, 2005; Jasperson et al., 2005; Benamati et al., 1997; Shaw, 2001; Boudreau and Robey, 2005; Compeau and Higgins, 1995; Ryu, Kim, Chaudhury and Rao, 2005; Sun, 2012). The contradiction of novel situations represents differences between the current and the new situations (Sun, 2012). Besides the somewhat obvious example of a new task, is the observation of an unfamiliar feature that is being used by others a possible way for leaders to stimulate their subordinates to perform adaptive system use behaviors on team level. This study argues that leaders can provoke novel situations not only by demonstrating new features to their team but also by providing their subordinates with tasks that forces them to apply adaptive system use. Nevertheless noticing someone else who is using features outside someone’s own features in use is not enough to trigger adaptive system use, because the individual should experience discrepancies with the individual performance of certain tasks before being triggered to change their behavior (Sun, 2012). Research has elaborated on other examples where managers played a role in provoking novel situations including task overload (Ahuja and Thatcher, 2005; Amabile, 1997) and making modifications of work processes Jasperson et al., 2005), both the create novel situations that awaken subordinates to adaptive system use in order to perform their daily tasks.

Discrepancies are outcomes of system use that differ from what you were expecting (Sun, 2012). People may be motivated to change their behavior, because of discrepancies between their expectations and reality (Hastie, 1984; Louis and Sutton, 1991; Wong and Weiner, 1981). Sun (2012) argues that discrepancies are the most important trigger of ASU due to his findings which confirmed his expectations resulting from the expectation-confirmation theory, which suggest that the disconfirmation of expectations is a salient factor in influencing people’s behavior (Bhattacherjee, 2001; Oliver, 1980; Oliver, 1993; Sun, 2012). Discrepancies represent situations where an unexpected failure, a disruption or a significant difference exists between expectations and the reality (Armstrong and Hardgrave, 2007, p. 456; Sun, 2012). The recognition for the need for adjusting behavior as Louis and Sutton (1991) described as a precondition for adaptive system use returns within discrepancies as Wong and Weiner (1981) have stated that discrepancies are results of the failure to recognize one’s experience in the present cognitive schema. Moreover, discrepancies as in the studies of Louis and Sutton (1991) and similarly Sun (2012) are interchangeable with the concepts of gaps discussed by Lyytinen and Newman (2008) in their theory of alignment among the components of any organization. The contradiction of discrepancies lies within the elements of the current system use activity (Burton-Jones and Straub, 2006). An example by Jasperson et al. (2005) is that a feature does not generate the expected or desired results, which is a contradiction between the feature and the task to be performed.

(12)

supervisors interpretation of the desirable system use activities (Sun, 2012). In a more general term people react to several external causes which may function as antecedents for active thinking (Hastie, 1984; Schön, 1983). In light of the mandatory nature of the deliberative initiatives as described by Sun (2012), it still depends on an individual’s personal effort to change their behavior whether or not this individual adapts to the new situations and changes his or her behaviors. Research confirms this notion of free will that usage intentions can vary of user’s willingness to adopt (Hartwick and Barki, 1994) and perceived usefulness of the features (Bala and Venkatesh, 2016). The interaction between freewill and mandatory actions is also represented in transactional leadership by which leaders influence their subordinates to act within a certain prearranged framework of behaviors (Burns, 1978; Buelens et al., 2011). In organizations where transactional leadership is dominant, employees can choose to operate within the borders of the expected proceedings or resist these (for any reason they might have) only risking internal conflicts and punishment (Janssen, 2003; Burns, 1978). Nevertheless, the trigger deliberate initiative is not included in the scope of this study as the findings of Sun (2012) state that deliberate initiative on itself are “not sufficient to motivate one to revise his/her system use” (Sun, 2012, p. 470). As Sun (2012) demonstrates that deliberate initiative only has an indirect effect through the perceptions of discrepancies, this research chose to focus on the other two triggers of adaptive system use. The study joins the proposition that the presented triggers are positively associated with adaptive system use as a starting point for my argument, because they enable individuals to move beyond existing patterns and create new ways of IT feature interaction.

(13)

Leadership styles

The goal of this study is to uncover the relationship between leadership styles and adaptive behavior in the context of IT in order to better understand adaptive system use. Before discussing the traits of the leadership styles, it seems fit to shortly present the academic looking glass this study is using in order to provide readers with some context.

Throughout literature, many writers have demonstrated that a manager’s effectiveness may be determined by the nature of the organization in which he or she operates as well as by their personal qualities and the nature of their relationship with subordinates (Arnold et al., 2010; Burnes, 1991; Griffin, 2002; Hales, 1986; Nahavandi, 2012; Sjostrand, 1997; Yukl, 2013). This study joins this conception by adopting the perspective of the contingency theory, using the contextual approach on leadership as a looking glass (Burnes, 2014). The contextual approach to effective leadership means that leadership effectiveness depends on the alignment between an individual’s leadership style and the organization in which the leader is operating (Burnes, 2014).

Among the contingency theories involving leadership Vroom and Yetton (1973) were the first to introduce the approach that suggested that leaders can change their behavior from situation to situation. Building on the original approach, Vroom and Jago (1988) later identified five styles of decision making, which differentiated on level of autocracy in comparison to democracy and identified key features of problem situations that leaders might need to consider like the need to resolve conflict or achieve goal congruence (Burnes, 2014). Burns (1978) was to first to introduce the transactional and transformational leadership styles within the context of political science and these were later applied to organizations by Bass (1985: 1995).

(14)

study strives to demonstrate that by combining transactional and transformational leadership behaviors leaders can create situations in which their followers will be rewarded for doing more than is expected of them.

Another reason for combining these leadership styles results from Burnes (2014) his statement on a limitation of Maslow’s pyramid of human needs (1943) in relation to the leadership styles. According to Maslow’s model (1943) there is a certain point in time where employees can no longer be motivated by basic needs anymore. As their needs move up the Maslow pyramid (1943), they are starting to take certain attributes for granted: in the context of the distinctive leadership styles transactional and transformational leadership this means that employees will take their financial compensation for their efforts to the organizational goals for granted and no longer be interested to put in extra effort for this reward. In order to motivate them to continue putting in extra effort, leaders can consider addressing higher levels of the pyramid (Maslow, 1943). In accordance with Vroom and Yetton (1973) leaders can shift from transactional leadership behaviors dominant to the situation to a more transformational leadership style that focuses on internal motivation and inspire the subordinates to the shared vision for the organization. Meanwhile, some researchers have warned that shifting leadership behavior too dramatically and frequently could damage the leader’s credibility and result in followers becoming skeptical towards the leader (Buelens et al., 2011; Cawsey et al., 2016). Nevertheless, balancing transactional and transformational leadership behaviors creates a leadership style that is applicable to a wider range of circumstances.

Transactional leadership focuses on maintaining the status quo (Burns, 1978). It means the leaders are primarily using social behavior exchanges for maximum benefit at minimum effort to lead their subordinates; motivating employees to perform tasks showing their personal responsibilities, goals and knowing the subordinates needs in order to enable rewards for good performance (Chaudhry and Javed, 2012).

(15)

Transactional leadership is defined in terms of three dimensions: 1) contingent reward, 2) active management by exceptions, 3) passive management by exception (Bass, 1990).

Table 3. Components of Transactional Leadership (Bass, 1985: 1999)

Component Definition

Contingent Reward The exchange of rewards for efforts, promises of rewards for good performance and recognition of accomplishments.

Active Management of Exception The process of managers watching and searching for deviations from rules and standards and taking corrective actions.

Passive Management of Exception The process of managers intervening when standards are not met.

The first dimension of transactional leadership contingent reward concerns contact exchange of rewards for efforts, promises of rewards for good performance and recognition of accomplishments (Bass, 1990; Tung, 2016). In light of adaptive system use, contingent reward contributes to the stimulation of organizational learning and performing innovative behavior in the form of adaptive system use. By acknowledging accomplishment of individuals in developing beneficial adaptive system use behaviors, the team is given an example of desirable behavior and motivated to fall in line in order to receive acknowledgement and reward (Chaudhry and Javed, 2012). Another example of stimulating adaptive system use on a more general level is rewarding individual and team’s self-reliance. Being oriented to solve problems hands-on is representing the triggers (novel situations and discrepancies) of adaptive system use. This study expects that by rewarding the overall concept of thinking before asking others, leaders motivate their subordinate to get involved in adaptive system use.

(16)

H1a: Transactional leadership, with its dimensions contingent reward and passive management of exceptions is positively associated with adaptive system use (in all its four components).

H1b: Transactional leadership, with its dimensions active management of exceptions is negatively associated with adaptive system use (in all its components).

Transformational leadership has become the most researched leadership theory in the last decade. This theory discusses changing organizations, transforming the firm following a (new) vision that will lead to the evolution of the organization’s culture (Tichy and Ulrich, 1984). Transformational leadership is often presented as a leadership style that seeks positive transformations ‘in those who follow’ and that achieves desired changes through the ‘strategy and structure’ of the organization (Geib and Swenson, 2013). Transformational leadership is the process of inspiring subordinates to share and pursue the leaders’ vision and motivating others to move beyond their self-interest and work for the aims of the team (Bass, 1985: 1999; Andreessen, Konradt and Neck, 2012, p. 70).

Transformational leadership is composed of four components: 1) idealized influence, 2) inspirational motivation, 3) individualized consideration and 4) intellectual stimulation. The first two components are concerning a leader’s level of charisma, showing the roots of transformational leadership lie in charismatic leadership, which gives meaningfulness to followers by developing deep commitment and providing a sense of moral purpose (Buelens et al., 2011, p.631). The final two components concern a leader’s ability to get subordinates in action.

Table 4. Components of Transformational Leadership (Bass, 1985: 1999)

Component Definition

Idealized Influence The capability of exerting influence by serving as a role model, demonstrating high performance as well as moral standards.

Inspirational Motivation The ability to inspire followers to move them to action by creating a vision and translating this vision so it is aligned with the vision percepted by the team members. Individualized Consideration The ability to connect with his or her subordinates,

understanding their strengths and being able to pull out those strengths to develop and exploit them.

Intellectual Stimulation The ability to stimulate user’s problem-solving skills by challenging them to address old problems using new perspectives by questioning established assumptions and working procedures

(17)

strive towards to a common goal. The components embody the positive correlation between transformational leadership and self-leadership (Andressen et al., 2012; Furtner, Baldegger and Rauthmann, 2013). Self-leadership as ‘the process of influencing oneself’ (Neck and Manz, 2010, p. 4) consists of a skill dimension with different cognitive, affective and motivational-volitional processes in “leading” one’s thoughts and behaviors (Furtner et al., 2013). Basically, this means the process of moving oneself conscious and rational as well as unwittingly thinking and behaving within the boundaries of the pre-established framework of desirable behavior. By personally and actively showing the desirable attitude and behavior themselves, leaders become a role model for their teams, demonstrating the feasibility and logic of achieving the common goal that is expected of them. In the context of this study it is expected that idealized influence being a role model for the team members the leader can stimulate innovative behavior in the form of adaptive system use.

The second component of transformational leadership inspirational motivation like idealized influence concerns the leader’s level of charisma. It covers the leader’s ability to inspire followers into action by creating a vision and translating his vision so it is aligned with the vision of the subordinates or team members (Cho, Park and Michel, 2011; Felfe et al., 2004). Within the context of IT, it enhances user’s confidence in the usage of the system by articulating an appealing vision and expressing high levels of expectation and optimism about the user’s ability to use information system (Cho et al., 2011). Inspirational motivation of transformational leadership in an IT context seems similar to the influence others have on the level of an individual’s computer self-efficacy (Compeau and Higgins, 1995; Porter and Lawler, 1968, Buelens et al., 2011). People need a certain level of confidence for their intrinsic motivation to perform any task, because without a sufficient level of confidence that they will successfully execute a certain task, they won’t participate in performing the task out of anxiety for failure (Buelens et al., 2011). Computer self-efficacy is defined as the individual’s perception of their ability to use IT in the accomplishment of tasks, rather than reflecting on the simple component of skills (Compeau and Higgins, 1995). Inspirational motivation involves scenarios where managers express their trust and confidence in the abilities of their subordinates to successfully perform certain tasks using IT. By doing so, the manager’s attitude affects the subordinate’ computer self-efficacy. In situations where it appears that high expectations are not achieved, discrepancies occur (Sun, 2012). It is important for the job satisfaction of the subordinates that the expected challenging tasks and are not unrealistic, because that might be counterproductive on the intrinsic motivation, computer self-efficacy and overall job satisfaction of the subordinates (Edmonds, Tsay and Olds, 2011; Buelens et al., 2011).

(18)

their internal motivation (Buelens et al., 2011). By giving them the opportunity to develop their skills, while considering the subordinates personal needs and interests, the leader creates a work environment of high involvement (Jasperson et al., 2005; Buelens et al., 2011). Nevertheless, leaders need to be careful that the challenging new tasks are feasible, otherwise it might cause opposite results on job involvement (Janssen, 2003). Unrealistic expectations and overwhelming task distribution may lead to task overload or discrepancies, which is not conducive for the work environment. Nevertheless Sun (2012) argues that these situations will trigger subordinates to adaptive system use in attempt to perform their tasks anyway. Assigning new tasks within the context of individualized consideration may trigger adaptive system use.

The last component of transformational leadership is intellectual stimulation (Cho et al. 2011; Felfe et al., 2004). Intellectual stimulation concerns leaders stimulating user’s problem-solving skills by challenging them to address old problems using new perspectives by questioning established assumptions and working procedures (Cho et al., 2011; Felfe et al., 2004). They stimulate others to be creative, take risks and solicit new ideas and new ways to perform tasks, without publicly correcting or criticizing them (Stewart, 2006; Felfe et al., 2004). Questioning assumptions and working procedures can be done by researching where things could be improved, wondering where the discrepancies between the expected and/or desired outcomes in comparison to the reality. Leaders can stimulate active thinking and even reward the development of new manners of performing working procedures in their extended role of a transformational as well as a transactional leader (Bass, 1999). Questioning assumptions might help to make the switch to cognitive gears and by discussing the discrepancies as a team, the team can stimulate innovative behavior and result in adaptive system use on the scale of teamwork instead of on the individual level.

H2: Transformational leadership, with its dimensions idealized influence, inspirational motivation, individualized consideration and intellectual stimulation, is positively associated with adaptive system use (in all its four components).

Personal Innovativeness in IT

(19)

with a relevantly higher level of personal innovativeness in IT are more likely to recognize emerging opportunities and tend to take greater risks in unstructured situations (Kirton, 1976). Therefore, this research expects that personal innovativeness in IT moderates the influence of leadership positively on adaptive system use, resulting in the following hypotheses:

H3: Personal innovativeness in IT positively moderates the influence of transactional leadership on adaptive system use (in all its components).

H4: Personal innovativeness in IT positively moderates the influence of transformational leadership on adaptive system use (in all its components).

Conceptual model

This section is finalized by the presentation of the conceptual model. This study strives to elaborate on the influence of leadership on ASU. As demonstrated above, there is enough information presented in past research for the presumed influence of leadership on ASU.

By doing so, this study contributes to the research field on the adaptation phase of IT within the research area of IT-enabled organizational change. Besides, it contributes to the academic research fields of IT and leadership studies. By elaborating on the possible effects of leadership on ASU, this study creates a deeper understanding of the complex process concerning post-adoptive IT implementations. This study might give researchers new perspectives on the approach of transactional and transformational leadership styles in an IT context and give researchers new insights on ways how to improve the adaptation of new IT.

(20)

Methodology

This section discusses the general preparation of the data and the data collection. The methodological components of this study are presented in subsections, including data collection, sample description and measurements.

This study adopts a quantitative research methodology, because as the theoretical background discussed the importance of adaptive system use and the effect of leadership on other important concepts in literature, the next logical step is to investigate whether leadership has influence on adaptive system use as well. The benefit of studying adaptive system use in this manner is that it gives the possibility to gain a deeper understanding about the concepts and maybe even expand the model of Sun (2012).

Data Collection

In this study, a survey instrument was assembled. The final draft version was presented to two subjects to verify whether statements from the questionnaire were self-explanatory and check how long it would take to complete the questionnaire. After receiving feedback, some explanatory text was added to introduce the different topics within the questionnaire. Furthermore, the average duration of answering the questionnaire was set on ten minutes. Afterwards, the improved pilot version was presented to two other subjects to verify whether the adjustments of the first pilot delivered the desired results and assisted the researcher to adjust the formulation regarding certain statements.

The questionnaire was digitally available both on desktop computer and portable devices in English and Dutch. The English version is based on earlier research, which will be elaborated on in the presentation of the measurements. The Dutch translation of the survey statements are translated by the researcher and evaluated in both evaluation rounds to ensure the statements cover the essence of the concepts just as well as the English version of the questionnaire. The primary reason for having both English and Dutch versions is that the main language of a substantial amount of the approached organizations is Dutch. Nevertheless, this study is also applicable on an international scale, as most participating organizations are active in international markets. The finalized questionnaire is added (see appendix A) and is to be available for future research.

The survey instrument was distributed to organizations that use IT on a daily basis. In light of this, this study considers not only the information systems that are provided to the respondents by their employers, but also all information technologies (for example Microsoft Office and the internet browsers) that may contribute to performing tasks. The broad definition of IT used in this study corresponds with the research methods of Sun (2012) as he focused on the features within Microsoft Office while introducing ASU. Despite the broad definition the usage of IT on a regular basis is a requirement for participants to be relevant for the sample.

(21)

Furthermore, I used my personal network to organize participation. From the 47 national and international active organizations that were contacted with general information int

researcher and the research goal, asking for collaboration, four large organization

enthusiastically to participate on a large scale through higher management. The department managers that functioned as contact person in these organ

anonymous, re-usable link and example content for an introductory email about the research topic and a reminder email. These organizations all met the daily IT

differ from industry sector (IT: 304 potent

participants; Public Services: 84 potential participants and Legal Services: 27 potential participants). Finally, family members and friends who are working for smaller organizations were asked for participation (37 potential participants; primarily active in the IT sector).

Sample Description

In the end, this led to a total representation of 129 responses, resulting in a response rate of 16.8%. However, 12 questionnaires were deleted, due to that they we

less than three minutes which seems unrealistic considering the number of items or were fragmentary completed. Thus, the final response was 117 valid responses (15.4% of the total potential number of participants). The demographi

presented in the table below.

Table 5. Demographic characteristics of the Sample Variable Sample composition

Gender Male Female Response Personal link

Anonymous link

Age Mean = 38 years; Std. dev. = 11 years ; range 18 The distribution of responses with regards to market sector majority of the participants are represente

33% 9%

my personal network to organize participation. From the 47 national and international active organizations that were contacted with general information int

researcher and the research goal, asking for collaboration, four large organization

to participate on a large scale through higher management. The department managers that functioned as contact person in these organizations were provided with an usable link and example content for an introductory email about the research topic and a reminder email. These organizations all met the daily IT-usage requirement and differ from industry sector (IT: 304 potential participants; Financial services: 87 potential participants; Public Services: 84 potential participants and Legal Services: 27 potential participants). Finally, family members and friends who are working for smaller organizations

ipation (37 potential participants; primarily active in the IT sector).

In the end, this led to a total representation of 129 responses, resulting in a response rate of 16.8%. However, 12 questionnaires were deleted, due to that they were either completed within less than three minutes which seems unrealistic considering the number of items or were fragmentary completed. Thus, the final response was 117 valid responses (15.4% of the total potential number of participants). The demographic characteristics of the sample population are

Table 5. Demographic characteristics of the Sample

Sample composition N=117

70 47 69 48 Mean = 38 years; Std. dev. = 11 years ; range 18-65 years

with regards to market sector is presented in the figure below. The represented in the IT sector.

34% 8% 12% 9% 4% IT Consultancy Fast Moving Consumer Goods Financial services Government None of the above

my personal network to organize participation. From the 47 national and international active organizations that were contacted with general information introducing the researcher and the research goal, asking for collaboration, four large organizations responded to participate on a large scale through higher management. The department izations were provided with an usable link and example content for an introductory email about the research usage requirement and ial participants; Financial services: 87 potential participants; Public Services: 84 potential participants and Legal Services: 27 potential participants). Finally, family members and friends who are working for smaller organizations

ipation (37 potential participants; primarily active in the IT sector).

In the end, this led to a total representation of 129 responses, resulting in a response rate of re either completed within less than three minutes which seems unrealistic considering the number of items or were fragmentary completed. Thus, the final response was 117 valid responses (15.4% of the total c characteristics of the sample population are

In % 59.8% 40.2% 58.9% 41.1%

(22)

Measurements

The measurements and scales are established from existing valid scales of literature and adjusted to fully cover the essence of the probable correlation between transformational leadership and ASU. The questions within the survey are presented in the format of statements and the phrasing is adjusted to allocate to the context of IT. The Likert-scale (from extremely disagree to extremely agree) is used throughout the whole questionnaire to establish uniformity (Blumberg, Cooper and Schindler, 2011). The construct validity is ensured by using only established research methods used in existing literature. Therefore, the validity and reliability is tested by other studies that used the same questions. The reliability is ensured by that all respondents receive the same level of explanatory assistants participating in the research and that all respondents fill in the same questionnaire.

Dependent variable. The dependent variable is adaptive system use. The questions concerning ASU are based on the questionnaire that Sun (2012) assembled. The questions are put in a grammatical past perfect tense. Each dimension is represented with several questions to ensure the reliability in case a question might need to be removed from the analysis.

Independent variables. The independent variables are: transactional leadership and transformational leadership. The measurements for transactional and transformational leadership are covered by the Multiple Leadership Questionnaire (Avolio and Bass, 1997: 2008). The statements of the questionnaire are a little adjusted to fit within the perspective of the subordinates in contrast to the original Multiple Leadership Questionnaire that follows statements through the perspective of the leader (evaluating their own leadership style). The adjusted perspective of the questionnaire allows a broader applicability of the questionnaire, by facilitating subordinates to judge the leadership style of their manager while working in teams.

Moderating variable. The moderating variable is personal innovativeness in IT (Agarwal and Prasad, 1999; Agarwal and Karahanna, 2000). Personal innovativeness in IT is included in the questionnaire by using the questions out of the original measurement tool about personal innovativeness on IT by Agarwal and Prasad (2000). The validity of the measurements are ensured by utilizing all four questions covering personal innovativeness in IT, so the moderating effect can still be measured in case of elimination of questions in case that questions might need to be removed from the analysis. Personal innovativeness in IT is commonly used as a moderating variable in relation to ASU.

(23)

Analysis and Results

This section presents a description of the performed analysis and the results of the research. The conducted steps are elaborated to show the process of conducting this research. Thereafter, the results will be discussed addressing the hypotheses introduced in the theoretical background. Analysis

For the analysis, I used IBM SPSS Statistics 20 to conduct an exploratory factor analysis (EFA) with a direct oblimin rotation. The exploratory factor analysis is intended to identify information about the constructs different items and the corresponding scores (Hinton, 2004). For the factor analysis the following criteria were applied: 1) each measure must have a minimum loading of 0.4; 2) each measure must not have a loading in more than one factor and 3) each measurement must load into the correct factor based on the predetermined components of the constructs (Song, Im, Van der Bij, Song, 2011). The final results of the EFA are presented in Table 6 (ASU) and Table 7 (Leadership styles and PIIT). An explanation for having to remove several items from the EFA is that the dimensions are quite similar which makes it hard to divide them properly in the EFA.

Table 6. Explanatory factor analysis loadings (Adaptive System Use)

Feature Repurposing Feature Combining Trying New Features Feature Substitution FR_3 FR_2 FC_2 FC_4 TR_1 TR_2 FS_3 FS_1 .912 .768 .159 .253 -.102 .873 .802 -.193 .124 .169 -.184 -.132 .274 .780 .771 .343 .150 -.296 -.867 -.522

Table 7. Explanatory factor analysis loadings (Leadership styles and PIIT) CR Passive

Man. of Exc.

Active Man. of Exc.

(24)

Table 8 presents the descriptive statistics and the Cronbach’s alphas. The reliability analyses and Cronbach’s alphas were established to ensure reliability of the provided measurement scales.

Table 8. Descriptive Statistics including Cronbach’s alphas Mean

(Standard

Deviation) 1 2 3 4 5 6 7 8 9 10 11 12

1. CR

2. Passive Man. Of Exc. 3. Active Man. of Exc. 4. TLIS

5. TLIC 6. TLIM 7. TLIF 8. PIIT

9. Trying New Features 10. Feature Substitution 11. Feature Combining 12. Feature Repurposing 4.70 (1.19) 3.99 (1.53) 4.99 (1.19) 5.17 (1.00) 5.29 (1.15) 5.35 (0.99) 5.65 (0.93) 5.29 (1.02) 5.74 (0.73) 5.33 (0.96) 5.47 (1.20) 4.56 (1.26) α:.74 .032 .635** .236* .471** .164 .380** -.080 .004 -.062 -.111 -.219* α: n/a .104 -.123 -.013 -.015 -.089 .076 -.136 -.109 .005 .029 α: n/a .276** .435** .163 .253** -.143 .106 -.035 -.081 -.125 α: .79 .278** .399** .481** -.080 .068 .063 .077 -.079 α: n/a .339** .274** -.126 .097 .038 -.086 -.115 α: n/a .424** -.006 .117 0.15 .205* .087 α: .70 .037 .092 -.008 -.053 -.099 α: .73 .379** .519** .274** .370** α: .75 .465** .384** .407** α: .67 .397** .534** α: .76 .558** α: .80 *. Correlation is significant at the 0.05 level (2-tailed).

(25)

The descriptive statistics (see Table 8) show that the components of transformational leadership are correlated on a significant level of 0.01. The same applies to the components of adaptive system use. These results are logical considering that the components are representing the same variable and are established by literature. The results of the reliability analysis indicate that all multiple-item constructs have good reliability scores between 0.67 and 0.80 (Hinton, 2004; Song et al., 2011).

(26)

Table 9a. Results from Hierarchical Regression Analyses Model 1 Coefficient Estimate (Standard Error) Model 2 Coefficient Estimate (Standard Error) Model 3 Coefficient Estimate (Standard Error) Contingent Reward (CR)

Passive Management of Exception (MBE_Pas) Active Management of Exception (MBE_Ac) Intellectual Stimulation (TLIS)

Individualized Consideration (TLIC) Inspirational Motivation (TLIM) Idealized Influence (TLIF

Personal Innovativeness in IT (PIIT) PIIT * CR PIIT * MBE_Pas PIIT * MBE_Ac PIIT * TLIS PIIT * TLIC PIIT * TLIM PIIT * TLIF Age Gender New Task Other’s Use

Changes in System Environment Discrepancies F value R square -0.62 (.057) -1.00 (.138) .180 (0.78)** .029 (.071) .009 (.083) .244 (.073)*** 3.498 .161 -.108 (.083) -.070 (.045) .113 (0.74) -.063 (.078) .079 (.069) .076 (.081) -.019 (.091) -.066 (.059) -.131 (.143) .176 (.081)** .007 (0.76) .064 (.087) .263 (.077)*** 2.255 .223 -.069 (.077) -.065 (.041) .126 (.067)* -.013 (.076) .073 (.069) .032 (.076) -.014 (.090) .311 (.075)*** .002 (.066) .094 (.041)** -.203 (.072)*** .003 (.098) .051 (.088) -.029 (.081) -.031 (.090) -.002 (.058) .026 (.135) .177 (.083)** .107 (.070) .008 (.083) .121 (.074) 3.682 .451 Dependent Variable: Trying New Features; N = 117.

* Significant at p < .10 (2-tailed test) ** Significant at p < .05 (2-tailed test) *** Significant at p < .01 (2-tailed test)

(27)

Table 9b. Results from Hierarchical Regression Analyses Model 1 Coefficient Estimate (Standard Error) Model 2 Coefficient Estimate (Standard Error) Model 3 Coefficient Estimate (Standard Error) Contingent Reward (CR)

Changes in System Environment Discrepancies F value R square -.044 (.075) -.177 (.180) .164 (.101) .130 (.093) .072 (.108) .226 (.096)** 2.807 .134 -.035 (.109) -.084 (.059) -.052 (.097) .018 (.103) .102 (.091) .093 (.107) -.125 (.120) -.065 (.077) -.207 (.188) .169 (.107) .105 (.101) .113 (.114) .249 (.102)** 1.762 .183 .045 (.104) -.066 (.056) -.040 (.090) .028 (.103) .072 (.094) .048 (.103) -.083 (.122) .402 (.102)* -.103 (.090) .063 (.056) -.107 (.098) .175 (.133) .071 (.119) -.051 (.110) -.177 (.121) .006 (.079) .014 (.182) .193 (.113)*** .245 (.095)** .073 (.112) .076 (.101) 2.874 .391 Dependent Variable: Feature Substitution; N = 117.

* Significant at p < .10 (2-tailed test) ** Significant at p < .05 (2-tailed test) *** Significant at p < .01(2-tailed test)

(28)

Table 9c. Results from Hierarchical Regression Analyses Model 1 Coefficient Estimate (Standard Error) Model 2 Coefficient Estimate (Standard Error) Model 3 Coefficient Estimate (Standard Error) Contingent Reward (CR)

Changes in System Environment Discrepancies F value R square -.059 (.092) -.339 (.220) .286 (.124)** .218 (.114)* .236 (.133)* -.041 (.118) 4.442 .196 -.192 (.129) .030 (.070) -.011 (.115) .134 (.122) -.023 (.108) .299 (.126)** -.191 (.142) -.074 (.092) -.341 (.222) .201 (.127) .271 (.119) .298 (.135)** -.003 (.120)** 3.308 .297 -.075 (.127) .062 (.068) -.019 (.110) .142 (.126) -.110 (.115) .260 (.126)** -.086 (.149) .320 (.125)** -.154 (.109) .028 (.068) -.047 (.119) .098 (.163) .219 (.146) -.110 (.134) -.275 (.148)* -.008 (.096) -.133 (.223) .238 (.138)* .414 (.116)*** .196 (.137) -.159 (.123) 3.487 .438 Dependent Variable: Feature Combining; N = 117.

(29)

Table 9d. Results from Hierarchical Regression Analyses Model 1 Coefficient Estimate (Standard Error) Model 2 Coefficient Estimate (Standard Error) Model 3 Coefficient Estimate (Standard Error) Contingent Reward (CR)

Changes in System Environment Discrepancies F value R square -.238 (.101)** -.375 (.243) .075 (.137) .195 (.147) .049 (.147) .133 (.130) 2.522 .122 -.267 (.143)* .055 (.077) .023 (.128) -.080 (.136) .009 (.119) .290 (.140)** -.139 (.157) -.235 (.102)** -.304 (.247) .004 (.141) .263 (.132)** .134 (.150) .203 (.134) 2.220 .221 -.179 (.131) .060 (.070) .041 (.114) -.034 (.129) -.002 (.118) .239 (.130)* -.165 (.153) .657 (.128)*** -.019 (.113) .199 (.070)* -.111 (.123) .205 (.168) .057 (.150) -.154 (.138) -.140 (.153) -.112 (.099) .059 (.230) -.012 (.142) .424 (.120)*** .090 (.141) -.054 (.127) 3.876 .464 Dependent Variable: Feature Repurposing; N = 117.

(30)

Results

Examining the content of four versions of Table 9, the following statements can be made concerning the predetermined hypotheses.

First of all, contingent reward is only significant (p < 0.1) in relation to feature repurposing (1/4). As passive management of exception is not significant to any of the components of ASU (0/4), therefore there is not enough empirical evidence to support H1a. Secondly, active management of exception is positively significant (p < 0.1) in relation to trying new features (1/4), and has no significant relation to the other components of ASU. Surprisingly, Table 9a shows a positive relation between active management of exception instead of the predicted negative relation. Besides that there is substantial empirical evidence to confirm a relationship between active management of exception and ASU, H1b is rejected. Future research could explore the determined positive relationship between active management of exception and ASU. Thirdly, regarding the association between transformational leadership and adaptive system use, inspirational motivation is the only component that is positively significant (p < 0.05) (both in Model 2 and Model 3) in relation to feature combining and feature repurposing (2/4). The other components of transformational leadership (intellectual stimulation, individualized consideration and idealized influence) are not significant to any of the components of ASU (0/12). Therefore, there is not enough empirical evidence to support H2.

Table 9a shows that the interaction between both passive and active management of exception with personal innovativeness in IT is significant (p < 0.05 and p < 0.01). In addition to this is the interaction between passive management of exception and PIIT is also significant in the model of feature repurposing (see Table 9d). Unfortunately, the moderating effect of PIIT does not seem to apply to the contingent reward as this interaction in not significant in any of the models. Therefore, there is only adequate empirical evidence to partially confirm H3.

(31)

Discussion

The main purpose of this research is to better understand adaptive system use by further examining how transactional and transformational leadership contribute to ASU in order to decrease underutilization of IT. In this final section the results will be discussed in relation to the research questions presented in the introduction. Furthermore, the implications regarding literature and practice are presented alongside the limitations of this study.

This section starts with the review of the first research question: “How do transactional and transformational leadership enable teams towards adaptive system use?”

First of all, transactional leadership positively influences ASU through active and passive management of exception. The positive effect of active management of exception on ASU can be explained by the Hawthorne studies. The Hawthorne studies showed that employees put more effort into their work when they felt like they were being involved in decision making and being monitored whilst performing their tasks (Buelens et al., 2011). Active management of exception considers managers actively searching for deviations in order to take corrective actions (Bass, 1985: 1999), which could also apply to performing adaptive system use under the condition that performing adaptive system use is customary and anticipated behavior in the organization. This would mean that within the organization it is expected to actively search for better ways to perform their tasks. For example to implement adaptive system use within the performance measurement of the employees by rewarding employees that take an active role in improving the current working methods.

The positive effect of passive management of exception on adaptive system use can be explained that individuals respond differently to the same approach, depending on their personality and social position in the group dynamic (Belbin, 2010). Passive management of exception is only significant through the moderating effect of PIIT, which also is a personal trait, making it logical to assume that other personality traits might also influence the relation between transactional leadership and ASU. Nevertheless, this assumption needs to be tested in future research.

Secondly, transformational leadership positively influences ASU through inspirational motivation and idealized influence.

The results show that inspirational motivation is positively associated with feature combining and repurposing. This is in line with the expectations presented in the theoretical background, which suggested that leaders showing confidence in the subordinates’ ability to use IT is positively associated with ASU (Cho et al., 2011). Furthermore, it is not surprising that inspirational motivation influences feature combining and feature repurposing positively, because both components belongs to revising the spirit of features in use (Sun, 2012). This implies that subordinates change the manner in which they use features they were already accustomed with and all that they needed to apply ASU was a confidence boost from their manager to start experimenting.

(32)

Finally, as the discussion above demonstrates PIIT has a positive moderating effect on the relations of transactional and transformational leadership on ASU. Thereby, the third research question “How does personal innovativeness in IT moderate the effect of leadership on adaptive system use?” has been answered accordingly. In summary, the results confirm the expected contribution that PIIT makes concerning the relation between leadership and ASU.

The findings demonstrate that transactional leadership is more significant to adaptive system use than transformational leadership. Looking at the number of significant relations in the result section, it is striking to see that transactional leadership is in more models significant than transformational leadership. “Which of these two leadership styles is the most effective in triggering adaptive system use?”

A reason for transactional leadership turning out to be more influential on ASU than transformational leadership could be rooted in the basic studies of human needs and the underlying concept changing behavior that Louis and Sutton (1991) introduced. Louis and Sutton (1991) state that in order for someone to change their behavior there needs to be some sort of discrepancies or contradiction between the current circumstances and the desirable situation. Or in other words, employees need external motivation in order for them to change their behavior. Transactional leadership is centered around controlling the external motivational factors in the working environment (e.g. reward and punishment) as opposed to transformational leadership which focuses on internal motivation (e.g. actively demonstrating desirable behavior). As the theoretical background elaborated on, transformational leadership is positively associated with innovative behavior. Nevertheless, these results show that transformational leadership does not have a significant influence on ASU. A possible explanation could be that ASU is more complex and comprehensive than innovative behavior. This makes it not only difficult to establish reliable statistical correlations, but this is also reason to lay the foundation for future research.

Control variables. The control variable age was negatively significant (p < 0.05) in the first two steps of Table 9d considering feature repurposing. These results suggest that the higher the age, the less likely it is that recipients apply feature repurposing behaviors. This corresponds with the common statement which states that young employees are more likely to show innovative behavior than older employees. The results show that gender is not significant in any of the models.