E
MPLOYEES PRE
-
IMPLEMENTATION EFFORT
EXPECTANCY AND COMMITMENT TO CHANGE OF CLOUD
COMPUTING
:
THE MODERATING EFFECT OF TRAINING
EVALUATION
Master thesis, Master of Science Business Administration: Change Management University of Groningen, Faculty of Economics and Business
June 23th 2014
2
Content
1. Introduction ... 3
2. Theory and hypotheses ... 6
2.1. IT adoption, acceptance and usage ... 6
2.2. Change commitment ... 7
2.3. Effort expectancy ... 11
2.4. End-user training evaluation ... 12
2.5. Conceptual model ... 14
3. Research Methodology ... 14
3.1. Research setting ... 14
3.2. Data collection and sample ... 15
3.3. Measurement ... 16
3.3.1. Independent variables ... 16
3.3.2. Dependent variables ... 17
3.3.3. Control variables ... 17
4. Analysis & results ... 18
4.1. Factor analysis... 18
4.2. Descriptive statistics ... 18
4.3. Linear regression analyses ... 20
4.4. Sobel test for partial mediation ... 20
4.5. Control variables ... 21
5. Discussion ... 22
6. Conclusion & Managerial implications ... 24
7. Research limitations & future research ... 26
8. References... 27
Appendix I – Coding table & Survey questions ... 30
Appendix II – Factor analysis ... 31
Appendix III- Sobel test ... 32
3
Abstract
Research on individual outcomes of IT implementation is relatively scarce. Although a lot of research has been done on the antecedents of IT user acceptance, there is less known about the implications of these implementations for the individual. Therefore this study introduces the change commitment model as more refined and complex explanation of behavioural intention. Change commitment will be measured in terms of affective commitment, normative commitment and continuance commitment. Therefore this study focuses on the effects of effort expectancy and training evaluation on the different change commitment components. In this study it was found that affective and normative commitment are positively related towards effort expectancy. The effect of training evaluation was only significant as a moderator variable for affective commitment.
Keywords: change commitment, information technology (IT), Google Apps for Education, cloud computing, effort expectancy, training evaluation.
---
1.
Introduction
As enterprise spending on cloud computing is speeding up from a 78.2 billion dollars in 2011 to a 235 billion dollars expected in 2017, cloud computing has become serious business (Muldoon, 2014). The most important advantage cloud computing has over traditional data storage is that prices are on average 70% lower. This enables enterprises to spend their budgets more strategically on information technology (IT). Thereby, cloud computing systems are flexible and setting up takes rather weeks than months. This trend is not only visible in business but it can also be found at institutional organizations such as Universities. Some researchers are even talking about cloud computing for education as ‘a new dawn’ (Sultan, 2010). Hence the questions remains, how can organizations ensure themselves that they are making the best out of these IT systems?
4 important. When employees do not feel comfortable using these systems, the implementation possibly can end up in a failure (Vella et al., 2013). With managers focussing only on technological aspects of IT systems, chances exist that users will be avoiding them or even creating workarounds which eventually leads to a significantly lower efficiency than intended. In literature this phenomenon is also called the productivity paradox (Venkatesh, 2000). Over the last years many research has been conducted on the acceptance of technology (Davis, 1989; Ajzen, 1991; Moore & Benbasat, 1991). One of the most known paper comes from Venkatesh et al (2003). They provided an extensive framework with different antecedents of user acceptance of IT systems. In this construct user acceptance has been explained by behavioural intention and facilitating conditions. These variables explain 60-65% of the variation in user acceptance (Venkatesh, 2008). Although the research field is already very mature there remain some challenges. One of them is the lack of theory on the effects of IT implementation on individual users (Venkatesh et al., 2003).
Change commitment is a construct which focuses on the different kinds of commitment that individuals can pertain towards a certain change situation (Meyer et al., 2002). So the focus is not on whether people accept or adopt a certain change but on what grounds they do adopt it. Research showed that the change commitment construct has been related to behavioural outcomes such as productivity (Meyer & Allen, 1997). Also the change commitment model does not only focus on mere acceptance since they make the distinction between cooperative and championing behaviour. Championing behaviour means that people will do their utmost best to make the change a success due to their inherent beliefs in this change (Herscovitch & Meyer, 2002).
One of the antecedents of change commitment is self-efficacy. People who pertain higher self-efficacy will have less difficulties adapting to a certain change (Herold et al., 2007). Whether people feel capable towards a new IT system also has been researched by Venkatesh et al. in terms of effort expectancy (2003). It would therefore be more interesting to combine these research streams in order to see whether they are complementary to each other.
Therefore this study aims to unravel what different intentions or dispositions people can have regarding the implementation of a new IT system and how these can be influenced by antecedents of IT user acceptance. This study thereby tries to open up the black box of individual responses to change due to implementation of a new IT system (Venkatesh et al., 2003).
5 opening the black box of individual user responses towards a new IT system. And secondly it introduces an existing model from change theory into the IT research stream. Thirdly, research on the antecedents of change commitment is broad and underdeveloped (Becker et al., 2009). And lastly in contributes to the research on employee adoption of cloud computing in educational settings (Sultan, 2010).
Data for this research was drawn from a training group of employees within a University. Currently the University is making a shift from Oracle towards Google apps for Education as the provider for email and agenda functions. Google Apps for Education (GAfE) is a free web-based email, agenda and data storage application for higher education. With GAfE it is possible to work in the cloud which can make data sharing more easily. Therefore GAfE is expected to improve collaboration between students and lecturers (Cahill, 2012). Primary results showed that when GAfE is used efficiently and instruction tools are being provided, the system has high benefits (Cahill, 2012). At the CIT of the University it is believed that for employees of the University the biggest change will be at the agenda application. Possibilities in administering meetings will be broadened. And from there on, people who do not have a University account can be invited to meetings in the application, also they can make changes themselves.
This paper is organized as follows. After this introduction the research question will be presented. Secondly, an overview of the theory on user acceptance, change commitment and end-user training will be provided. Also a conceptual model will be provided. Thereafter research methodology will be explained. Lastly the findings of this research will be discussed in the conclusion and discussion. Also limitations and possible directions for further research will be discussed.
As stated before the aim of this study is to provide insights on the influence of training evaluation on change commitment during IT implementation. Therefore the research question will be;
6
2.
Theory and hypotheses
2.1. IT adoption, acceptance and usage
When an IT system is being implemented employees evaluate whether they will accept or reject it, based on the associations they have of the system (Kim & Kankanhalli, 2007). This association can be subdivided into three considerations towards the change; behavioural beliefs, normative beliefs and control beliefs (Kim & Kankanhalli, 2007). Probably the most well-known IT user acceptance model is Venkatesh’s et al. Unified Theory of Acceptance and Usage of Technology (2003). With the UTAUT Venkatesh et al. provided an extensive model on the antecedents of user acceptance of IT (2003). The usage of technology in this model is influenced directly by behavioural intention and facilitating conditions. Indirectly and through behavioural intention, usage of technology is influenced by; performance expectancy, effort expectancy and social influence (2003). In this study user acceptance was measured low-high continuum.
Recently, Van Offenbeek et al. (2012) showed that acceptance and resistance cannot be considered as two opposites. By combining usage and acceptance/resistance theory they propose a two-factor model (2012). Acceptance of IT systems is defined on a high-low continuum of usage as done in the studies of Venkatesh & Davis (2000). Resistance has been defined as supportive or resistant behaviour as explained in the studies of e.g. Markus (2004). This two factor model shows that although people accept a technology they can still be non-users and vice versa. Therefore this model provides a more refined view on user acceptance.
Another more complex view on user acceptance comes from Beaudry & Pinsonneault (2005). They distinguish between two streams of research within IT. The first stream concerns variance research which comprises the antecedents that lead to IT acceptance and usage (Venkatesh et al., 2003). The second stream comprises process research which is more interested in the process of adaptation behaviours (Orlikowski, 2007). Especially the process research paid attention to the complex character of user acceptance. The Coping Model of User Acceptance (CMUA) tries to bridge the gap between these two approaches (Beaudry & Pinsonneault, 2005). CMUA states that there are four adaptation strategies; benefits maximizing, benefits satisficing, disturbance handling, and self-preservation.
7 the upcoming paragraphs the change commitment construct is introduced in order to contribute to this discussion by giving a different view on IT user acceptance.
2.2. Change commitment
Change commitment is considered to be one of the most important factors with respect to change successes (Herscovitch & Meyer, 2002). Due to its three-component model it became easier for researchers to make more precise predictions of the impact of change commitment upon change behaviour (Herscovitch & Meyer, 2002). The upcoming paragraphs will explain the theoretical groundings of change commitment. Moreover it gives the definitions of the different change commitment components and it connects change commitment to the conceptual model of this study.
In literature there exist two ways of thinking about change. First, there is overcoming problems and fears. Second, there is a way to see change as an opportunity to improve and motivate people to work harder and better (Bouckenhooghe, 2010). Change commitment can be related to the latter since it regards change as an opportunity instead of a threat (Bouckenhooghe, 2010). The theoretical grounding of change commitment can be found in the Theory of Planned Behaviour (TPB)(Bouckenhooghe, 2010). The TPB states that people’s intentions and behaviours are shaped by two variables. First, the costs and benefits associated with the change. And second, the pressure to follow the norm. Moreover they state that people’s intentions are determined by their feelings and thoughts, social pressure and self-efficacy (Bouckenhooghe, 2010). Change commitment finds it origins in the organizational commitment literature. Meyer & Allen (1997) provided a model of organizational commitment which would be later extended to a construct of change commitment by Herscovitch & Meyer (2002). The research on organizational commitment started with the question on what volunteers made so committed to non-profit work and their organization. Moreover the researchers wanted to find out what makes people committed to a certain kind of action and how former experiences contribute to this commitment (Meyer & Allen, 1997).
8 As stated earlier change commitment consists out of three components, which allows researchers to make more precise predictions on change behaviour. These three components, continuance, normative and affective commitment, will be discussed later in more detail.
With respect to change behaviour, researchers found that commitment to change is a good predictor of behavioural support for change (Herscovitch & Meyer, 2002). Moreover change commitment displays the willingness to work on behalf of a successful implementation of a change. Also it leads to proactively supporting the change by making it a success (Batistelli et al., 2014; Fedor et al., 2006). But most and foremost change commitment construct goes beyond a favourable disposition towards change (Bouckenhooghe, 2010). Even though it was found that commitment to change leads to behavioural support, the relationship between affective commitment, normative commitment and continuance commitment seems to be more complex that initially thought. For this reason the three different change commitment constructs should not be considered apart from each other since it is the combination which makes the construct relevant (Meyer & Allen, 1997).
2.2.1. Affective commitment
Affective commitment can be defined as; ‘supporting the change initiative based on the belief of the inherent benefits of the change initiative’ (p. 475, Meyer & Herscovitch, 2002). Meyer & Allen (1997) state that affective commitment is related to emotional attachment of the employee, the identification with the organization and involvement with the organization. Also Neubert & Cady (2001) demonstrated that affective commitment to a new program initiative was positively related to the level of participation in the program and program-relevant performance. Possible antecedents of affective commitment are; personal characteristics, work experiences (Meyer et al. 2002).
2.2.2. Normative commitment
9 accept the change because they feel that they ought to accept the change (Meyer & Allen, 1997). Possible antecedents of normative commitment are; personal characteristics, socialization experiences and organizational investments (Meyer et al. 2002).
2.2.3. Continuance commitment
Continuance commitment is described as the perceived costs of not complying with the proposed change (Cunningham, 2006). People who engage in continuance commitment do so because they need to and have no other options(Meyer & Allen, 1997). Therefore continuance commitment is best defined by stating that it is the individual who is weighing its other options. Possible antecedents of continuance commitment are; personal characteristics, alternatives and investments (Meyer et al. 2002).
2.2.4. Antecedents of change commitment
Recently, Batistelli et al. found that besides the existing antecedents that concerns about the change itself influence change commitment (2014). They defined concerns about change as; an individual’s appraisal of how the change will affect his/hers work (p. 952; Batistelli et al., 2014). Concerns about change can be divided into; concerns about the content of the change, concerns on the benefits of the change, and concerns about mastering the change (Batistelli et al., 2014). Moreover, the antecedents of change commitment can be divided into situational variables and change specific variables (Herold et al., 2007).
The most important antecedent in this study is self-efficacy. Armenakis et al. state that this is one of the five components for change to be effective (1999). Hence self-efficacy can be defined as; ‘’judgments about how well one thinks they can execute courses of action required to deal with prospective situations’’ (p. 122 Bandura, 1982). In literature there is some inconvenience about the concept of self-efficacy. Not all authors tend to make the distinction between general self-efficacy and self-efficacy towards a certain change. As stated before, there exists a difference between situational and change specific antecedents of commitment to change. Due to these different findings, this distinction is worth noticing. Also these findings are supported by the research of Venkatesh who found that effort expectancy and general self-efficacy are conceptually distinct (2000).
10 and affective commitment (2002). Also, Li & Whang found that knowledge self-efficacy, defined as the self-assessment of the ability to provide knowledge to others, was positively related to affective commitment and continuance commitment (2014). They found that high levels of self-efficacy encourage employees to contribute to and use IT systems, which relates to affective and normative commitment. Neves stated initially that self-efficacy would be related to commitment since it enhances the ability to cope with the change more easily (2009). However he found that the relationship between self-efficacy and affective commitment was not significant (Neves, 2009). Contrary Herold et al. found a stronger relationship between self-efficacy and change commitment in general (2007). However, Herold et al. used the four-component Caldwell et al. (2004) construct whereas Meyer et al. used the Herscovitch 18-component construct. Therefore this effect might be due to a difference in measurement. Also reasons for the mixed findings can be related to the effect of other contextual factors such as change characteristics (Neves, 2009).
2.2.5. Commitment and behavioral outcomes
Researchers found that commitment to change is related to behavioural outcomes (Meyer & Allen, 1997). However, Herscovitch & Meyer state that although commitment to change has an overall positive influence, the constructs can have different implications for the actual behaviour of people (2002).
11 Looking at outcomes within the organization it was found that employees with profiles reflecting strong affective commitment and normative commitment reported the most favorable work conditions, stay intentions, and wellbeing (Meyer, 2013). Employees who showed lower levels of commitment and employees with continuance commitment profiles showed the least favorable behavioral outcomes (Meyer, 2013).
To conclude, change commitment is a mind-state that binds employees to a certain path of action. The different aspects of change commitment cannot be seen separately from each other. The most relevant antecedent of change commitment in this study is self-efficacy. Although overall it seems to have positive influence on change commitment, the influences on the different constructs are not totally clear yet. Change commitment is related towards different behavioural outcomes such as productivity and favourable work conditions.
2.3. Effort expectancy
In 1989 Davis laid the foundations for the theory of effort expectancy. He hypothesized that when users find themselves having little difficulties using new IT systems, behavioural intention increases (Venkatesh et al., 2003). Perceived control over systems and situations has an influence on both behavioural intention and behaviour. Therefore effort expectancy is being defined as; the degree of ease associated with the use of the system (p. 450; Venkatesh et al., 2003). The effort expectancy construct originated from three existing concepts; perceived ease of use (TAM/TAM2), complexity (MPCU) and ease of use (IDT). Effort expectancy is considered, besides performance expectancy, as one of the most important antecedents regarding IT user acceptance (Venkatesh et al., 2003). Antecedents of effort expectancy can be divided into internal and situational drivers (Venkatesh, 2000). Internal drivers focus on individual differences such as self-efficacy and computer anxiety (Vella et al., 2013). Situational drivers can be defined as facilitating conditions such as the resources and opportunities to use the system that are being offered (Venkatesh et al., 2003). However effort expectancy is not always an antecedent for behavioural intention. Researchers found that effort expectancy only influences intention right after the first training and in the very early stages of IT implementation. Thereafter the effects will merge into the performance expectancy construct (Venkatesh et al., 2003). This can be explained by the fact that people first need to overcome hurdles in using the system and thereafter focus on instrumental aspects of the system.
12 with perceived ease of use. Vella et al. state that people who have a tighter bond with their organization find it easier to overcome complexity issues (2013). This is also holds for people with high normative commitment levels. No significant relationship was found between perceived ease of use and continuance commitment. Continuance commitment is coming with obedience, which means that it is just a matter of other opportunities and therefore people feel less committed to overcome any complexity problems (2013).
Concluding this study expects effort expectancy to be positively related to affective and normative commitment. Continuance commitment is expected to be not related to effort expectancy. Therefore the following hypothesis are proposed;
H1a: Effort expectancy positively influences affective commitment H1b: Effort expectancy positively influences normative commitment H1c: Effort expectancy does not influence continuance commitment 2.4. End-user training evaluation
End-user training is an important but costly intervention in implementing new IT systems (Compeau et al., 2005). However training promotes the acceptance of the implementation and thereby tries to increase the return on investment of implementation (Hazen, 2014). Earlier research showed that training positively influences skill acquisition and knowledge retention (Tannenbaum & Yukl, 1992). However there is a scarce amount of literature available on the influence of training on change commitment (Sahinidis & Bouris, 2007).
13 Shum et al. (2008) conducted qualitative research on what factors influenced affective commitment. They found that training was related to increased affective commitment. Even more interesting was the fact that respondents mentioned training as the most important factor in a change situation. This can be explained by the fact that knowledge assimilation hurdles are overthrown and employees thereby show higher levels of commitment to change (Shum et al., 2008).
Looking at how people evaluate their training Sahanidis & Bouris (2007) found that perceived training effectiveness positively relates to motivation, employee commitment and job satisfaction. This is also confirmed by Bulut who found that people who have a higher perception of access to training display higher levels of affective commitment (2010). These findings can be explained by the fact that employees feel supported by their organization (Bulut, 2010). Moreover, Bouckenhooghe states that change commitment can be enhanced by providing knowledge and decreasing uncertainty (2010). Therefore the amount of control that people feel can be of great explanatory power for how commitment to change is shaped. Most important is the relationship between benefits of the training and organizational commitment that Bulut & Culha (2010) found. Training has both positive influences on the company as well as for the employees themselves. Companies see benefits in terms of performance, productivity and employee development. On the other hand employees see benefits in terms of promotions or with respect to their future careers (Cahill, 2012).
As training is mentioned as one of the most important factors in change situation, it is expected that this influences change commitment. Moreover it is found that perceived benefits of training influences employee’s commitment to change. Therefore this study assumes that training evaluation will have an influence on affective and normative commitment. Hence training evaluation will not influence continuance commitment;
H2a: Training evaluation positively mediates between effort expectancy and affective commitment
H2b: Training evaluation positively mediates between effort expectancy and normative commitment
14 2.5. Conceptual model
Figure 1. Conceptual model
3.
Research Methodology
The change commitment construct has already been researched and tested extensively by several researchers (Armenakis, 1996; Bouckenhooghe, 2012; Herscovitch, 2002). Therefore this study connects the already existing theories on user acceptance and adoption of IT-systems (Davis & Davis, 1990; Venkatesh & Davis, 2002; Venkatesh et al., 2003) with the theory of change commitment (Meyer et al., 2002; Herscovitch & Meyer, 2002). Because these topics are already researched but not yet related to each other a theory testing approach is suitable for this research (Van Aken et al, 2012).
3.1. Research setting
The gathering of the data was conducted at a University. June 2013 this University made the transition from Oracle to Google Apps for Education (GAfE) for students and alumni as a provider for University email. This transition was a predecessor for the transition of University staff, planned end of May 2014. There are three main reasons why the University switched to GAfE. First, the Oracle system was fully developed and became outdated. Second, Universities do not have to pay for GAfE, which allows them to spend their budgets different. And last, with GAfE there is the possibility for students and staff to work in the cloud and share information more easily (“Google Apps for education”, 2014). In order to make the transition from Oracle to GAfE easier, the Centre of Information Technology (CIT) of the University offers two different workshops for
15 University staff. The first workshop is being called; Slimmer werken met Google mail (Smarter working with Google mail). The second one is being called; Slimmer werken met Google Calender (Smarter working with Google calendar). Both courses were taught from May until June 2014 at different faculties of the University. 22 workshops are planned with almost 700 participants who subscribed to them. Workshops took about 3 hours and were supported by a training manual handed out to each participant. The rooms where training was held were provided with a computer for each participant. Each hour of training consisted out of 45 minutes of teaching and 15 minutes of trying/playing with GAfE. Additionally the trainers explain what employees can expect at the day of transition and what they have to do in order to complete the transition. During the training not only attention is paid to the mere functional components. Participants are also pointed to other features which are not necessarily in the package but can be downloaded. An example is Google Labs where experimental functions are being offered.
3.2. Data collection and sample
The data will be collected with an online survey which will be spread by the trainers at the end of each workshop session. Online surveys have been set up with help of Google Forms. In general Qualtrics is the preferred system for online surveys, however in this particular case it was interesting to test GAfE in a research setting. Since it was not possible to provide the survey link digitally, trainers wrote it down on the board. Therefore a customized link with Bitly was made which really increased the response rate because people did not have to type in a difficult url. At May 26 data collection stopped since GAfE was implemented at that date.
16 Table 1. Descriptive statistics
Demographic variable Data (%)
Gender Male 42 (21,3) Female 155 (78,8) Age <25 12 (6,6) 25-29 12 (12,8) 30-34 13 (19,4) 35-39 15 (27,0) 40-44 26 (29,1) 45-49 38 (59,7) 50-54 20 (69,9) 55-59 35 (87,7) 60-65 24 (100) Experience Low 108 (54,8) Moderate 84 (42,6) Experienced 5 (2,5) 3.3. Measurement
The questions in the survey are already well-validated in literature. With respect to the topic of change commitment, items from Herscovitch & Meyer (2001) are being used. With respect to the control variable factors three concept; age, gender, experience of the research of Venkatesh et al (2003) are being used. Seven questions related to training evaluation are measured on a 5-item and 20 questions related to effort expectancy and change commitment are measured on a 7-item Likert scale. Likert scales are especially useful because they have the same response style and therefore increase the information processing capabilities (Edwards et al., 1997).
3.3.1. Independent variables
3.3.1.1. Effort expectancy
The effort expectancy construct as used by Venkatesh et al. (2003) consists out of the perceived ease of use construct (Davis, 1989), a complexity construct (Thompson et al., 1991) and the ease of use construct (Moore & Benpasat, 1991). Effort expectancy is being measured by two items from the research of Venkatesh et al. (2003) on a 7-point Likert Scale. One item is derived from the ease of use component and the other from the perceived ease of use component.
3.3.1.2. Training evaluation
17 3.3.2. Dependent variables
3.3.2.1. Change commitment
Items (appendix II) on change commitment are measured by a seven-point Likert scale (1 = strongly disagree, 7 = strongly agree). Each commitment construct is covered by six questions as founded in the study by Herscovitch & Meyer (2002). In previous studies it was found that affective commitment did not correlated with continuance commitment. Moreover it was found that normative commitment correlated with both affective commitment and continuance commitment (Herscovitch & Meyer, 2002; Vella et al., 2013).
3.3.3. Control variables
3.3.3.1. Age
In general people ought to think that when people age, their ability to learn decreases. However this is not completely true as some abilities will remain the same, while others decline (Venkatesh, 2000). With respect to age and perceived behaviour control Venkatesh found that age is related to initial acceptance decisions (2000). Younger employees are more likely to focus on attitudes using the new technology, while the older employees focus more on the perceived behavioural control. However Davis & Davis (1990) found that age has no significant influence on the ability to become a good end-user. In this study age will be coded as a continuous variable, this was found convenient in the study of Venkatesh et al. (2000; 2003).
3.3.3.2. Gender
18 behaviour intention was mostly explained by perceived behavioural control factors. In order to measure gender a dummy variable will be used, which was also found convenient in the study of Venkatesh et al. (2000; 2003).
3.3.3.3. User experience
In IT user acceptance theory, user experience is considered to be an important control variable (Venkatesh & Davis, 2000). It decreases uncertainty about the system and it helps overcoming the ‘first time using the product’ hurdles (p. 190 Venkatesh & Davis, 2000). User experience will be coded categorical in three time sections, this was found convenient in the study of Venkatesh et al. (2000; 2003). However measuring user experience in this study as Venkatesh (2003) did is not convenient. The studies on user acceptance were mostly conducted at three different times. One right after training and two times later in time. Theory behind this approach was that user acceptance can evolve over time because people are already working with the system (Venkatesh et al., 2003). However in this study this study only measures at one time, right after training and before implementation of GAfE.
4.
Analysis & results
4.1. Factor analysis
Orthogonal factor analysis (n=5) was conducted using a Varimax rotation. Although a sample size of n=300 is considered comfortable for factor analysis, smaller sample sizes (n=150) are also adequate when there are high loadings on the items (Tabachnick & Fidell, 2007). Therefore the sample size of this study (n=200) was found convenient for factor analysis. After the factor analyses there were no items deleted from the effort expectancy construct (EE1, EE2), there are two out of seven items deleted in the training construct (TR1, TR2, TR3, TR5, TR6), two out of six items were deleted in the affective commitment construct (AC2, AC4, AC5, AC6), four out of six items were deleted in the normative commitment construct (NC4, NC5) and also four out of six items were deleted in the continuance commitment construct (CC1, CC2) (Appendix I).
4.2. Descriptive statistics
19 (Cunningham, 2006; Batistelli, 2014). In this study we found the following Cronbach’s Alpha; affective commitment α=.88, continuance commitment α=.81 and normative commitment α= .85. Effort expectancy showed a Cronbach’s Alpha of α=.88 and training evaluation α=.88.
Looking at these correlations, it can be concluded that there exists a relation between age and how experienced people consider themselves with Google Apps (r=-.214). This also counts for gender (r=.225). Therefore it is expected that young males will be more likely to consider themselves capable with Google Apps. This is also true for the correlation between age and effort expectancy (r=-.187). And for gender and effort expectancy (r=.150). In between correlations of the change commitment do exist. Continuance commitment correlates significantly with affective commitment (r= -.337) and with normative commitment (r=.181). When looking at effort expectancy and change commitment correlations are found with affective commitment (r=.314) and normative commitment (r=-.163). Also correlations are found between effort expectancy and training evaluation (r=.185). Training evaluation correlates positively with affective commitment (r=.156). Correlations that were found among the different commitment constructs were different than found in the study of Herscovitch & Meyer (2002). In this study we found a significant relation between affective and continuance commitment, contrary to their non-significant results. Also this research did not yield a significance relationship between affective and normative commitment as was found by the study of Herscovitch & Meyer (2002).
Table 2. Correlations between variables
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
20 4.3. Linear regression analyses
In order to test the hypotheses linear regression analyses were conducted. First the relationship between the dependent variables; affective commitment, continuance commitment, normative commitment and between the independent variable effort expectancy was estimated. The findings show that there is a significant correlation between effort expectancy and affective commitment (β = .203, p< 0.01). Therefore hypothesis H1a is accepted. There is a significant relationship between effort expectancy and normative commitment (β = -.193, p<.05) found. However this relationship is negative (Appendix IV), hence hypothesis H1b can be rejected. Moreover there was no significant relationship found between effort expectancy and continuance commitment (β = -.010, p = .802). Therefore hypothesis H1c is accepted.
Next the relationship between training evaluation and effort expectancy is estimated. A significant positive relationship was found between training evaluation and effort expectancy (β = .185, p<0.01). Thereafter the relationship between training evaluation and change commitment is estimated. The correlation shows a significant relationship between training evaluation and affective commitment (β = .178, p<0.05). However in order to estimate for mediation a Baron & Kenny test needs to be conducted. This will be done in the upcoming paragraph. There was no significant correlation between training evaluation and normative commitment found (β = .095, p = .201). Therefore hypothesis H2b is rejected. Lastly, no significant correlation was found between training evaluation and continuance commitment (β = -.024, p = .736). Therefore hypothesis H2c is accepted.
4.4. Sobel test for partial mediation
21 test (β= .210, p=.005). This means that there is evidence for a partial mediator (Baron & Kenny, 1986). The Baron & Kenny method is being used by many researchers to test for partial correlation (Mackinnon et al., 2007). However this test does not confirm the significance of the indirect pathway. There are two ways to measure the significance of this indirect pathway; Judd & Kenny difference of coefficients approach (1981) or the Sobel product of Coefficients Approach (1982). Both approaches yield the same outcomes, in this case was chosen for the Sobel approach. The results showed that training evaluation no significant partial mediator between effort expectancy and affective (p= 0.198). Hence hypothesis 2a is rejected (Appendix II).
4.5. Control variables
Not all of the control variables were found significant in this research model. With respect to the model of affective commitment only experience was found significant (β=.204, p < 0.01). Age had a significant (β=-0.167, p<0.05) weak negative relationship with affective commitment. However when introducing experience as a control variable into the construct this effect was diminished. It is interesting to see that in the affective commitment model experience explains more than effort expectancy. However as was stated earlier, in this research experience is irrelevant since we are only measuring at one time, namely right before introducing GAfE. Therefore experience can be considered similar to effort expectancy. This was also the case for continuance commitment (β=-.207, p < 0.01). Also the effect of age became non-significant after introducing experience as a control variable. However with affective commitment there was a positive relationship with experience. But this was reversed for continuance commitment. So people who had more experience with Google Apps showed lower continuance commitment. For normative commitment none of the control variables had a significant influence.
Table 3. Results of the hypotheses
Variable Hypothesis Supported/ Not supported Relationship performance Effort expectancy positively influences affective commitment H1a Accepted
Effort expectancy positively influences normative commitment H1b Rejected negative
Effort expectancy does not influences continuance commitment H1c Accepted n.s.
Training evaluation positively mediates between effort
expectancy and affective commitment H2a Rejected
n.s. partial moderation Training evaluation positively mediates between effort
expectancy and normative commitment H2b Rejected n.s. Training evaluation does not mediate between effort
22
5.
Discussion
The findings show that there is a relationship between effort expectancy and change commitment. Hence the relation will be further explained as well as the influence of training evaluation. As expected effort expectancy has a positive influence on affective commitment. Hereby this study confirms the findings of Meyer et al. (2002) and Herold et al. (2001) who found a weak positive significant relationship between self-efficacy and affective commitment. Moreover it confirms the findings of Vella et al. (2013) who found a positive relationship between affective commitment and perceived ease of use. In contrast, the study of Neves reported a non-significant relationship between self-efficacy and affective commitment (2009). The explanation is twofold. Firstly, Neves attributed these to the fact that affective commitment can be influenced by other situational variables. Secondly, these findings can be related to the time of measurement (Neves, 2009). The fact that employees were not able to test the new IT system can explain for difference in outcomes for self-efficacy. In this study employees neither had the chance to work with Google Apps on forehand. However employees could already open a Gmail account themselves to practice with. When we look at the correlations in this study it shows that we found an even stronger relationship than the aforementioned studies. This difference can be explained by the fact that self-efficacy related to the object of change is stronger than self-efficacy in general (Argawal et al., 2001). Moreover this study was conducted right after the training, the moment where effort expectancy influences behavioural intention most strongly (Venkatesh et al., 2003). Why employees with higher effort expectancy also display higher levels of commitment can be explained by the fact that when people feel convenient to use the system it is more likely that they are more positive about the change just because people think it provides benefits for the organization (Herold et al., 2007). Although this relationship sounds straightforward, some studies found that people can have high levels of commitment without actually knowing how to use a system or what a change is about (Cawsey et al., 2011).
23 commitment and vice versa. Recently, Lin & Hwang (2014) found no significant relationship between knowledge self-efficacy and normative commitment. The explanation they put forward is that there might be other situational factors that are more relevant to normative commitment. A sound explanation in literature is hard to find. However it could be reasoned employees with a low effort expectancy are more committed in terms of moral obligation .This might be due to the fact that although employees have a lower level of self-efficacy towards the system and therefore a lower understanding of the system benefits, they still might be committed because they believe they have the moral obligation to do the right thing for their company (Meyer & Herscovitch, 2002).
There was no significant relationship found between effort expectancy and continuance commitment. This is convenient with previous studies on self-efficacy (Vella et al. 2013; Meyer et al. 2002; Herold et al., 2001). These findings can be explained by the fact that people who show high levels of continuance commitment are not inherently interested in the organization itself, neither in the IT system. Therefore effort expectancy cannot have an influence on continuance commitment (Vella et al., 2013). Moreover this result can be partially due to the fact that this survey was taken in at the classroom in the University right after training (Edwards et al., 1997). Items in the continuance commitment scale can be perceived as discordant to people. With this in mind it could be that people have felt adversely with respect to these items in this setting.
Regarding the effects of the training evaluation, initial results indicated that training evaluation was a partial moderator for the relationship between effort expectancy and affective commitment. However, after conducting the Sobel test for significance of the mediating relationship it was found to be insignificant. And although training evaluation explains 10% of the variance for affective commitment, it diminishes after introducing effort expectancy(Appendix IV). These findings are in line with the findings of Shafiq et al. (2013) who found that the perceived benefits of training are positively related towards affective commitment.
24 In general this study confirms the findings of Batistelli et al. (2014) who state that change commitment is being influenced by change concerns. Especially the change concern about mastering the change explains the relationship between effort expectancy and affective and normative change commitment.
Figure 2. Conceptual model with results
6.
Conclusion & Managerial implications
Although the vast majority of companies implements technology into their business, less of them are concerned with how employees adopt these systems (Vella et al., 2013). This is at least noteworthy because in order for these systems to work properly employees have to feel comfortable in using these systems. First of all this study aims to unravel what different intentions or dispositions people can have regarding the implementation of a new IS system and how these can be influenced by antecedents of IT user acceptance. Thereby, this study aims to open up the black box of individual responses to change due to implementation of a new IT system (Venkatesh et al., 2003).
By combining the change commitment model with the aspects of effort expectancy different hypotheses have been developed. The findings show that there is a relation between effort expectancy and change commitment. This is consistent with the research question; what is the influence of effort expectancy and training evaluation on change commitment. In the discussion it was found that effort expectancy has a significant positive influence on affective commitment and a negative influence on normative
TRAINING EVALUATION CONTINUANCE COMMITMENT EFFORT EXPECTANCY .314** n.s. -.163* n.s.
AGE GENDER EXPERIENCE
25 commitment. Training evaluation can serve as a control variable but has no significant influence on the relationship between effort expectancy and change commitment. Hence it can be concluded that employees who have higher levels of self-efficacy related to change content are more committed due to the fact that they believe in the inherent benefits of the IT system. Contrary it was found that employees with lower levels of self-efficacy show higher levels of normative commitment. Which means that employees with higher levels of self-efficacy related to change content are less committed in terms of moral obligation towards the organization.
In terms of organizational outcomes it is extremely important to have employees with high levels of effort expectancy. Researchers found that employees with higher levels of affective commitment participate more intensely in new change programs (Meyer et al., 2002). High levels of affective commitment can therefore, in the light of this study, be considered a key success factor for IT implementation. Moreover the findings demonstrate that employees pertain different behavioural intentions towards change (Herscovitch & Meyer, 2002). Also this study contributed to the literature by unravelling the difference in measures of self-efficacy. In search of literature on self-efficacy, great differences were found in terms of intentional and behavioural outcomes. This was partly due to the fact that researchers measured self-efficacy in general and self-efficacy towards content of change. Lastly this study found that time of measurement is critical, this is due to the fact self-efficacy towards the content of change to have the same problem that Venkatesh et al. encountered with their effort expectancy construct. It diminishes after time and merges into expectations of the change (2003).
26 expected that these employees will feel more stimulated to also use these voluntary features in order to work more innovative (Herscovitch & Meyer, 2002).
7.
Research limitations & future research
The results of this study should be interpreted with its limitations in mind. First of all, the study is limited by the fact that it is only focussing on the relationship between effort expectancy and change commitment. However as stated earlier, there are many more antecedents that can have an influence on change commitment. Second, in this study organizational tenure as a control variable was not included. This could be influencing the outcomes, since employees with less experience tend to find it more difficult to assess their feelings towards a change due to less comparison material (Meyer & Herscovitch, 2002). Third, the research setting needs to allow for social desirable responding. Respondents were asked to fill in the surveys right after training, when they were still in the classroom with the trainer (Edwards et al., 1997). Fourth, this sample consisted mostly out of women (78.8%), which means that there is more focus on self-efficacy aspects instead of performance aspects of the IT system (Venkatesh, 2000). Also women tend to grade themselves lower in terms of self-efficacy with respect to the change content (Venkatesh et al., 2003). Lastly, it must be stated that, in relation to former limitation, this sample does not give an adequate representation of the research population (Edwards et al., 1997).
27
8.
References
Agarwal, R., Sambamurthy, V. V., & Stair, R. M. (2000). Research Report: The Evolving
Relationship Between General and Specific Computer Self-Efficacy--An Empirical Assessment. Information Systems Research, 11(4), 418.
Aken J.E. van, Berends, H., Bij, H. van der (2012). Problem solving in organizations – A methodological handbook for business and management students. Cambridge: Cambridge University Press
Armenakis, A. A., Harris, S. G., & Mossholder, K. W. (1993). Creating readiness for organizational change. Human Relations, 46, 681–703.
Bandura, A. (1982). Self-efficacy mechanism in human agency. American psychologist, 37(2), 122. Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of personality and social psychology, 51(6), 117.
Bartlett, K. R. (2001). The relationship between training and organizational commitment: A study in the health care field. Human Resource Development Quarterly, 12(4), 335-352.
Battistelli, A., Montani, F., Odoardi, C., Vandenberghe, C., & Picci, P. (2014). Employees' concerns about change and commitment to change among Italian organizations: The moderating role of innovative work behavior. The International Journal Of Human Resource Management, 25(7), 951-978.
Beaudry, A., & Pinsonneault, A. (2005). Understanding User Responses to Information Technology: A Coping Model of User Adaptation. MIS Quarterly, 29(3), 493-524.
Bouckenooghe, D. (2012). "The role of organizational politics, contextual resources, and formal communication on change recipients' commitment to change: A multilevel study." European Journal Of Work & Organizational Psychology 21, no. 4: 575-602.
Bouckenooghe, D. (2010). Positioning Change Recipients’ Attitudes Toward Change in the Organizational Change Literature. Journal Of Applied Behavioral Science, 46(4), 500-531.
Brynjolfsson, E., & Hitt, L. (1995). Information Technology as a Factor of Production: The Role of Differences among Firms. Economics Of Innovation And New Technology, 3(3-4), 183-199. Brynjolfsson, E., & Hitt, L. M. (2000). Beyond Computation: Information Technology,
Organizational Transformation and Business Performance. Journal Of Economic Perspectives, 14(4), 23-48.
Bulut, C., & Culha, O. (2010). The effects of organizational training on organizational commitment. International Journal Of Training & Development, 14(4), 309-322.
Cahill, J. L. (2012). The collaborative benefits of Google Apps Education Edition in higher education. Dissertation Abstracts International Section A, 73.
Cawsey, T. F., Deszca, G., & Ingols, C. (2011). Organizational change: An action-oriented toolkit. New York: Sage.
Coetsee, L. (1999). From resistance to commitment. Public Administration Quarterly, 23, 204– 222.
28 Davis, F. D. (1989). Perceived Usefulness, Perceived Ease of Use, and User Acceptance of
Information Technology. MIS Quarterly, 13(3), 319-340.
Edwards, J.E., Thomas, M.D., Rosenfeld, P., & Booth-Kewley, S (1997). How to conduct organizational surveys: a step-by-step guide. New York: Sage Publications.
Fedor, D. B., Caldwell, S., & Herold, D. M. (2006). The Effects of Organizational changes on Employee Commitment: A Multilevel Investigation. Personnel Psychology, 59(1), 1-29. Google Apps for Education. (2014, May 12). Retrieved from http://www.google.com.
Hazen, B. T., Kung, L., Cegielski, C. G., & Jones-Farmer, L. (2014). Performance expectancy and use of enterprise architecture: training as an intervention. Journal Of Enterprise Information
Management, 27(2), 180-196.
Herold, D. M., Fedor, D. B., & Caldwell, S. D. (2007). Beyond change management: A multilevel investigation of contextual and personal influences on employees' commitment to change. Journal Of Applied Psychology, 92(4), 942-951.
Herscovitch, L., & Meyer, J. P. (2002). Commitment to Organizational Change: Extension of a Three-Component Model. Journal Of Applied Psychology, 87(3), 474-487.
Lin, H., & Hwang, Y. (2014). Do feelings matter? The effects of intrinsic benefits on individuals' commitment toward knowledge systems. Computers In Human Behavior, 30191-198.
Mackinnon, D.P., Fairchild, A.J., & Fritz, M.S. (2007). Mediation analysis. Annual review of Psychology, 58, 593-614.
Markus, M.L. 2004. "Techno-change Management: Using IT to Drive Organizational Change," Journal of Information Technology (19:1), 4-20.
Meyer, J. P., & Herscovitch, L. (2001). Commitment in the workplace: Toward a general model. Human Resource Management Review, 11(3), 299.
Meyer, J. P., Stanley, D. J., Herscovitch, L., & Topolnytsky, L. (2002). Affective, Continuance, and Normative Commitment to the Organization: A Meta-analysis of Antecedents, Correlates, and Consequences. Journal Of Vocational Behavior, 61(1), 20-52.
Meyer, J. P., Stanley, D. J., Jackson, T. A., McInnis, K. J., Maltin, E. R., & Sheppard, L. (2012). Affective, Normative, and Continuance Commitment Levels across Cultures: A Meta-Analysis. Journal Of Vocational Behavior, 80(2), 225-245.
Meyer, J. P., Srinivas, E. S., Lal, J. B., & Topolnytsky, L. (2007). Employee commitment and support for an organizational change: Test of the three-component model in two cultures. Journal Of Occupational & Organizational Psychology, 80(2), 185-211.
Meyer, J. P., Kam, C., Goldenberg, I., & Bremner, N. L. (2013). Organizational commitment in the military: Application of a profile approach. Military Psychology, 25(4), 381-401.
Morris, M. G., & Venkatesh, V. (2010). Job Characteristics and Job Satisfaction: Understanding the Role of Enterprise resource planning system implementation. MIS Quarterly, 34(1), 143-161. Muldoon, M. (2014, April 15) The fight gets serious in the cloud. Retrieved from:
http://www.thefinancialist.com/the-fight-gets-serious-in-the-cloud/.
29 Neves, P. (2009). Readiness for Change: Contributions for Employee's Level of Individual Change and Turnover Intentions. Journal Of Change Management, 9(2), 215-231.
Noe, R. A., & Wilk, S. L. (1993). Investigation of the factors that influence employees' participation in development activities. Journal of applied psychology, 78(2), 291.
Oreg, S., Vakola, M., & Armenakis, A. (2011). Change Recipients’ Reactions to Organizational Change: A 60-Year Review of Quantitative Studies. Journal Of Applied Behavioral Science, 47(4), 461-524.
Orlikowski, W. J. (2000). Using Technology and Constituting Structures: A Practice Lens for Studying Technology in Organizations. Organization Science, 11(4), 404-428.
Orlikowski, W. J. (2007). Sociomaterial Practices: Exploring Technology at Work. Organization Studies (01708406), 28(9), 1435-1448.
Preacher, J.K & Leonardelli, G.J.. (2010-2014) Calculation for the Sobel Test; An interactive calculation tool for mediation tests. Retrieved from: http://quantpsy.org/sobel/sobel.htm.
Sahinidis, A. G., & Bouris, J. (2008). Employee Perceived Training Effectiveness Relationship to Employee Attitudes. Journal Of European Industrial Training, 32(1), 63-76.
Shafiq, M., Zia-ur-Rehman, D., & Rashid, M. (2013). Impact of Compensation, Training and Development and Supervisory Support on Organizational Commitment. Compensation & Benefits Review, 45(5), 278-285.
Shum, P., Bove, L., & Auh, S. (2008). Employees' affective commitment to change: the key to successful CRM implementation. European journal of marketing, 42(11/12), 1346-1371. Song, M. Im, S., Bij, H. van der, Song, L.Z. (2011). Does strategic planning enhance or impede innovation and firm performance? Journal of Product Innovation Management 28 (4): 503-520. Sultan, N. (2010). Cloud computing for education: A new dawn?. International Journal Of Information Management, 30(2), 109-116.
Tabachnick, B.G. & Fidell, L.S. (2007). Using multivariate statistics (5th edn). Boston: Pearson
Education.
Tannenbaum, S. I., Mathieu, J. E., Salas, E., & Cannon-Bowers, J. A. (1991). Meeting trainees' expectations: The influence of training fulfillment on the development of commitment, self-efficacy, and motivation. Journal Of Applied Psychology, 76(6), 759-769.
Vella, J., Caruana, A., & Pitt, L. (2013). Organizational commitment and users' perception of ease of use: a study among bank managers. Journal Of Management Development, 32(4), 351-362. Venkatesh, V. (2000). Determinants of Perceived Ease of Use: Integrating Control, Intrinsic Motivation, and Emotion into the Technology Acceptance Model. Information Systems Research, 11(4), 342.
Venkatesh, V., & Davis, F. D. (2000). A Theoretical Extension of the Technology Acceptance Model: Four Longitudinal Field Studies. Management Science, 46(2), 186.
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User Acceptance of Information Technology: Toward a Unified View. MIS Quarterly, 27(3), 425-478.
Appendix I – Coding table & Survey questions
Coding Table
1 Wat is uw leeftijd Age
2 Bent u man of vrouw Gender
3 Wat is uw ervaring met GoogleApps Experience
4 Het gebruik van het systeem is voor mij helder en begrijpelijk EE1 5 Leren hoe Google apps werkt is makkelijk voor mij EE2
6 Ik zie de waarde van het gebruik van Google Apps AC1
7 De overgang naar Google Apps is een goede keuze AC2
8 Het gebruik van Google Apps heeft een belangrijk doel AC3
9 REVIk denk dat we een fout maken door Google Apps te gebruiken AC4 10 REVHet zou beter zijn als we het oude systeem blijven gebruiken AC5 11 REV Ik vind de overgang naar Google Apps niet juist AC6 12 Ik heb geen andere keuze dan Google Apps te gebruiken CC1
13 Ik voel mij verplicht Google Apps te gebruiken CC2
14 Het is voor mij te risicovol Google Apps niet te gebruiken CC3
15 Het zou niet verstandig zijn voor mij om Google Apps niet te gebruiken CC4
16 Het is niet verstandig mij negatief uit te spreken over Google Apps CC5
17 Het is geen optie om mij te verzetten tegen Google Apps CC6
18 Het gebruik van Google Apps voelt als een verplichting NC1
19 Het zou niet goed zijn om het gebruik van Google Apps te belemmeren NC2
20 Ik zou mij niet vervelend voelen als ik Google Apps niet gebruik NC3
21 Het zou onverantwoordelijk zijn voor mij Google Apps niet te gebruiken NC4 22 Ik zou mij schuldig voelen als ik Google Apps niet zou gebruiken NC5
23 Ik voel mij absoluut niet verantwoordelijk het gebruik van Google Apps NC6
24 Ik volgde een cursus over Googlemail, Google Calendar, beide C1
25 De leerdoelen van de training waren helder TR1
26 Participatie en interactie tijdens de training werd aangemoedigd TR2 27 De onderwerpen van de training waren relevant voor mij TR3
28 De inhoud van de training was overzichtelijk en makkelijk te volgen TR4
29 Het trainingsmateriaal vond ik nuttig TR5
30 De training is nuttig voor mijn eigen werkzaamheden TR6
31
Appendix II – Factor analysis
Rotated Component Matrixa
Raw Rescaled Component Component 1 2 3 4 5 1 2 3 4 5 EE1 1,042 ,885 EE2 1,202 ,931 AC2 1,009 ,733 AC4 (R) 1,285 ,819 AC5 (R) 1,448 ,898 AC6 (R) 1,367 ,859 TR1 ,557 ,719 TR2 ,432 ,582 TR3 ,650 ,805 TR5 ,502 ,689 TR6 ,529 ,737 CC1 1,632 ,909 CC2 1,535 ,912 NC4 1,295 ,831 NC5 1,524 ,883
Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a. Rotation converged in 5 iterations.
32
Appendix III- Sobel test
33
Appendix IV – Multiple Hierarchical Regression analysis
Dependent variable:
Affective commitment Model 1 Model 2 Model 3 Model 4 Model 5 Control variables Age -0.167* -0.169* -0.108 -0.136 -0.102 Gender 0.041 -0.023 0.017 -0.012 Experience 0.274*** 0.260*** 0.204** Training evaluation 0.178* 0.131 Independent variables Effort expectancy 0.203** Adjusted R2 0,023 0,02 0,083 0,109 0,138 Dependent variable:
Normative commitment Model 1 Model 2 Model 3 Model 4 Model 5 Control variables Age 0.106 0.109 0.114 0.099 0.067 Gender -0.087 -0.092 -0.071 -0.044 Experience 0.023 0.015 0.068 Training evaluation 0.095 0.14 Independent variables Effort expectancy -0.193* Adjusted R2 0.006 0.009 0.004 .007 0,033* Dependent variable:
Continuance commitment Model 1 Model 2 Model 3 Model 4 Model 5 Control variables Age 0.143* 0.142* 0.095 0.099 0.097 Gender 0.008 0.057 0.051 0.053 Experience -0.212*** -0.210*** -0.207*** Training evaluation -0.024 -0.022 Independent variables Effort expectancy -0.010 Adjusted R2 0.015* 0.010 0.046*** 0.042* 0.037* *p <.05, **p<.01, ***p<.001
