1
The Impact of Positive and Negative Management Control System
Packages on Autonomous Motivation in the Public Sector
By Miranda Schenk S3845613 m.a.schenk@student.rug.nl 18 January 2021 University of Groningen Faculty of Economics and Business Management Accounting & Control
Supervisor: prof. dr. ir. Paula van Veen-Dirks
Word count: 11.192
Abstract
2 Introduction
Managers and employees around the world suddenly need to discover new ways of working due to consequences of COVID-19. To continue work as normal as possible it is necessary to “create a more meaningful and lasting emotional connection between the employee and their employer” (Meister, 2020). Prior studies on work performance stated that employees’ motivation is one of the critical factors in encouraging work performance (Pinder, 2011). Therefore, managers constantly try to find methods to effectively support and enhance employees’ motivation (Imran, Arif, Cheema, and Azeem, 2014). In order to improve motivation, it is important to understand different types of motivation and how to influence motivation.
In literature the distinction is made between two types of motivation: intrinsic and extrinsic (Cerasoli, Nicklin, and Ford, 2014; Ryan and Deci, 2000). Cerasoli et al. (2014) suggest that intrinsic motivation is directed by personal interest, and extrinsic motivation is conducted by incentives or punishments (external mechanisms). Other studies state that external mechanisms can cause two effects; on the one hand these mechanisms can enhance motivation (Gagné and Deci, 2005; Zuckerman, Porac, Lathin, and Deci, 1978), while on the other hand, external incentives can cause a so-called crowding out effect. Which means that due to the incentives intrinsic motivation is weakened (Gneezy, Meier and Rey-Biel, 2011; Gagné and Deci, 2005). In recent management accounting literature, there is a growing attention for management control systems (MCS) in relation to employees’ motivation (Groen, Wouters, and Wilderom, 2017; Chen, Lill, and Vance, 2019; Van der Kolk, Van Veen-Dirks and Ter Bogt, 2019). The study of Groen et al. (2017) found a positive indirect relation between the use of operational performance measurement and employees’ motivation. Similar results are found in the study of Chen et al. (2019) whereby several combinations of management control systems enhance motivation. Van der Kolk et al. (2019) investigated the influence of MCSs on intrinsic and extrinsic motivation and found positive relations. The above-mentioned studies did not take into account the different types of motivation. By making such a distinction between intrinsic and extrinsic motivation, it is possible to explicitly investigate which MCS influences the different motivations of employees.
3
Deci, 2005). When people are feeling autonomous, they find an activity interesting and they do the activity voluntarily. With controlled motivation there is an influence of extrinsic rewards. Intrinsic motivation is prototypically autonomous, whereas extrinsic motivation can be both controlled and autonomous. Extrinsic motivation consists of four different regulations that increase gradually in autonomy, respectively: external, introjected, identified, and integrated. The first two, external and introjected regulation are part of controlled motivation. Identified and integrated regulation belong to autonomous motivation. According to the SDT there are three basic psychological needs before someone is able to feel intrinsic or internalized with an activity (Gagné and Deci, 2005). These three basic needs are: autonomy (own choices), competence (self-deployment) and relatedness (belonging to a group). As stated before, the study of Van der Kolk et al. (2019) made a distinction between intrinsic and extrinsic motivation. It was concluded that it would be interesting for future research to distinguish the four types of extrinsic motivation and explore the influence of different MCSs on these extrinsic motivations. With this research, I will complement their study by making a distinction between controlled and autonomous extrinsic motivation. Autonomous motivation is assumed to be more powerful than controlled motivation (Groen et al., 2017). In their study they found that positive outcomes, such as well-being, job satisfaction, and affective commitment, are positively related to autonomous motivation and negative outcomes such as psychological distress and turnover intentions are negatively related to autonomous motivation (Groen et al., 2017). Besides, controlled motivation is negatively related to work satisfaction and positively related to turnover intentions (Gillet, Gagne, Sauvagere and Fouquereau, 2013). Especially in the public sector, employees are more stimulated by autonomous motivation compared to controlled motivation, such as extrinsic rewards (Perry, Hondeghem, and Wise, 2010). In this study I will continue on this literature by using data from employees of higher professional educational organisations.
4
also on how management controls fit with each other (Bedford et al., 2016). Management control packages can consist of many different combinations and various studies (Malmi and Brown, 2008; Sutton and Brown, 2016) show evidence that certain combinations have a positive influence on organisational outcomes. Different terminology is used to address diverse types of MCSs. This study uses the distinction between MCSs that generate positive forces and MCSs that generate negative forces, which is perceived from the study of Kruis, Speklé, and Widener (2016).
According to Simons (1995), these MCSs that generate positive and negative forces should be used side by side to create the right balance and thus generate the best results. He developed a framework called the Levers of Control (LoC). The framework consists of four control systems: beliefs, boundary, diagnostic, and interactive. Beliefs and interactive control systems are associated with generating positive forces, whereas the other two are associated with generating negative forces (Simons, 1995; Kruis et al., 2016). Inter-dependency between the control systems is a characteristic of the framework. Malmi and Brown (2008) assumed that MCS should be used simultaneously, and that situation and the type of organisation determines which balance between the different MCSs should be used. However, the study of Bedford et al. (2016) suggests that most of the MCSs are not interdependent. This is in line with the claim of Indjejikian and Matĕjka (2012) that the interdependency between MCSs is likely overstated. Those studies demonstrate that MCSs that have been observed to be beneficial in isolation do not need to be simultaneously present with other MCSs to achieve effective outcomes (Bedford et al., 2016; Indjejikian and Matĕjka 2012). An explanation for MCS packages that stimulate effective results could be that MCSs that generate positive forces are more strongly present in the package than MCSs that generate negative forces and that is does not matter whether all of the MCSs are present in the MCS package.
5
I will contribute to the existing literature in two ways. First, by making a distinction between autonomous and controlled extrinsic motivation. Autonomous motivation is of great importance to organisations because it is related to positive outcomes, such as performance and job satisfaction (Gagné, Forest, Gilbert, Aubé, Morin, and Malorni, 2010). Secondly, I will contribute to existing literature by investigating MCS as a package. On the one hand, by examining MCS packages whereby MCSs that generate positive forces dominate and on the other hand, by examining MCS packages whereby MCSs that generate negative forces dominate. I expect that MCS packages that generate positive forces are better suited to create autonomous motivation within the public sector compared to MCS packages that generate negative forces. In order to investigate this relation, I stated the following research question:
To what extent do MCS packages that generate positive or negative forces influence employees’ autonomous motivation in higher professional educational organisations?
The remainder of this paper is structured as follows. In the next chapter, scientific literature is reviewed to develop four hypotheses. Subsequently, the method section presents the research design. Afterwards, the results are presented based on correlation analysis and multiple regressions. The last chapter provides a conclusion of the results, discussion, limitations and recommendations for future research.
Theoretical Background
Self-Determination Theory
6
autonomous motivation (Gagné and Deci, 2005). In contrast, with controlled motivation people show behaviour because of an external factor. Extrinsic motivation can be both autonomous and controlled motivation. Besides, autonomous motivation can also consist partly from intrinsic and partly from extrinsic motivation (Figure 1). For example, when a waiter wants to deliver good care and hospitality to his guest in order to receive a good tip. In this case, wanting to receive a high tip from the guest generates extrinsic motivation. Furthermore, the waiter can be intrinsically motivated because he gets satisfaction from making his guest happy. Even if he will not receive a tip it will fulfil him to give people a good time. This encompasses a voluntary effort. In short, on one side when intention to act is lacking and people are acting without intent or not acting at all, we refer to amotivation. On the other side, when people find an activity interesting and acting because they are completely self-motivated, we refer to intrinsic motivation. Between these two extremes, we refer to extrinsic motivation, whereby there is a difference to what extent the regulation is autonomous (Baard, Deci, and Ryan, 2004).
Extrinsic motivation can be divided in four types; external, introjected, identified and integrated regulation. The first extrinsic motivation type, external regulation, is a prototype of controlled motivation (Gagné and Deci, 2005). In this case extrinsic motivation of someone is in contrast with his or her intrinsic motivation. People act with the intention of avoiding an undesired outcome or earning a desired one. Secondly, introjected regulation refers to behaviour that is engaged by a person, but not recognized as his or her own behaviour (Gagné and Deci, 2005). Introjected is
7
a relative controlled form of extrinsic motivation although regulation is within the person itself. In other words, people are not accepting rules as their own, but obeying them in order to gain pride or avoid guilt (Gagné et al., 2010). The third extrinsic motivation type is identified regulation. With this regulation, people are more congruent with their personal goals and identities, which result in, people feeling greater freedom and volition (Gagné and Deci, 2005). It is autonomously regulated, because of the identification with people’s personal values and therefore they will accept it as their own (Gagné et al., 2010). Finally, the fourth extrinsic motivation type is integrated regulation. This regulation is the most autonomous form. Integrated regulation is self-determined because people have the full impression that their behaviour is an intrinsic part of who they are (Gagné and Deci, 2005). It is quite similar to intrinsic motivation, but it differs in the interest towards the activity. Both integrated and identified regulation, differ from intrinsic motivation because people find the activity instrumentally important for personal goals, and not find the activity itself interesting. Intrinsic motivation is driven by emotion, integrated and identified are driven by values and goals (Gagné et al., 2010).
8
and absence of the three needs will result in people being controlled motivated (Gagné et al., 2010).
Former researchers indicate that high autonomous motivation results in better job satisfaction and better performance evaluations at the individual level (Deci, Connell, and Ryan, 1989; Baard et al., 2004). It is important for organisations to understand the differences between autonomous and controlled motivation. Using the wrong extrinsic incentive to motivate employees could have an undesired effect on autonomous motivation. This “crowding-out” effect appears when motives are reduced because of extrinsic incentives (Gneezy et al., 2011). Autonomous motivation is of greater importance than controlled motivation and therefore organisations should pay attention to avoid crowding out of autonomous motivation. The shift of motivation from autonomous to controlled is detrimental to organisations and difficult to reverse (Groen et al, 2017). Management control systems (MCS) are designed, among other things, to motivate employees. In order to provoke autonomous motivation among employees the right balance of MCSs should be determined. Simons (1995) developed a framework to rank different control systems, called Levers of Control. This framework will be used to better understand what kind of MCSs there are and how to use them in order to create the right balance.
Levers of Control
9
researches suggested that the beliefs and interactive control systems generate positive forces, while boundary and diagnostic systems generate negative forces (Simons, 1995; Henri, 2006; Widener, 2007; Tessier and Otley, 2012; Kruis et al., 2016). MCSs that generate positive forces tend to guide, reward, motivate, and promote learning whereas MCSs that generate negative forces tend to punish, coerce, prescribe and control (Tessier and Otley, 2012). Although, MCSs that generate positive forces should not be confused with “good” controls whereas MCSs that generate negative forces should not be confused with “bad” controls (Simons, 1995; Tessier and Otley, 2012). The quality of controls indicates to whether a control is efficient, effective, economical, etcetera and whether it has undesirable effects such as slowing down innovation, causing dysfunctional behaviour, etcetera (Tessier and Otley, 2012). Every organisation has its own optimal balance of control systems; this depends among other things on the industry and the selected strategy of the organisation (Kruis et al., 2016). This research discusses the influence of the composition of MCSs on autonomous motivation in the public sector.
10
2007). Organisations that focus on strategy as a pattern highly emphasize interactive control (Kruis et al., 2016).
Public Sector
11 MCS as a package
There are many studies indicating that MCSs are interdependent and should not be used alone but in a package (Malmi and Brown, 2008; Henri, 2006; Widener, 2007; Mundy, 2010). However, finding the right balance seems difficult because this balance is assumed to be different for every organisation (Bedford et al., 2016). The research of Mundy (2010) concludes that how managers use MCSs forms the balance. All four LoC have a significant role in establishing the right balance (Mundy, 2010; Speckle, Van Elten, and Widener, 2017). Balance within a MCS package gets growing attention among researchers (Kruis et al, 2016; Henri, 2006; Bedford et al., 2016). Although, a specific description of how a balance of MCS should look like is not yet be found. The statement that all four LoC have a significant role does not mean that every individual MCS must be equally present. A manufacturer for example, will benefit from clear composed critical success factors, so that employees know what is expected from them. This indicates that a diagnostic control system should be strongly present within the MCS package. Kruis et al., (2016) argued that the strategy of an organisation determines how the right balance of MCSs should look like. Within the public sector employees seem to prefer intrinsic over extrinsic rewards (Perry et al., 2010; Wright, 2007; Georgellis, Iossa, and Tabvuma, 2011). Hence, for public organisations it is necessarily to make sure that MCSs that increase intrinsic rewards should be strongly present within the MCS package. Similarly, public organisations should avoid MCS packages whereby MCSs that cause the “crowding out” effect are strongly present.
MCS and autonomous motivation
12
can provoke the three innate psychological needs that will create the feeling of autonomous motivation. I expect that MCSs that generate positive forces (beliefs and interactive control) will enhance the need-supportive environment, and thus should be more strongly present in a MCS package compared to MCSs that generate negative forces (boundary and diagnostic control) in order to generate autonomous motivation within the public sector.
The beliefs system focuses on inspiring employees and gaining commitment. Examples of these controls are core values of an organisation or the mission and vision statement. Simons (1995) argues that this MCS is related to the strategy of an organisation which is used to inspire, empower and encourage employees to come up with new ideas. By using the beliefs system, employees feel related and more emotionally attached to the organisation (Merchant & Van der Stede, 2007). According to the SDT autonomous motivation will increase when people feel more related with an organisation. When employees internalize the values and mission of an organisation, it will increase the self-regulated and autonomous feelings (Ryan and Deci, 2000). Furthermore, supportive behaviour of a manager can affect the satisfaction of autonomy, competence, and relatedness (Deci, Ryan, Gagné, Leone, Usunov, and Kornazheva, 2001). With autonomous motivation people consider goals and values of an organisation important and they feel connected, as it is a part of who they are. The boundary and diagnostic control system are both based on constraints (Chen et al., 2019). The SDT suggest that an individual’s sense of autonomy could be reduced by environmental features that constrain behaviour (Ryan and Deci, 2017). For this reason, I think these MSCs that generate negative forces should not be strongly present in the MCS package. Therefore, I expect that beliefs systems have a positive effect on autonomous motivation as long as they are stronger present in the MCS package than the MCSs that generate negative forces (boundary and diagnostic control). Hence, my first hypothesis is:
H1: Beliefs systems are positively associated with autonomous motivation of employees
13
The boundary systems rely on restraints in order to prevent high-risk behaviour among employees. According to Gagné et al. (2010) are rules in an organisation used to ensure an incentive to show appropriate behaviour. The rules will extrinsically motivate employees to avoid certain consequences. Once the focus is on telling employees what not to do, they will feel controlled by their managers. This could cause an increased stress level among employees (Aiello, 1993). This suggests that, related to the SDT, feelings of autonomy are reduced, because employees do not experience choices as their own. Moreover, boundary systems do not focus on intrinsic rewards but on restraining people from risky behaviour (Widener, 2007). Linking intrinsic rewards to the MCSs, it is likely that the beliefs and interactive control system stimulates this rather than the boundary system. Hence, a boundary system tries to prevent employees from high-risk behaviour by the use of constraints (Simons, 1995). In the public sector employees prefer intrinsic rewards that are not likely to occur in a highly regulated environment were boundary systems dominate and MCSs that generate positive forces are weakly present. Therefore, I expect that boundary systems have a negative effect on autonomous motivation as long as they are more strongly present in the MCS package than the MCSs that generate positive forces (beliefs and interactive control). Hence, my second hypothesis is:
H2: Boundary systems are negatively associated with autonomous motivation of
employees as long as they are more strongly present in the MCS package than the beliefs and interactive control systems.
14
systems foster controlled motivation more than autonomous motivation. As stated before, the beliefs and interactive control system tend to generate positive forces by stimulating and promoting employees to learn (Tessier and Otley, 2012). For public organisations this suggest that diagnostic control systems, that are based on constraints and critical success factors should not dominate over the MCSs that generate positive forces. Therefore, I expect that diagnostic control systems have a negative effect on autonomous motivation as long as they are more strongly present in the MCS package than the MCSs that generate positive forces (beliefs and interactive control). Hence, my hypothesis is:
H3: Diagnostic control systems are negatively associated autonomous motivation of
employees as long as they are more strongly present in the MCS package than the beliefs and interactive control systems.
15
benefits of the learning environment created by the interactive control system it will decrease the autonomous motivation. Therefore, I expect that interactive control systems have a positive effect on autonomous motivation as long as they are more strongly present in the MCS package than the MCSs that generate negative forces (boundary and diagnostic control). Hence, my last hypothesis is:
H4: Interactive control systems are positively associated with autonomous motivation
of employees as long as they are more strongly present in the MCS package than the boundary and diagnostic control systems.
16 Methodology Data collection
To investigate the influence of MCSs that generate positive and negative forces within a package, I will use a quantitative approach to test the hypotheses described in the prior section. For this approach, I use a secondary dataset collected by the University of Groningen through surveys from four different higher professional educational organisations (A, B, C, and D) in the north of the Netherlands. The dataset consists of 524 completed surveys obtained from employees and managerial positions. The survey contains 28 items with multiple sub-questions and was designed in the online program Qualtrics. Before the surveys were provided, the board of directors was approached to obtain permission. The board of directors showed a high interest in the results of the study, which resulted in good commitment to the study and a relatively high response rate among the employees. Organisation A has a response rate of 88%; organisation B 52%; organisation C 50%, and organisation D 60%. For this research I use a sample consisting of 250 respondents from organisation B (91) and organisation C (159). To comply with privacy regulations anonymity is guaranteed in the survey.
The questions from the survey are based on constructs of existing literature. The LoC framework by Simons (1995) is used to measure MCSs. This framework distinguishes four MCSs that are used for the different statements in the survey; beliefs, boundary, diagnostic control, and interactive control. Furthermore, the constructs regarding autonomous motivation are based on the SDT. This theory divides autonomous motivation in: identified regulation, integrated regulation, and intrinsic motivation.
Sample
17
have been working at their organisation for 10 years or less. Only 28 of the respondents (11.2%) have a temporary contract, the other 88.8% has a fixed contract.
Analyses
In order to execute the statistical test for this research I used the program IBM SPSS version 25. A few key assumptions of the dataset are checked before testing the hypotheses to make sure the dataset is suitable. First, I checked the data for system-missing or extreme values, however non were found. Most of the items in the survey are measured with Likert-scales. This is a measurement instrument that contains response categories that have a rank order, but the differences between the intervals are not equal (Jamieson, 2005). For example, the value 4 is not two times “better” than the value 2. In this research one of the variables is motivation which is measured by statements with the answers that vary from (1) “not at all” to (7) “completely/entirely”. For the items regarding to job-related tension (JRT) there is an answer possibility “does not apply (DNA)”, this gives the respondent the opportunity to indicate that the statement does not apply to that person. These respondents do not answer the question by saying it is not applicable to them and therefore they create missing values (Hair, Black, Babin, Anderson, and Tatham, 1998). There are 37 cases (14.8%) within this
Table 1. Sample Characteristics
Variable Frequency Percentage Variable Frequency Percentage
Gender (n =250) Organisation (n =250)
Male 122 48.8% B 91 36.4%
Female 128 51.2% C 159 63.6%
Age distribution (n =250) Tenure (n =250)
25 and younger 1 4.0% 0 to 5 years 78 31.2%
26 to 35 30 12.0% 6 to 10 years 68 27.2%
36 to 45 60 24.0% 11 to 20 years 58 23.2%
46 to 55 83 33.2% 21 to 30 years 35 14.0%
Older than 55 76 30.4% 31 years and longer 11 4.4%
Educational Background (n =250) Type of contract (n =250)
Secondary vocational education 21 8.4% Temporary contract 28 11.2%
Secondary education 4 1.6% Fixed contract 222 88.8%
Bachelor degree 71 28.4%
18
sample that have answered with DNA at least one time. However, 25 of those cases replied DNA only once and the maximum times DNA answered per case is 4 out of the 15 statements. I decided to keep the cases in the sample because most of the 37 cases only have answered DNA once and otherwise the remaining data of these cases will also be lost.
Common method bias is one of the main sources of measurement error. Validity of the conclusions about the relationships between measures is threatened by measurement errors (Podsakoff, Mackenzie, Lee, and Podsakoff, 2003). To prevent common method bias, several procedures can be done. One is to guarantee respondents that their answers are anonymous. Hence, when sensitive questions about someone’s position or the organisation are asked people might give social desired responses if the responses are not anonymous (Podsakoff et al., 2003). In this study the anonymity of the respondents is ensured. In addition, a statistical approach, the Harman’s single factor test, is performed to detect possible common method bias. This technique is one of the most widely used among researchers to address this issue. There are two possibilities when a substantial amount of common method variance is existing; “(a) a single factor will emerge from the factor analysis or (b) one general factor will account for the majority of the covariance among the measures” (Podsakoff et al., 2003, p. 889). An unrotated factor analysis with one factor from all the variables shows a variance of 27.25% which is far below the threshold of 50% (Eichhorn, 2014). These two procedures suggest there is no common method bias present in this research.
19
to the normal distribution. The agreed threshold for a skewness test of a normal distributed construct is between -2 and 2. All the skewness and kurtosis outcomes of the constructs in this study are between -1 and 1. Based on the Q-Q plots and the skewness and kurtosis test it is assumed that the data of the constructs are normally distributed.
Afterwards I checked the data for reliability and validity with several tests. First, I performed a KMO-test and a Barlett’s Test of Sphericity test to check whether the data is suitable for an exploratory factor analysis. The KMO-test has a score of .875 and the Barlett’s Test of Sphericity was significant (p < .01), which indicated an exploratory factor analysis should be useful for this sample. In order to ensure construct validity, I performed the principal component analysis (PCA) with a direct oblimin rotation (Appendix B). Most of the items from the survey did properly represent the constructs from the literature. I performed the PCA several times and removed three items from the variable JRT due to cross loadings. Besides, the factor analysis showed that the JRT variable consists of two factors, which I label with JRTC 1 and JRTC 2. These will be included in the study as two separate control variables. Next, I tested the internal consistency of the statements for each construct with the Cronbach’s Alpha. This test is measuring how close the items within a construct are related. The threshold for a construct in order to be considered as internally reliable is above .70 (Nunnally, 1978). The Cronbach’s Alpha scores of the constructs in this study are all above the threshold. At last, I checked the data for multicollinearity with the variance inflation factor (VIF). The VIF values in this study are all between 1.38 and 3.09 which is far below the threshold of 10. This indicates there is no evidence for multicollinearity issues.
Measurements Independent variables
20
performed a Cronbach’s Alpha test. The four statements have a Cronbach’s Alpha of 0.873, which is above the threshold of 0.7.
Boundary system: The second independent variable is also based on the study of Kruis et al. (2016). In their study they explain that boundary systems “ensuring that the strategy domain is firmly set, and that behavioural hazards are recognized and dealt with in codes of conduct” (Kruis et al., 2016, p. 28). For this study I used four statements about the organisation’s code of business conduct developed by Widener (2007) to measure boundary systems (Boundary). The statements were answered with a seven-point Likert scale, varying from (1) “strongly disagree” to (7) “strongly agree”. The Cronbach’s Alpha of the four statements related to boundary systems is 0.849. This is above the threshold of 0.7 which suggest the construct is reliably measured.
Diagnostic control system: This independent variable is based on the study of Bedford and Malmi (2015). In their study they use the definition perceived from Simons (1995) “Monitoring activity through deviations from pre-set standards of performance” (Bedford and Malmi, 2015, p. 7). The items used in the study are based on Henri (2006), Widener (2007) and the descriptions of Simons (1995). To measure diagnostic control system (Diagnostic) there are five statements included in the survey. The statements were answered with a seven-point Likert scale, varying from (1) “to an extremely low extent” to (7) “to an extremely high extent”. The Cronbach’s Alpha of the four statements related to diagnostic control systems is 0.965. This is above the threshold of 0.7 which suggest the construct is reliably measured.
21 Dependent variable
Autonomous motivation: The dependent variable is based on the study of Gagné et al. (2015). The study of Deci and Ryan (2008) is used to state that “autonomous motivation has been found to yield the most desirable behavioural, attitudinal, and affective out-come” (Gagné et al., 2015, p. 179). To measure autonomous motivation the Multidimensional Work Motivation Scale is used (MWMS). The MWMS is based on the SDT continuum from Deci and Ryan (2005). It differs from other scales that measure work motivation by assessing work motivation at the domain level of analysis (Vallerand, 1997). The MAWS contain five items: amotivation, external regulation, introjected regulation, identified regulation, and intrinsic motivation. These items have an increasing sense of autonomy. Autonomous motivation can be measured by identified regulation and intrinsic motivation. First, I start the analysis with autonomous motivation as combined construct of identified regulation and intrinsic motivation. The Cronbach’s Alpha of this construct is 0.829, which is above the threshold of 0.7 and thus considered to be reliable. In additional analyses a distinction is made between identified regulation and intrinsic motivation to examine if this shows any different results. Identified regulation is measured by three statements in the survey, which were answered with a seven-point Likert scale varying from (1) “not at all” to (7) “completely/entirely”. The Cronbach’s Alpha of this construct regarding identified regulation is 0.810, which is above the threshold and thus considered to be reliable. Integrated regulation is not existing in the MAWS and therefore not included in this study. Intrinsic motivation is also measured by three statements in the survey, with the same seven-point Likert scale as identified regulation. The Cronbach’s Alpha of the three statements related to intrinsic motivation is 0.899. This is above the threshold of 0.7 which suggest the construct is reliably measured.
Control variables
22
Age: This control variable is used because age can have a significant influence on the motivation of employees (Boumans, De Jong, and Janssen, 2011). Younger employees tend to be more motivated when career opportunities are rising. As employees turn older their interest shifts to maintaining their position (Boumans et al., 2011). They can fear for their position because of competition of younger employees or changing environments. Therefore, this study takes into account that age can influence the employee’s perception on the four MCS. This variable will be measured in number of years.
Job-related tension: This variable is about suffering stress in working environments by employees. Job-related tension (JRT) indicates psychological bothering of an individual due to work related factors (Wooten, Fakunmoju, Kim and LeFevre, 2010). In the study of Wooten et al. (2010) a Job-Related Tension Index (JRTI) based on 15 measures is used. The study uses a JRTI that is measured on a 6-point Likert scale varying from (1) “never” to (5) “always” or (6) “does not apply”. Three items had to be removed (JRT 6, JRT 8 & JRT 10), due to their low factor loadings with their components (Appendix B). Besides, the factor analysis showed that the variable is divided into two components. The first component consists of JRT 1, JRT 2, JRT 3, JRT 7, JRT 11 & JRT 12, and will be called JRTC 1. The second component consists of JRT 4, JRT 5, JRT 13, JRT 14 & JRT 15, and will be called JRTC 2.
Data analysis
23
motivation, to investigate whether the results were different when autonomous motivation was split up.
Results
Descriptive statistics
Table 2 shows the descriptive statistics of the independent variables, dependent variable, and the control variables. These statistics include the minimum, maximum, median, mean, standard deviation, skewness, and kurtosis. The values in Table 2 are based on the average of the statements that together form a construct. As described in the methodology, three items are removed from the JRT statements and two constructs have been formed, based on the exploratory factor analysis. The means of the different control systems are between 3.44 and 3.85, which is a little below the Likert midscale point of 4. All the MCSs have a median of 4. Remarkable are the high values (6) of the median of autonomous motivation, identified regulation, and intrinsic motivation. For the dependent variable, autonomous motivation, the mean is 5.55. Both, the high mean and median suggest that the feeling of autonomous motivation is quite high among the respondents.
Table 2. Descriptive Statistics
Variable Minimum Maximum Median Mean SD Skewness Kurtosis
24 Correlation analysis
Table 3 reports the correlations between the constructs of the independent variables, dependent variable, and the control variables. Beliefs system has a significant positive correlation with the other MCSs; boundary system (r = .38, p < .01), diagnostic control system (r = .39, p < .01), and interactive control system (r = .55, p < .01). Furthermore, boundary system correlates positively significant with diagnostic control system (r = .47, p < .01) and interactive control system (r = .51, p < .01). The last combination of MCSs between diagnostic control system and interactive control system is also positively significant correlated (r = .75, p < .01). Besides, the construct of autonomous motivation has a significant positive correlation with the beliefs system (r = .24, p < .01). Autonomous motivation consists of the constructs identified regulation and intrinsic motivation. Intrinsic motivation has a positive significant correlation with the beliefs (r = .30, p < .01), and interactive control system (r = .15, p < .05). Job-related tension is divided in the constructs JRTC 1 and JRTC 2. The MCSs correlate negatively to JRTC 1; beliefs system (r = -.43, p < .01), boundary system (r = -.27, p < .01), diagnostic control system (r = -.27, p < .01), and interactive control system (r = -.36, p < .01). JRTC 2 too has a negative correlation with the MCSs; beliefs system (r = -.28, p < .01), boundary system (r = -.19, p < .01), diagnostic control system (r = -.13, p < .05), and interactive control system (r = -.26, p < .01).
** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).
Table 3. Correlation Analysis
25 Regression analyses
The correlation analysis showed correlations between the four MCSs, and between the beliefs system some motivation types. To further investigate the hypothesis, several multiple regressions are performed. For the multiple regression analyses, the sample was split into two groups for each hypothesis: one that did meet the conditions and one that did not (control group). There are two models for the multiple regression analyses. The first model tests the relation between the control variables and autonomous motivation. Furthermore, the second model tests the relation between the different MCSs and autonomous motivation. In the additional analyses, autonomous motivation is split up in identified regulation and intrinsic motivation.
Model 1 includes the control variables (age, JRTC 1, and JRTC 2) to test their association with autonomous motivation. All the groups of which one MCS is more strongly present than two other MCSs, did not show a significant relationship with autonomous motivation. For example, the group from hypothesis 1: beliefs systems are more strongly present in the MCS package than boundary and diagnostic control systems. Besides, the outcomes of F-tests are not significant. This indicates that age, and stress perceived by employees do not influence the autonomous motivation. For the control groups there are negative significant relations between JRTC 1 and autonomous motivation (Appendix C, D, E and F). The beliefs control group (consisting of employees that scored lower or equal on the beliefs system than on the boundary and/or diagnostic control system) showed a negative correlation between JRTC1 and autonomous motivation (β = -.202, p < .05). The other control groups did show similar correlations between JRTC 1 and autonomous motivation: boundary control group (β = -.292, p < .01), diagnostic control group (β = -.357, p < .01), and
interactive control group (β = -.207, p < .05). These results suggest that stress perceived by employees decrease autonomous motivation.
26
supported. However, the beliefs control group showed a positive significant relation between beliefs system and autonomous motivation (β = .189, p < .05).
Hypothesis 2 hypothesized that boundary systems are negatively associated with autonomous motivation as long as they are more strongly present in the MCS package than the beliefs and interactive control system. Table 4 depicts that there is no relation between boundary system and autonomous motivation, which leaves hypothesis 2 unsupported. Nevertheless, the group that meets these conditions, and the boundary control group, did have a significant positive association between beliefs system and autonomous motivation (β = .398, p < .01 and β = .167, p < .05, respectively).
Hypothesis 3 states that diagnostic control systems are negatively associated with autonomous motivation as long as they are more strongly present in the MCS package than the beliefs and interactive control system. Table 5 shows there is no relation between diagnostic control system and autonomous motivation, which leaves
Table 4.
Regression analysis of beliefs and boundary system
Beliefs High Beliefs Low
Boundary High Boundary Low Autonomous motivation Autonomous motivation Autonomous motivation Autonomous motivation Controls Age .010 .003 .004 .001 JRTC 1 -.248 -.127 -.031 -.216* JRTC 2 .268* .119 .135 .155 Main effects Beliefs .146 .189** .398*** .167** Boundary .139 -.029 -.183 .022 Diagnostic -.279 .018 .146 -.085 Interactive .206 -.032 -.219 .027 n 61 189 96 154 R2 .127 .063 .133 .091 ΔR2 .011 .027 .065 .047 F-value 1.098 1.744 1.936* 2.085
27
hypothesis 3 unsupported. However, the group that meets these conditions as well as the diagnostic control group did have a significant positive association between beliefs system and autonomous motivation (β = .262, p < .10 and β = .205, p < .01, respectively).
Hypothesis 4 hypothesized that interactive control systems are positively associated with autonomous motivation as long as they are more strongly present in the MCS package than the boundary and diagnostic control system. The group that meets these conditions and the interactive control group did not show a significant relation between interactive control system and autonomous motivation (Table 5). Therefore, hypothesis 4 is not supported. However, Table 5 depicts that the interactive control group showed a positive significant relation between beliefs system and autonomous motivation (β = .235, .01).
Table 5.
Regression analysis of diagnostic and interactive control system
Diagnostic High
Diagnostic
Low Interative High Interactive Low
Autonomous motivation Autonomous motivation Autonomous motivation Autonomous motivation Controls Age .006 .004 .001 .005 JRTC 1 .106 -.289** -.345 -.121 JRTC 2 .034 .189* .483 .107 Main effects Beliefs .262* .205*** .213 .235*** Boundary .028 -.046 .170 -.015 Diagnostic .031 -.128 -.341 .036 Interactive -.189 .105 .230 -.120 n 81 169 25 225 R2 .064 .125 .208 .094 ΔR2 -.025 .086 -.118 .065 F-value .718 3.271*** .638 3.225***
28 Additional analyses
The group in which beliefs systems are more strongly present than boundary and diagnostic control systems did have a positive significant relation between beliefs system and identified regulation (β = .282, p < .10), and a negative significant relation
between diagnostic control system and intrinsic motivation (β = -.504, p < .05) (Table 6). The beliefs control group did have a positive significant relation between beliefs system and intrinsic motivation (β = .240, p < .05). These relations do not support hypothesis 1.
The group in which boundary systems are more strongly present than beliefs and interactive control systems did have a positive significant relation with beliefs system and identified regulation (β = .375, p < .05), and with beliefs system and intrinsic motivation (β = .422, p < .05). Moreover, Table 6 depicts that within this group there is a negative significant relation between interactive control system and identified regulation (β = -.431, p < .05). The boundary control group did have a positive significant relation between beliefs system and intrinsic motivation (β = .208, p < .05). These relations do not support hypothesis 2.
The group in which diagnostic control systems are more strongly present than beliefs and interactive control systems did not have any significant relations with identified regulation or intrinsic motivation. The diagnostic control group did have a positive significant relation between beliefs system and intrinsic motivation (β = .289, p < .01), a negative significant relation between boundary system and identified regulation (β = -.110, p < .10), and a negative significant relation between diagnostic control system and intrinsic motivation (β = -.309, p < .10) (Table 7). These relations do not support hypothesis 3.
The group in which interactive control systems are more strongly present than boundary and diagnostic control systems did not have any significant relations with identified regulation or intrinsic motivation (Table 7). The interactive control group did have a positive significant relation between beliefs system and identified regulation (β
29 Table 6.
Additional analyses - Regression
Beliefs High Beliefs Low Boundary High Boundary Low
Identified regulation Intrinsic motivation Identified regulation Intrinsic motivation Identified regulation Intrinsic motivation Identified regulation Intrinsic motivation Controls Age .013 .006 -.006 .011 -.003 .011 -.004 .005 JRTC 1 .013 -.510** -.034 -.221* .024 -.086 -.078 -.354** JRTC 2 .216 .320 .186 .051 .230 .040 .170 .139 Main effects Beliefs .282* .009 .138 .240** .375** .422** .127 .208** Boundary .065 .212 -.086 .029 -.073 -.292 -.077 .121 Diagnostic -.055 -.504** .017 .019 .201 .091 -.042 -.127 Interactive .104 .308 -.058 -.006 -.431** -.007 .070 -.015 n 61 61 189 189 96 96 154 154 R2 .16 .137 .026 .145 .088 .184 .036 .135 ΔR2 .049 .023 -.011 .112 .016 .119 -.010 .094 F-value 1.437 1.202 .704 4.374*** 1.215 2.840** .781 3.267***
30 Table 7.
Additional analyses - Regression
Diagnostic High Diagnostic Low Interactive High Interactive Low
Identified regulation Intrinsic motivation Identified regulation Intrinsic motivation Identified regulation Intrinsic motivation Identified regulation Intrinsic motivation Controls Age -.002 .014 -.002 .010 .008 -.005 -.002 .012* JRTC 1 .185 .027 -.123 -.455*** -.154 -.536 .003 -.244** JRTC 2 .032 .036 .195 .184 .365 .601 .160 .054 Main effects Beliefs .248 .277 .122 .289*** .134 .293 .206*** .263*** Boundary -.006 .062 -.110* .018 -.034 .375 -.072 .043 Diagnostic .179 -.118 .054 -.309* -.312 -.370 .070 .002 Interactive -.279 -.100 -.028 .238 .532 -.072 -.151 -.089 n 81 81 169 169 25 25 225 225 R2 .082 .058 .047 .198 .171 .260 .051 .141 ΔR2 -.006 -.032 .006 .163 -.171 -.045 .020 .113 F-value .936 .641 1.142 5.69 .499 .853 1.663 5.082***
31
Discussion and Conclusion
The goal of this research is to investigate the influence of MCS packages that generate positive or negative forces on autonomous motivation. Currently, it is unclear how a MCS package should look like for organisations in the public sector to stimulate autonomous motivation. Autonomous motivation is important to investigate because it is related to positive organisational outcomes, such as well-being, job satisfaction, and affective commitment (Groen et al., 2017; Gagné et al., 2010). In order to promote these positive organisational outcomes, it is important to stimulate the autonomous motivation of employees. To investigate this, four hypotheses have been tested on a sample of 250 employees.
The regression analyses of the groups of which one MCS is more strongly present than two other MCSs did not show a relation between the control variables and autonomous motivation. However, the control groups did show a negative relation between JRTC 1 and autonomous motivation.
32
the MCS package than boundary and diagnostic control systems. Findings show no evidence for this relation. However, the interactive control group did show positive relations between beliefs system and autonomous motivation.
Discussion
The control variable JRTC 1 did show a negative relation with autonomous motivation, but only when the test was performed in the control groups. An explanation for this difference could be the variance in sample size. Since, the control groups consisted of bigger samples. Jenkins and Quintana-Ascencio (2020) argued that the problem of a small sample size is that contradictory or inconclusive results are more likely.
Beliefs systems are used by managers to share the organisation’s vision and implant it deeply throughout the organisation. Hypothesis 1, regarding the beliefs system was not supported. However, a possible explanation for the positive relation that was found between beliefs systems and identified regulation is that when employees feel related with the organisations values, they will be able to be oriented to the long-term meaning of a current activity even if they do not find it particularly interesting. When employees experience a less strong presence of the beliefs system, a positive relation between beliefs system and intrinsic motivation arises. This could be explained as follows. The core values and mission of public organisations are based on helping others, just like the values and mission of many employees who work in this sector (Wright, 2007). When top management share the organisations’ vision more, employees acknowledge their shared values and therefore experience interest and excitement.
Boundary systems use rules in an organisation to ensure an incentive to show acceptable behaviour. Hypothesis 2 regarding the boundary system was not supported. Nevertheless, there was a relation within the group that was used for this hypothesis. A possible explanation for the positive relation between beliefs system and autonomous motivation is that when employees experience a high level of defined acceptable or unacceptable behaviours, they approve when they notice that the values and mission of an organisation matches with their own values, which, in turn leads to an increase in autonomous motivation.
33
it will decrease the feelings of doing tasks voluntarily, and thus the intrinsic motivation will decrease.
Interactive control systems stimulate employees to think along and make their own decisions. Hypothesis 4 regarding the interactive control system was not supported. However, there was a positive relation between beliefs system and autonomous motivation in the control group. A possible explanation could be that when employees do not experience a high level of involvement in decision-making, the beliefs system fulfils the need for relatedness more. If the values of the organisation and the employee align it is likely to be of more importance than involvement in decision-making.
Limitations and future research
34 Conclusion
35 References
Aiello, J. R. (1993). Computer‐based work monitoring: electronic surveillance and its effects. Journal of Applied Social Psychology, 23(7): 499-507.
Baard, P. P., Deci, E. L., & Ryan, R. M. (2004). Intrinsic need satisfaction: a motivational basis of performance and well‐being in two work settings. Journal of applied social
psychology, 34(10): 2045-2068.
Becker, T. E. (2005). Potential problems in the statistical control of variables in organizational research: A qualitative analysis with recommendations. Organizational Research
Methods, 8(3): 274-289.
Bedford, D. S., & Malmi, T. (2015). Configurations of control: An exploratory analysis. Management Accounting Research, 27: 2-26.
Bedford, D. S., Malmi, T., & Sandelin, M. (2016). Management control effectiveness and strategy: An empirical analysis of packages and systems. Accounting, Organizations and
Society, 51: 12-28.
Besley, T., & Ghatak, M. (2005). Competition and incentives with motivated agents. American
economic review, 95(3): 616-636.
Bisbe, J., Batista-Foguet, J. M., & Chenhall, R. (2007). Defining management accounting constructs: A methodological note on the risks of conceptual misspecification. Accounting,
organizations and society, 32(7-8): 789-820.
Boumans, N. P., De Jong, A. H., & Janssen, S. M. (2011). Age-differences in work motivation and job satisfaction. The influence of age on the relationships between work characteristics and workers' outcomes. The international journal of aging and human development, 73(4): 331-350.
Cerasoli, C. P., Nicklin, J. M., & Ford, M. T. (2014). Intrinsic motivation and extrinsic incentives jointly predict performance: a 40-year meta-analysis. Psychological Bulletin, 140(4): 980-1008.
36
De Vocht, A. (2019). Basishandboek SPSS 25 – Voor SPSS & SPSS Subscription. Bijleveld
Press.
Deci, E. L., & Ryan, R. M. (2008). Facilitating optimal motivation and psychological well-being across life's domains. Canadian psychology/Psychologie canadienne, 49(1): 14.
Deci, E. L., Connell, J. P., & Ryan, R. M. (1989). Self-determination in a work organization.
Journal of applied psychology, 74(4): 580.
Deci, E. L., Ryan, R. M., Gagné, M., Leone, D. R., Usunov, J., & Kornazheva, B. P. (2001). Need satisfaction, motivation, and well-being in the work organizations of a former eastern bloc country: A cross-cultural study of self-determination. Personality and social psychology
bulletin, 27(8): 930-942.
Donaldson, L. (2001). The contingency theory of organizations. London: Sage Publications. Eichhorn, B. R. (2014). Common method variance techniques. Cleveland State University,
Department of Operations & Supply Chain Management. Cleveland, OH: SAS Institute Inc,
1-11.
Fisher, C. D. (1978). The effects of personal control, competence, and extrinsic reward systems on intrinsic motivation. Organizational behavior and human performance, 21(3): 273-288. Frey, B. S. (2012). Crowding out and crowding in of intrinsic preferences. Reflexive governance
for global public goods, 75-78.
Frey, B. S., Homberg, F., & Osterloh, M. (2013). Organizational control systems and pay-for-performance in the public service. Organization studies, 34(7): 949-972.
Gagné, M., & Deci, E. L. (2005). Self-Determination Theory and Work Motivation. Journal of
Organizational Behavior, 26: 331-362.
Gagné, M., & Deci, E. L. (2005). Self‐determination theory and work motivation. Journal of Organizational behavior, 26(4): 331-362.
Gagné, M., Forest, J., Gilbert, M. H., Aubé, C., Morin, E., & Malorni, A. (2010). The motivation at work scale: Validation evidence in two languages. Educational and psychological
37
Gagné, M., Forest, J., Vansteenkiste, M., Crevier-Braud, L., Van den Broeck, A., Aspeli, A. K., Bellerose, J., Benabou, C., Chemolli, E., Guntert, S. T., Halvari, H., Indiyastuti, D. L., Jognson, P. A., Molstad, M. H., Naudin, M., Ndao, A., Olafsen, A. H., Roussel, P., Wang, Z., & Westbye, C. (2015). The Multidimensional Work Motivation Scale: Validation evidence in seven languages and nine countries. European Journal of Work and
Organizational Psychology, 24(2): 178-196.
Georgellis, Y., Iossa, E., & Tabvuma, V. (2011). Crowding out intrinsic motivation in the public sector. Journal of Public Administration Research and Theory, 21(3): 473-493.
Gillet, N., Gagné, M., Sauvagère, S., & Fouquereau, E. (2013). The role of supervisor autonomy support, organizational support, and autonomous and controlled motivation in predicting employees' satisfaction and turnover intentions. European Journal of Work and
Organizational Psychology, 22(4), 450-460.
Gneezy, U., Meier, S., & Rey-Biel, P. (2011). When and why incentives (don't) work to modify behavior. Journal of Economic Perspectives, 25(4): 191-210.
Groen, B. A. C., Wouters, M. J. F., & Wilderom, C. P. M. (2017). Employee participation, performance metrics, and job performance: A survey study based on self-determination theory. Management Accounting Research, 36: 51-66.
Hackman, J. R., & Oldham, G. R. (1976). ‘’Motivation through the design of work: test of a theory’’. Organizational Behavior and Human Performance, 16(2): 250–279.
Hackshaw, A. (2008). Small studies: strengths and limitations. European Respiratory Journal, 32(5): 1141-1143
Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (1998). Multivariate data analysis. Upper Saddle River, NJ: Prentice hall, 5(3): 207-219.
Henri, J. F. (2006). Management control systems and strategy: A resource-based perspective. Accounting, organizations and society, 31(6): 529-558.
Hood, C. (1995). The “new public management” in the 1980s: Variations on a theme. Accounting,
38
Imran, H., Arif, I., Cheema, S., & Azeem, M. (2014). Relationship between job satisfaction, job performance, attitude towards work, and organizational commitment. Entrepreneurship and
Innovation Management Journal, 2(2): 135-144.
Indjejikian, R. J., & Matĕjka, M. (2012). Accounting decentralization and performance evaluation of business unit managers. The Accounting Review, 87(1): 261-290.
Jamieson, S. (2005). Likert Scales: How to (ab) Use Them? Medical education, 38(12): 1217-1218.
Kruis, A. M., Speklé, R. F., & Widener, S. K. (2016). The levers of control framework: An exploratory analysis of balance. Management Accounting Research, 32: 27-44.
Latham, G. P., & Pinder, C. C. (2005). Work motivation theory and research at the dawn of the twenty-first century. Annual Review of Psychology, 56: 485-516.
Malmi, T., & Brown, D. A. (2008). Management control systems as a package -Opportunities, challenges and research directions. Management accounting research, 19(4): 287-300. Meister, J. (2020). Employee Experience Is More Important Than Ever During The Covid-19
Pandemic. Forbes. Retrieved from: https://www.forbes.com/sites/jeannemeister /2020/06/08/employee-experience-is-more-important-than-ever-during-the-covid-19-pandemic/#56b1b48d34bc at 10-06-2020.
Merchant, K. A., & Otley, D. T. (2006). A review of the literature on control and accountability.
Management Accounting Research, 2: 785-802.
Merchant, K. A., & Van der Stede, W. A. (2007). Management control systems: performance measurement, evaluation and incentives. Pearson Education.
Mundy, J. (2010). Creating dynamic tensions through a balanced use of management control systems. Accounting, Organizations and society, 35(5): 499-523.
Nunnally, J. C. (1978). Psychometric Theory (2nd ed.) New York: MbecGraw-Hill Otley, D. T., & Berry, A. J. (1994). Case study research in management accounting and
39
Perry, J. L., Hondeghem, A., & Wise, L. R. (2010). Revisiting the motivational bases of public service: Twenty years of research and an agenda for the future. Public administration
review, 70(5): 681-690.
Pinder, C. C. (2011). Work Motivation in Organizational Behavior. Psychology Press, New York,
NY.
Podsakoff, N. P., MacKenzie, S. B., Lee, J. Y., Podsakoff, P. M. (2003). Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal
of Applied Psychology, 88(5): 879-903.
Pollitt, C., & Bouckaert, G. (2011). Public management Reform - A comparative analysis (3rd ed.). Oxford University Press, USA
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation. American Psychologist, 55(1): 68-78.
Ryan, R. M., & Deci, E. L. (2017). Self-determination theory: Basic psychological needs in
motivation, development, and wellness. Guilford Publications.
Schmitt, N. W., & Klimoski, R. J,(1991). Research methods in human resources management.
South-Western Pub.
Simons, R. (1995). Levers of Control. Harvard University Press, Boston.
Speklé, R. F., & Verbeeten, F. H. (2014). The use of performance measurement systems in the public sector: Effects on performance. Management Accounting Research, 25(2): 131-146. Speklé, R. F., van Elten, H. J., & Widener, S. K. (2017). Creativity and control: A paradox -
Evidence from the levers of control framework. Behavioral Research in Accounting, 29(2): 73-96.
Sutton, N. C., & Brown, D. A. (2016). The illusion of no control: management control systems facilitating autonomous motivation in university research. Accounting & Finance, 56(2): 577-604.
40
Vallerand, R. J. (1997). Toward a hierarchical model of intrinsic and extrinsic motivation.
Advances in experimental social psychology, 29: 271-360.
Van den Broeck, A., Vansteenkiste, M., De Witte, H., Soenens, B., & Lens, W. (2010). Capturing autonomy, competence, and relatedness at work: Construction and initial validation of the Work-related Basic Need Satisfaction scale. Journal of occupational and organizational
psychology, 83(4): 981-1002.
Van der Kolk, B., van Veen-Dirks, P. M., & ter Bogt, H. J. (2019). The impact of management control on employee motivation and performance in the public sector. European Accounting
Review, 28(5): 901-928.
Widener, S. K. (2007). An empirical analysis of the levers of control framework. Accounting,
organizations and society, 32: 757-788.
Wooten, N. R., Fakunmoju, S. B., Kim, H., & LeFevre, A. L. (2010). Factor structure of the job-related tension index among social workers. Research on Social Work Practice, 20(1): 74-86.
Wright, B. E. (2001). Public-sector work motivation: A review of the current literature and a revised conceptual model. Journal of public administration research and theory, 11(4): 559-586.
Wright, B. E. (2007). Public service and motivation: Does mission matter? Public administration
review, 67(1): 54-64.
Zuckerman, M., Porac, J., Lathin, D., & Deci, E. L. (1978). On the importance of self-determination for intrinsically-motivated behavior. Personality and social psychology
41 Appendix A
43 Appendix B
Principal Component Analysis
Component 1 2 3 4 5 6 7 Belief_1 .803 Belief_2 .690 Belief_3 .822 Belief_4 .803 Boundary_1 .839 Boundary_2 .906 Boundary_3 .552 Boundary_4 .820 Diagnostic_1 .936 Diagnostic_2 .932 Diagnostic_3 .915 Diagnostic_4 .914 Diagnostic_5 .905 Interactive_1 .792 Interactive_2 .606 Interactive_3 .603 Interactive_4 .589 JRT_1 .619 JRT_2 .698 JRT_3 .726 JRT_4 -.817 JRT_5 -.436 JRT_7 .651 JRT_9 JRT_11 .499 JRT_12 .573 JRT_13 -.888 JRT_14 -.480 JRT_15 -.851 Identified_1 .867 Identified_2 .837 Identified_3 .687 Intrinsic_1 .760 Intrinsic_2 .902 Intrinsic_3 .883
44
Appendix C
Belief High - Regression analysis
Autonomous motivation Identified regulation Intrinsic motivation Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
Controls Age .002 .010 .010 .013 -.005 .006 JRTC 1 -.249 -.248 -.114 .013 -.383* -.510** JRTC 2 .162 .268* .105 .216 .219 .320 Main effects Belief .146 .282* .009 Boundary .139 .065 .212 Diagnostic -.279 -.055 -.504** Interactive .206 .104 .308 R2 .045 .127 .026 .160 .059 .137 ΔR2 -.005 .011 -.025 .049 .009 .023 F-value .898 1.098 .512 1.437 1.182 1.202 * p < .10 ** p < .05 *** p < .01
Belief Low - Regression analysis
Autonomous motivation Identified regulation Intrinsic motivation Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
45 Appendix D
* p < .10 ** p < .05 *** p < .01
* p < .10 ** p < .05 *** p < .01
Boundary High - Regression analysis
Autonomous motivation Identified Regulation Intrinsic motivation Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
Controls Age .013 .004 .005 -.003 .020* .011 JRTC 1 -.119 -.031 .022 .024 -.259 -.086 JRTC 2 .060 .135 .162 .230 -.042 .040 Main effects Belief .398*** .375** .422** Boundary -.183 -.073 -.292 Diagnostic .146 .201 .091 Interactive -.219 -.431** -.007 R2 .036 .133 .019 .088 .092 .184 ΔR2 .005 .065 -.013 .016 .062 .119 F-value 1.16 1.936* .583 1.215 3.101** 2.840**
Boundary Low - Regression analysis
Autonomous motivation Identified regulation Intrinsic motivation Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
46 Appendix E
* p < .10 ** p < .05 *** p < .01
* p < .10 ** p < .05 *** p < .01
Diagnostic High - Regression analysis
Autonomous motivation Identified regulation Intrinsic motivation Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
Controls Age .009 .006 .001 -.002 .016 .014 JRTC 1 .073 .106 .189 .185 -.044 .027 JRTC 2 .015 .034 .041 .032 -.010 .036 Main effects Belief .262* .248 .277 Boundary .028 -.006 .062 Diagnostic .031 .179 -.118 Interactive -.189 -.279 -.100 R2 .014 .064 .025 .082 .026 .058 ΔR2 -.024 -.025 -.013 -.006 -.012 -.032 F-value .372 .718 .667 .936 .678 .641
Diagnostic Low - Regression analysis
Autonomous motivation Identified regulation Intrinsic motivation Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
47 Appendix F
* p < .10 ** p < .05 *** p < .01
* p < .10 ** p < .05 *** p < .01
Interactive High - Regression analysis
Autonomous motivation Identified regulation Intrinsic motivation Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
Controls Age -.005 .001 -.002 .008 -.007 -.005 JRTC 1 -.355 -.345 -.231 -.154 -.480 -.536 JRTC 2 .346 .483 .191 .365 .500 .601 Main effects Belief .213 .134 .293 Boundary .170 -.034 .375 Diagnostic -.341 -.312 -.370 Interactive .230 .532 -.072 R2 .107 .208 .045 .171 .144 .260 ΔR2 -.020 -.118 -.091 -.171 .021 -.045 F-value .843 .638 .330 .499 1.175 .853
Interactive low - Regression analysis
Autonomous motivation Identified regulation Intrinsic motivation Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
