Amsterdam Business School
The effect of results control tightness on work tension and unit performance in professional service firms
Name: Ryon E. Dijkhoff Student number: 10884351
Thesis supervisor: D.M. Swagerman Date: June 8, 2016
Word count: 12712
MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam
Statement of Originality
This document is written by student Ryon Dijkhoff who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
This study uses the survey method to investigate the influence of several dependent variables on the extent of results controls which are part of the MCS design in Professional Service Firms. MCS was conceptualized in four dimensions with two classifications: explicit behavioral control tightness & implicit behavioral control tightness, explicit results control tightness & implicit results control tightness, explicit cultural control tightness & implicit cultural control tightness, and explicit personnel control tightness & implicit personnel control tightness. The interaction that was studied in this study is of explicit and implicit result control tightness on professional tension in a Professional Service Firm, and on the unit performance of a Professional Service Firm. Questionnaires were administered to professionals with a certain level of work experience in Professional Service Firms operating in different countries. Using linear regression analysis the results indicated that (1) explicit results control tightness increases the level of professional tension/stress in Professional Service Firms, and (2) explicit results control tightness increases the level of unit performance in Professional Service Firms.
Table of Contents
1 Introduction ... 8
2 Theory ... 10
2.1 Literature review ... 10
2.1.1 Agency theory ... 10
2.1.2 Management control systems ... 11
2.1.3 Results controls ... 12
2.1.4 Explicit and implicit results control tightness ... 13
2.1.5 Work tension and unit performance ... 15
2.2 Hypothesis development ... 15
3 Research methodology ... 17
3.1 Sample and survey procedure ... 17
3.2 Operationalization of the variables ... 18
4 Statistical analysis ... 21
4.1 Demographics ... 21
4.2 Measurement of variables ... 24
4.2.1 Result controls tightness ... 24
4.2.2 Work tension ... 26 4.2.3 Unit performance ... 27 4.3 Control variables ... 28 4.3.1 Firm size ... 28 4.3.2 Firm reputation ... 28 4.3.3 Firm ownership ... 29 5 Findings ... 30 5.1 Descriptives ... 30 5.2 Linear regression ... 33 5.2.1 Assumptions testing ... 33 5.2.2 Hypotheses findings ... 34 6. Concluding discussion ... 40
6.1 Limitations of this study and directions for future research ... 41 References ... 44 Appendices ... 48
Preface
This thesis is written as a completion to the master Accountancy and Control, at the University of Amsterdam. The master programme focuses on diverse core competences that accounting professionals need to possess, one of them being management control. The subject of this thesis, reflects the effects that management control systems can have on professional tension or the performance of a department in a professional service firm, as will be explained later.
The subject selected is in co-operation with the Faculty of Economics and Business of the University of Amsterdam, which began a study on professional service firms in cooperation with master students. This in combination of personal intrigue in how management control systems affect individuals in an organization led to this thesis.
Since November of last year the data collection has begun through surveys from my personal network in professional service firms. This process went well, communication was the key. After that the in depth reading of theories gave a broader view of the subject. Then the research was operationalized through survey items into statiscical models to test hypothesis in this research. As of February I have been conducting research on this topic. I have experienced this period as very intensive but very rewarding as well.
Acknowledgements
This master thesis has been completed also thanks to many persons which contributed with their suggestions, thoughts, and constructive criticisms. I take therefore the occasion to briefly mention them here.
I am grateful to my thesis supervisor Mr. Dirk Swagerman, for being a motivator and guide during this research, and has been a very available supervisor whenever he was needed. Many thanks also to Mr. Sander van Triest whose statistical knowledge has encouraged me to broaden my knowledge further. Also a big thanks to Mrs. Helena Kloosterman for setting up the survey so that this research could be done and for her patience and guidance when it was needed.
Many thanks to all my friends and ex-colleagues to have taken time out of their busy schedule to give their input on this thesis. The input you had is invaluable and I am really appreciative for your time and efforts.
Several people helped me in a way or another during these years of study abroad. Therefore I take the occasion to thank, in no particular order, Davina Habibe, Leonie Nicolaas, Nabhilach de Palm, Anthony Uden, Meredith Plet, Indy Feliciano, Sennett Stamper, Shawn Lopez, Cecily Fingal, Tony-John Fingal, Canilia Geerman, Leonella de Palm, Anna Maduro, Jourenne Werleman, Shadey Figaroa, Lionel de Cuba, Cristelle Rafine, Xanron Dijkhoff.
I am greatly indebted to my family, especially my father and mother, Frederick and Greta Dijkhoff for always supporting me during my studies abroad, you are the reason I strive and keep going forward, both of you are the wind beneat my wings. To my brothers, Myron and Tyron Dijkhoff, I couldn’t have done it without you, thanks for your unconditional support, patience and understanding.
1 Introduction
There are very few studies involving Management control systems (hereinafter: MCS) in Professional Service Firms. According to Merchant and Van der Stede (2012) management control tightness is a subject that has received relatively little attention in the literature. Professional Service Firms are considered special types of firms, therefore it is very current to do a research in this group of firms because there are still a lot to be researched which would contribute greatly to the academic literature. By definition, Professional Service Firms are organizations where expert knowledge workers (the ‘operating core’) exercise to a greater or lesser extent control over both the means and ends of services delivery (Kipping et al., 2013, p. 778).
Due to the lack of research on these types of firms and their management control systems, the following research question is formulated: what is the effect of results control tightness on work tension and unit performance in Professional Service Firms? There has not been extensive research yet in this area and there is room for discussion on this topic. Results control tightness could be a strong predictor of work tension and unit performance in Professional Service Firms.
Derived from the agency theory is the fact that agents (employees) will not always act in the best interest of the principal (owners of a firm) (Jensen and Meckling 1976, p. 5). According to Merchant and Van der Stede (2012) it is essential to align the MCS to the interest of agents and principals of a firm because failure can lead to large financial losses, reputation damage, and possibly even to organizational failure.
There are four management control system variables in a company defined by different authors and used in this survey research construct namely: behavioral, personnel, results and cultural controls, these could have significant impact on the outcome of more or less work tension due to tight or loose controls and also it may have a significant impact on the firm performance outcome due to it’s tight or loose controls in Professional Service Firms which may cause agency problems (Ouchi, 1979). The right balance of the MCS can alleviate these potential agency problems.
Auziar and Langfield-Smith (2005) mention in their study that there is a need for studies that not only offer useful understandings to the MCS studies in service organizations, but rather studies that also can assist in the development of instruments to measure the relevant constructs. This research focuses on new ways of measurement items of the different levels
of controls with existing items from literature and new items, therefore it is a clear request of this suggestion of Auziar and Langfield-Smith (2005).
This research focuses to what extent results controls tightness have a positive or negative effect on unit performance and also to what extent results controls tightness have a positive or negative effect on work tension. Jaworski et al. (1993) found evidence that high control systems produce positive effects and low control systems produce negative effects on the different management control system variables. Relevant to this research are the evidence found on the output control variable on job performance and person-role conflict which is similar to the construct of the effect of results controls on unit performance and work tension. Simons (1987) found that Prospectors generally use a lot of forecast data, set tight budget goals, monitor outputs carefully and emphasize frequent reporting with uniform control systems. He also states the need for more extensive research of the effects of the management control systems separately rather than grouped. The purpose of this study is to add to the limited body of knowledge of the effects of the design of MCS in service firms.
By researching this possibility, the contribution could be, if proven, that results control tightness has an influence on work tension and unit performance in Professional Service Firms. Professional Service Firms can have insight of the effects of control tightness on work tension and unit performance.
2 Theory
2.1 Literature review
To provide rationale for this research a literature review has been performed, this in relevance to this research and its objectives and to provide a framework of what it is composed out of. In the first part of this section an overview of the current state of knowledge regarding the agency theory, results control tightness of MCS, work tension, firm performance is given. Following this theoretical framework the hypothesis are presented in the second part of this section.
2.1.1 Agency theory
In companies there are bound to be conflicts between the principals (firm-owners) and the agents (managers), the agency theory looks at conflicts of interest that can arise between these parties with different interests in the same assets. The main conflicts of interest that the theory predicts are goal conflicts between the principals and agents, due to self-interest people will inevitably try to increase their own wealth.
The focus of this research is based on the tightness of results controls and there effect on work tension and unit performance.
According to Eisenhardt (1989, p. 3) “the theory is concerned with resolving two problems that can occur in agency relationships”. She also clarifies that the first is the agency problem that arises when the desires or goals of the principal and agent conflict and it is difficult or expensive for the principal to verify what the agent is actually doing. The problem here is that the principal cannot verify that the agent has behaved appropriately. The second is the problem of risk sharing that arises when the principal and agent have different attitudes toward risk. The problem here is that the principal and the agent may prefer different actions because of the different risk preferences. Because the unit of analysis is the contract governing the relationship between the principal and the agent, the focus of the theory is on determining the most efficient contract governing the principal-agent relationship given assumptions about people (e.g., self-interest, bounded rationality, risk aversion), organizations (e.g., goal conflict among members), and information (e.g., information is a commodity which can be purchased). Eisenhardt (1989) questions if a behavior-oriented contract (e.g., salaries, hierarchical governance) would be more efficient than an
outcome-oriented contract (e.g., commissions, stock options, transfer of property rights, market governance).
Since Jensen and Meckling developed the agency theory in 1976 for which they won a Nobel Prize, a lot has happened. It is still very difficult after so many years to assess the effects of the agency theory and moreover to solve these effects accordingly. For some time it has been apparent that the agency theory has its shortcomings. Jensen is a leading agency theorist who carried out empirical work together with Murphy (1990) in which they didn’t find a conclusive link between CEO pay and stock price performance. After several years Tosi et al. (2000) found similarly in a meta-analysis of 137 empirical studies, that incentive alignment as an explanatory agency construct for CEO pay was at best weakly supported by the evidence.
Up until now this has been the source of much discussion in academic literature and in practice. Incentives that are being awarded to CEO’s or employees are constantly being questioned, if there effective or that they rather have a negative effect on the performance of employees and also the cause of work tension on the work floor.
Results controls are based on outcome-oriented contracts, focusing on goals, targets and performance measures. There is a gap in prior research of both behavior based- as well as outcome based contracts that can potentially lead to undesirable actions by managers that affect unit performance and work tension. Therefore it is interesting to look at these relationships.
2.1.2 Management control systems
As mentioned in the introduction, management control systems (MCS) could be the solution for agency problems such as better unit performance and less work tension. MCS are implemented to control these problems and align the goal of the principal and the agent.
Firstly, a general explanation of MCS is given and the potential reasons that this can be the answer to agency problems and or unit performance and work tension. Thereafter a further discussion on more specific MCS such as described in the introduction.
There are different definitions in literature for MCS, but many of them have the same significance. MCS are present in organizations to align the interests of employees with organizational interests by influencing behavior. One of the earliest definitions comes from Lowe (1971, p. 5): “A management control system might be defined as a system of
organizational information seeking and gathering, accountability, and feedback designed to ensure that the enterprise adapts to changes in its substantial environment and that the work behavior of its employees is measured by reference to a set of sub-goals (which conform with overall objectives) so that the discrepancy between the two can be reconciled and corrected for.”
According to Merchant and Van der Stede (2012) there are three types of MCS: results controls, action controls and personal/cultural controls. Results controls mainly focus on goals and targets performance measures. With these measures the management can reward or punish their employees, by promoting or demoting them. Also the risk of dismissal can be a motivation for them to achieve the targets and goals set by management. Hall (1968) found that the type of profession matters in the outcome, if the profession receives relatively few rewards in a material sense, the level of dedication is likely to be higher. This demonstrates an opposite to the common perception of result controls, that if there were more rewards in a material sense that the level of dedication would likely be higher than when there isn’t.
In line with the behavioral- and outcome contracts as deliberated in the previous paragraph, in the next section the characteristics of implicit and explicit results control tightness will be discussed. Lastly I will discuss the effects of tight results controls on unit performance and work tension.
2.1.3 Results controls
As discussed in the previous section Merchant and Van der Stede (2012) differentiate between results controls, action controls, personnel controls, and cultural controls as four types of MCS. The first type of MCS, i.e. results controls, influences “actions because they cause employees to be concerned about the consequences of their actions they take” (p. 23). That means that the outcome of employee’s behavior is the objective of MCS. Pay-for-performance is a prominent example of a type of management control that can be called results control because it involves rewarding employees for generating good results, or punishing them for poor results. The rewards linked to results go far beyond monetary compensation and include, among others, job security, promotions, autonomy, and recognition. These controls encourage employees to discover and develop their talents and get placed in jobs in which they will be able to perform well. Thus, a well designed results control system should help produce the desired results.
According to Straub and Zecher (2013) results controls typically create meritocracies in which the highest reward is given to the person (or business unit) with the highest results. However they state that results controls like every other type of MCS cannot be used in every situation. A necessary requirement is that those whose actions influence the results, i.e. the controllability principle, and where the results can be measured effectively can control the results. Nevertheless, these requirements are fulfilled in many organizational situations and hierarchy levels, and nearly all managers in the firm can potentially use results controls. Merchant and Van der Stede (2003) contribute that results controls are particularly effective in addressing motivational problems because they “induce employees to behave so as to maximize their chances of producing the results the organizations desires” (p. 26)— without upper-level manager supervision. They also state that as with all other forms of management control, results controls cannot be used in every situation. They are effectively only where the desired result areas can be controlled by the employee whose actions are being influenced and where the controllable result areas can be measured effectively.
Furthermore, results controls besides influencing performance also influences the behavior of employees and thus it influences work tension as well.
2.1.4 Explicit and implicit results control tightness
To maximize the relevance of this research the focus is specifically on the explicit and implicit results control tightness. This survey-based research is set up of existing and new constructs and items developed by the lead of the research project at the Faculty of Economics and Business.1 The following definitions are taken from other studies and have newly added items as well by the leader.
Tight control is not defined consistently in the literature (Merchant and Van der Stede, 2011), and relatively few academics choose to define tight/loose control as a whole. Instead, most prefer to discuss the components or characteristics of tight control. Merchant and Van der Stede (2007) define control tightness as the “degree of certainty that employees will act as the organization wishes“ (pp. 118). Amigoni (1978) utilizes the definition of Dearden (1971) and Dalton (1971) who characterized a tight control system as one in which administrative control dominates over several (social) and individual control. Hopwood (1974) further expanded his own concept of tight/loose arguing that in tight control systems
1 Mrs. Kloosterman of the Faculty of Economics and Business at the University of Amsterdam sets up the
employee participation in setting objectives is low, targets are imposed on employees and seen as firm commitments and performance is based only on accounting measures. Loose control on the other hand is evidenced by managers which as conscious to the social side of control. Employee participation is high, targets are negotiated with employees and even when agreed upon are only seen as reference points to be considered in the context of other available information. Anthony, Dearden and Govindarajan (1992) base their definition of control of the extent of monitoring whereby a firm can be said to exercise tight control if management monitors the activities of the business unit frequently. This approach is also adopted by Hunton, Mauldin and Wheeler (2008) and Campbell, Epstein and Martinez-Jerez (2011). Whitley (1999) utilizes the interpretation of Lerner and Wanat (1983) and Butler (1991) where tight control implies that decision rules are precisely defined whereas in loose control systems precise rules may exist but the idiosyncrasies of the particular situation and the people involved are taken into account when deciding a course of action.
Tightness is defined as the degree of flexibility in the control system.2 Tightness can be created in two ways: 1) increasing the extent or scope of the MCS or 2) expanding the level of tolerance for deviations from the MCS. In the first case, tightness is achieved by creating more controls, more rules and more procedures. Increased tightness achieved through control selection, definition and completeness, is defined as explicit tightness. In the second case tightness is achieved by minimizing the difference in scope between the actions defined by the control system and those deemed acceptable within the organization. Tightness created by decreasing the level of tolerance for deviations from the MCS is defined as implicit tightness. Implicit and explicit tightness is then defined and operationalized for each type of control (behaviour, results, cultural, personnel) based on existing survey items (when available) as well as new items.
Explicit results control tightness can be defined in this survey based research construct as the extent of use of goals, targets and performance measures as part of the MCS, where a tight system is defined as one with a lot of controls in terms of amount and scope (Jaworski et al., 1993).
Implicit results control tightness can be defined as the degree to which deviation from goals, targets and performance measures is tolerated and/or encouraged, where a tight system
2 The explicit and implicit results control tightness construct is designed by Mrs. Kloosterman of the Faculty of
is defined as one which doesn’t permit any deviation from established goals, targets and performance measures (Hage and Aiken 1968, Simons, 1987, Van der Stede, 2001).
2.1.5 Work tension and unit performance
The effects of results control tightness on work tension and unit performance are the backbone of this research. Consequences of tight or loose MCS are split into two concepts.
Firstly, professional tension/stress, which is the degree to which the employees feel they cannot perform work in the way they believe it should be done (Rizzo et al., 1970, Aranya and Ferris, 1984).
Secondly, unit performance, which specifically measures the unit performance in an organization (King and Clarkson, 2015).
These consequences are relevant measures and issues in the working place, therefore they are very practical to revise with the information that will be collected.
2.2 Hypothesis development
There is a gap in literature regarding the effects of results controls on work tension and unit performance, specifically in Professional Service Firms. “The unique characteristics of pure services include (1) intangibility of services, (2) inseparability of production from consumption, where customers are also involved in the production of services, (3) perishability of services, where services not consumed are lost, and (4) heterogeneity in service products, where services provided by the same person may differ between customers or differ at different times” (Fitzgerald et al., 1991; Hope and Muhlemann, 1997; McColl et al., 1998 in Auzair and Langfield-Smith, 2005). Human participation in the delivery of services is an important feature in service organizations. Case studies in service organizations have suggested that the role of management control systems are different than manufacturing organizations, thus a reorientation is needed for an management control system to effectively be implemented within these types of organizations. (Abernethy and Stoelwinder, 1991, 1995). It is very interesting to research this topic because researchers have encouraged this area of research for future research. There is very little known on the behavioral element, thus tension, which is caused by results controls, performance has been measured more extensively in literature but not in the context that it is presented in this research, which is the influence of results controls on unit performance of Professional Service Firms.
Jaworski et al. (1993) found evidence that high control systems produce positive effects and low control systems produce negative effects on the different management control system variables. Relevant to this research are the evidence found on the output control variable on job performance and person-role conflict which is similar to the construct of the effect of results controls on unit performance and work tension.
Simons (1987) found that Prospectors generally use a lot of forecast data, set tight budget goals, monitor outputs carefully and emphasize frequent reporting with uniform control systems. He also states the need for more extensive research of the effects of the management control systems separately rather than grouped.
To conduct this research it has been chosen to do hypothesis testing, this because via the survey that was send out to the population regarding this research area there could be statistical evidence that can clarify the main question of this research which is:
What are the effects of results control tightness on work tension and unit performance in Professional Service Firms?
The above stated combined with a personal intrigue on this matter has led to the development of the following hypothesis:
H1a: Explicit results controls tightness increases work tension in Professional Service Firms.
H1b: Implicit results controls tightness decreases work tension in Professional Service Firms.
H2a: Explicit results controls tightness increases unit performance in Professional Service Firms.
H2b: Implicit results controls tightness decreases unit performance in Professional Service Firms.
These hypotheses are based on the explanations in the theory paragraph and expectations that flow out of theory (Jaworski, 1993; Simons 1987). If after the testing the hypotheses are not all supported, thus rejected, this has a meaning, and this matter will be clarified later on in this research. Due to the lack of research specifically on explicit and implicit results controls tightness in Professional Service Firms it is difficult to find them in existing literature. Nonetheless the formulated hypothesis can be measured with the survey construct, which provides a unique and original view on this area of research.
3 Research methodology
3.1 Sample and survey procedure
To test the hypotheses a survey-based research is conducted. First, the motivation for this topic was due to the gap in literature in management control systems and their effects on work tension and unit performance. Secondly, with the construct sources used in the questionnaire, the hypotheses can be operationalized. Thirdly, survey-based research is well suitable to examine the effects of control tightness on work tension and business unit performance in Professional Service Firms because the information gathered is from professionals that deal with this in their daily job. The dataset that is being used is original data.
Prior to sending the survey out to the professionals there was two separate pre-tests done to maximize the quality and effectiveness of the survey. Firstly, a pre-test was done to establish the quality of the items used to measure the variables for the types of management controls. By using an item sort task, subjects were provided with two separate documents, which of the first contained the construct definitions for the eight constructs used in the survey regarding the implicit and explicit nature of the results, behavior, cultural and personnel controls. The second documents stated fifty-two statements designed to test the constructs. Herewith, the subjects were asked to pair each statement to the definition they felt it most closely resembled. The fourteen subjects that participated in this pre-test were from a diverse array of professionals from different fields. Forthcoming out of this pre-test are the thirty-two statements for inclusion in the survey for each type of control with the least number of incorrect pairs identified by the participants. Secondly, another pre-test was done to assess the quality of the survey as a whole. For this testing an additional number of twenty participants took part of this pre-test. They were asked to view the survey online and assess the quality of the content, clarity and appearance of the survey as well as the amount of time required to complete the survey. Flowing out of this pre-test came comments from the participants regarding the quality of these points and resulted in minor changes in wording and the inclusion of additional answers by a few multiple choice questions.
The sample for this research consists of professionals that have three years work experience or more, but less than ten years work experience. The respondents are out of the apprenticeship and or learning phase of their job and are able to work relatively independently with minimal guidance from their supervisor. However, the respondents are
not in a position to exercise power over the direction of the organization. The respondents are not an owner/partner or board members of the organization. In other words, the respondents need to be a subject to the management accounting and control system rather than design it. The respondents work in medium/large size organizations (> 50 employees). The respondents must speak and understand English at a business level. The collected information through the survey is original and relevant for this research.
The survey was send out to the personal network of students that took part in this research project of the Faculty of Economics and Business at the University of Amsterdam, which adhered to the target group of this research. Due to the efforts of these students, a total of 372 responses were received of which 75,26 % were usable. A data screening procedure was undertaken to analyze missing responses. Also for the following reasons some answers of respondents were removed: surveys that weren’t finished (47), respondents with less than three years work experience in their current employment field (36) and respondents that stated that they did not read the terminology pertaining to the survey (9).This step further reduced the sample to 280 responses.
3.2 Operationalization of the variables
The dependent variables in this research are work tension and unit performance. For each of the hypotheses the operationalization is sketched below:
H1a: Explicit results controls tightness increases work tension in Professional Service Firms.
Control tightness is operationalized as explicit results control tightness and work tension is operationalized as professional tension in the survey (figure 1.1).
Figure 1.1 Operationalization of H1a
H1b: Implicit results controls tightness decreases work tension in Professional Service Firms.
Control tightness is operationalized as implicit results control tightness and work tension is operationalized as professional tension in the survey (figure 1.2).
Figure 1.2 Operationalization of H1b
H2a: Explicit results controls tightness increases unit performance in Professional Service Firms.
Control tightness is operationalized as explicit results control tightness and firm performance is operationalized as unit performance in the survey (figure 1.3).
Figure 1.3 Operationalization of H2a
H2b: Implicit results controls tightness decreases unit performance in Professional Service Firms.
Control tightness is operationalized as implicit results control tightness and firm performance is operationalized as unit performance in the survey (figure 1.4).
The measurements of these variables are taken place via the collected data from the survey. These variables are being linked to questions in the survey, in different constructs, these constructs of measurement will be further discussed in the following chapter.
4 Statistical analysis
In this chapter we will start with the statistical analysis. First by showing demographics of the respondents of the survey used for this study. Secondly we will share some information on the nature of the Professional Service Firms. Thirdly we will explain how the independent, dependent and control variables in this study were measured in the survey and if the constructs are appropriate for further analysis.
4.1 Demographics
Table 1 presents the demographics of the respondents of the survey. The respondents are 61,3% male and 38,7% female; the mean age of the respondents is 35 years with a standard deviation of 7,8. Fifty-two percent of the respondents have a Master degree, thirty-six-point eight have a Bachelor degree or lower and eleven-point twenty percent have a PhD or other professional doctorate degree. The occupations are also given in Table 1, the largest groups are: accounting professionals (25%), consulting management or strategic professionals (8,9%), medicine or physician practices professionals (8,2%), also a lot of respondents chose “other” as their occupation because it was not in the list given (19,3%).
range mean SD median % Gender
Male N/A N/A N/A Male 61,3
Female 38,7
Age N/A 35,00 7,80 33,00 N/A
Education
Bachelor degree or lower N/A N/A N/A Master 36,8
Master degree 52,0
PhD or other professional doctorate degree 11,2
Occupations
Accounting N/A N/A N/A Accounting 25,0
Consulting Management/ Strategic 8,9
Medicine/ Physician practices 8,2
Consulting IT 5,4
Marketing/public relations 4,6
Risk management services 3,6
Consulting HR 3,2 Law 3,2 Financial advising 2,5 Investment banking 2,1 Engineering 1,8 Recruiting - executive 1,8 Software development 1,8 Project management 1,4 Real estate 1,4
Investment management (hedge funds) 1,1
Pharmaceutical 1,1
Research/R&D 1,1
Architecture 0,7
Consulting Technology 0,7
Media production (film, TV, music) 0,4
Graphic design 0,4
Advertising 0,4
Other 19,3
Table 1: Demographics of the respondents
Other interesting factors to know about the respondents are about their firms, in table 2 we sum up some of these factors. The firms are located for the majority of 61,8% in The Netherlands, 10% in Germany, 6,4% in Aruba, 5% in the USA and the rest compromises of different other countries. The total size in regards to employees of the organizations, 37,4%
of the respondents are from an organization of more than five thousand employees, followed by 26,6% of the respondents are from an organization with more than on hundred but less than five hundred employees, 21,6% from an organization with more than five hundred but less than five thousand employees and the smallest group is from respondents from organizations of that employ less than one hundred employees (14,4%). Another aspect important to know are the size of the organizational unit that the respondent works in, 38,7% of the respondents work in a unit which consists of more than 10 but less than 50 employees, 26,9% work in a unit which consists of more than 500 employees, 17,9% in a unit with less than 10 employees and 16,5% in a unit of more than 50 but less than 500 employees. The ownership structure of a firm says a lot of the type of MCS that is in place as mentioned in the control variable paragraph, therefore the results from the survey are that 61,1% of respondents work in a firm that is owned by a partner that is working in the firm, 27,5% are from a firm that is owned by investors or shareholders and 11,4% is part of a non-profit or public organization.
Table 2: Information of the Professional Service Firms
range mean SD median %
Firm location N/A N/A N/A Netherlands
Netherlands 61,8
Germany 10,0
Aruba 6,4
United States 5,0
Other countries outside the Netherlands 16,8
Size of the organization 1 - 4 N/A N/A > 5000
Less than 100 employees 14,4
More than 100 but less than 500 employees 26,6
More than 500 but less than 5000 employees 21,6
More than 5000 employees 37,4
Size of the organizational unit 1 - 4 N/A N/A > 10 < 50
Less than 10 employees 17,9
More than 10 but less than 50 employees 38,7
More than 50 but less than 500 employees 16,5
More than 500 employees 26,9
Ownership structure 1 - 3 N/A N/A Partnership
Partnership 61,1
Investors/ Shareholders 27,5
4.2 Measurement of variables
In this paragraph the constructs to measure the variables will be elaborated. There are different items in the survey that are expected to measure a certain construct. As mentioned in chapter three pre-testing of the survey was done to test these items, and also some of them are based on previous literature. All the reason that the probability that a construct is wrongfully measured is minimized. The survey gives opportunity for different areas of research of MCS in Professional Service Firms, due to the measurement of a big number of items on specific constructs. Statements in the survey were positively and negatively formulated on some constructs. By doing this it is checked for the probability that a respondent fills every statement positively or negatively without noticing the difference (Bryman, 2012). These items are recoded for further analysis. In the following sub paragraphs the most important variables for this research are discussed and the underlying constructs are tested for reliability and internal validity.
4.2.1 Result controls tightness
The independent variable in this research is results control tightness. The constructs to measure control tightness are adapted from prior research by (Bodewes, 2000), (Jaworski et al., 1993), (Hage and Aiken, 1968), (Simons, 1987), (O’Reilly and Chatman, 1986) and (Van den Stede, 2001).The degree of flexibility of the control system is measured via its tightness. The degree of tightness is measured via different variables with statements regarding the extensiveness of the MCS (explicit tightness) and statements regarding the degree of tolerance for deviation from the standards (implicit tightness). The tightness is measured via implicit and explicit tightness of the four forms of controls out of the objects-of-control (OOC) framework (Merchant and Van der Stede, 2012) and are constructs that have not been previously used. For the tightness of result controls, in some items it was possible to base them on measurement instruments from previous scientific research.
The Keyser-Meyer Olkin criteria en the Bartlett’s test of Sphericity show that the data is appropriate for factor analysis. Out of this analysis analysis forthcomes that there are two underlying components from the eight result control tightness items. The screeplot also illustrates that two factors should be kept. These two factors explain 61,9% of the variance. Followingly it is assessed if these two factors and there underlying items load well on the respective component.
For the independent variable the respondent has to answer eight questions on a five-point Likert scale, these are divided in implicit and explicit result controls tightness measures.
Explicit results control tightness is measured by the responses of four questions on a five-point Likert scale, e.g. “In my job, there is a performance measure for everything” (appendix 1). Respondents were asked to indicate the extent to which the four statements best described the type of results controls in their organization by choosing a number on a five-point Likert scale. The score of 1 and 2 (strongly disagree, disagree) indicated loose explicit results control, 3 neutral, and a score of 4 and 5 (agree, strongly agree) indicated the presence of tight explicit results control in their organization. To assess the discriminant validity, factor analysis was conducted on explicit results control tightness constructs. By clustering of items in one dimension, errorneous items that do not load on the same factor can be removed to improve validity and reduce the dataset for it to be more structured appropriately (Bryman, 2012). The varimax rotated principle component analysis (based on eigen values greater than 1) gave one underlying factor for explicit result control tightness, but one item “in my organization employees are expected to meet pre-established goals/targets with no exceptions” loaded with 0,651 on the explicit factor but originally in the survey construct this should measure implicit result control tightness. After analysis of the question there was concluded that this item isnt a good fit with the explicit result control tightness construct and therefore there was decided to remove this item from the analysis. After deletion of this item, another factor analysis was done, the factor loadings are recorded in table 3. Although the results control tightness framework developed in this study was based on theoretical foundations, no study has tested whether the items when grouped together, measure the same underlying concept, especially in Professional Service Firms. Therefore, a test of the reliability of the scale was conducted in which the Cronbach’s alpha is .813, which is above the limit of acceptability for exploratory research, generally considered to be around .70 (Nunnally, 1978). The variance explained by the scale is 11,85%. The scores for the four items related to explicit results control were afterwards summated into a composite score variable for further analysis
Also, the independent variable implicit results control tightness is measured by the responses of four questions on a five-point Likert scale. Respondents were asked to indicate the extent to which the four statements best described the type of implicit results controls in their organization by choosing a number on a five-point Likert scale, e.g. “in our organization, goals/targets are essentially a guideline rather than a true commitment”
(appendix 1). The score of 1 and 2 (strongly disagree, disagree) indicated loose implicit results control, 3 neutral, and a score of 4 and 5 (agree, strongly agree) indicated the presence of tight implicit results control in their organization. Three of the four items were negatively worded and had to be reverse coded for further analysis. It is probable that the negatively worded statements were not well interpreted amongst the respondents. This shows due to a low cronbach’s alpha of -1,844, after deletion of the item “responding to new, unforeseen opportunities is considered more important by my supervisor than achieving pre-established goals/targets” the cronbach’s alpha coefficient went up to ,299 which is also too low and there is no more possibilities to increase this coefficient. Also there is a low correlation amongst the three items that remained in the construct (less than 0,30). Therefore it is concluded that there isnt sufficient reliability for this scale, which underlying items should measure the implicit result control tightness construct. As a consequence of this analysis it is decided that it was not possible to compose a variable for implicit result control tightness with the underlying items used in this construct of the survey. Moreover, it is assesed that we cannot continue with the testing of hypothesis H1b and H2b due to this discovery.
4.2.2 Work tension
The constructs to measure work tension (dependent variable) are adapted from prior research by (Rizzo et al., 1970) and (Aranya and Ferris, 1984). For the dependent variable work tension, the respondent has to answer eight questions on a five-point Likert scale, e.g. “In my organization, there is a conflict between the work standards and procedures of the organization and my own ability to act according to my professional judgment” (appendix 2). Respondents were asked to indicate the extent to which the eight statements best described the professional tension emphasized in their organization by choosing a number on a five-point Likert scale. The score of 1 and 2 (strongly disagree, disagree) indicated a low work tension, and a score of 4 and 5 (agree, strongly agree) indicated the presence of high work tension in their organization. One item was negatively worded, this was recoded to be consistent with the construct.
The Keyser-Meyer Olkin criteria (0,892) and the Bartlett’s test of Sphericity (p < ,000) show that the data is appropriate for factor analysis. Out of this analysis analysis forthcomes there is only one underlying component from the eight profesional tension items in the survey construct. The screeplot also illustrates that one factors should be kept, exactly as predicted by the construct.
The varimax rotated principle component analysis (based on eigen values greater than 1) gave one underlying factor for work tension (table 3). Although the professional tension framework developed in this study was based on theoretical foundations, no study has tested whether the items when grouped together, measure the same underlying concept, especially in Professional Service Firms. A test of the reliability of the scale was conducted in which the Cronbach’s alpha was .819, above the upper limit of acceptability for exploratory research, generally considered to be around .70 (Nunnally, 1978), this indicates a high level of internal consistency with this specific sample. The variance explained by the scale was 31,60% forthcoming out of factor analysis on convergent validity. The scores for the eight items were summated into a composite score for further analysis.
4.2.3 Unit performance
The constructs to measure unit performance (dependent variable) are based on prior research by (King and Clarkson, 2015) for unit performance.
For the dependent variable unit performance, the respondent has to answer six questions on a five-point Likert scale (appendix 3). Respondents were asked to indicate the extent to which the six statements best described the unit performance in their organization by choosing a number on a five-point Likert scale. The score of 1 and 2 (strongly disagree, disagree) indicated low unit performance, 3 neutral, and a score of 4 and 5 (agree, strongly agree) indicated the presence of high unit performance in their organization.
The Keyser-Meyer Olkin criteria (,783) and the Bartlett’s test of Sphericity (p < ,000) show that the data is appropriate for factor analysis. Out of this analysis analysis forthcomes there are two underlying components from the six items in the survey construct. The screeplot also illustrates that two factors should be kept, different from what the construct is supposed to show (table 3).
The varimax rotated principle component analysis (based on eigen values greater than 1) gave two underlying factors for unit performance. After analysis of the two questions “compared to other organizations, generally my organization: is more innovative” and “compared to other organizations, generally my organization: is larger in size” that loaded on a different factor with the other four, was concluded that these items are a good fit with the unit performance construct on basis of the inter item correlations and high cronbach’s alpha that would not improve if items are deleted.
Although the unit performance framework developed in this study was based on theoretical foundations, no study has tested whether the items when grouped together, measure the same underlying concept, especially in Professional Service Firms. Therefore, a test of the reliability of the scale was conducted in which the Cronbach’s alpha was .81, above the upper limit of acceptability for exploratory research, generally considered to be around .70 (Nunnally, 1978). The variance explained by the scale was 21,13%. The scores for the six items were summated into a composite score for further analysis.
4.3 Control variables
A survey in general, but specifically the comprehensive design of this survey offers many opportunities to control for other factors and thus improve internal validity. The selected control variables for this research are firm size, firm reputation and firm ownership. These were selected based on the potential influence they can have on control tightness and firm performance. A correlation test was done in relation to unit performance and work tension, here outflowing it was decided to include these variables in the statistical model.
4.3.1 Firm size
A size of a firm can say a lot about the MCS that is employed and for that matter to the relation between results control tightness with unit performance and work tension. To test firm size there are two constructs in the survey (appendix 6). The first item asks in which of the following groups the firm the respondent works for in belongs, namely in the size of the organization: (1) less than 100, (2) more than 200 but less than 500, (3) more than 500 but less than 5000 and (4) more than 5000. The second item requires the respondents to select a group on how many employees are there in the organization unit that the respondents works in, the options are: (1) less than 10, (2) more than 10 but less than 50, (3) more than 50 but less than a 100, (4) more than 100. A test of the reliability of the scale was conducted in which the Cronbach’s alpha was .71, above the upper limit of acceptability for exploratory research, generally considered to be around .70 (Nunnally, 1978). The descriptives can be seen in table 3 and also table 2.
4.3.2 Firm reputation
Firm reputation is an interesting measure in this research context, the reputation of a firm can influence the performance of the firm by the means of attracting new clients and attaining current clients, the negative effect of reputation can be losing clients due to occurrences that
damage a firm’s reputation. Combs and Ketchen (1999) found that the reputation of a firm is positively associated with financial performance. The construct used in the survey to measure firm reputation consists of four items, e.g. “my organization is well respected in its field” and three more (appendix 4). Respondents had to state their opinion on the statement on the basis of a five point Likert scale: (1) strongly disagree, (2) disagree, (3) neutral, (4) agree, (5) strongly agree. Meaning that if the respondent responds with a high number on the scale (4 or 5) it would mean that the firm has a high reputation. The Keyser-Meyer Olkin criteria (.729) and the Bartlett’s test of Shpericity (p <.000) show that the data is adequate for factor analysis. The screeplot shows a confirmation with literature, there is a clear pattern in the data, there is one underlying factor that measures 37,17% of the variance. A varimax rotated principal component analysis was done, showing one underlying factor with eigen values higher than 1. All items load above 0.5 on a factor. This construct has a high Cronbrach’s alpha of .80, above the upper limit of acceptability for exploratory research, generally considered to be around .70 (Nunnally, 1978). The scores for the four items were summated into a composite score for further analysis.
4.3.3 Firm ownership
Another aspect that has an impact on the choices that are made in regards of the MCS of a firm is the firm ownership structure. King and Clarkson (2015) state that dependent on the ownership structure of a firm differential MCS choices should be made. Therefore, the firm ownership structure can possibly have an effect on the relationship between results control tightness and firm performance and work tension in the firm. To acquire this information from the respondents, a construct in the survey asks of the respondents to state if the organization is a partnership, shareholders/investors or non-profit/public organization (appendix 5). The descriptives can be seen in table 2.
5 Findings
In the last chapter of this research, the findings on the basis of the hypothesis are done and also the statistical tests used for these testing are explained. Thereafter the limitations of the study are stated in a separate paragraph in combination with guiding avenues for future research on this topic and target group.
5.1 Descriptives
Followingly, table 3 highlights all the relevant descriptive statistics for the underlying constructs of this research and there reliability, internal validity. For the constructs that complied with the requirements a composite score was summated with the underlying items for further analysis.
Hereafter the correlations between the variables were tested, the correlation analysis is reported in table 4. Herein the possible lineair association between two variables is measured via the Pearson correlation method and the Spearman tho’s method.
Examination of the correlation matrix indicates that there are no correlations greater than .70 between the contingent variables. Thus, there is some reason to believe that multicollinearity is unlikely to be a problem (see Tabachnick and Fidell, 2001, p. 84). The Pearson correlations of explicit result control tightness and work tension (r=0.107) are positively but yet not significantly correlated, on the other hand explicit result control tightness and unit performance are positively significant correlated via the pearson correlation (r=0.149; p<0.05). The Spearman correlation coefficients state a similar outcome, explicit result controls tightness and work tension are positively significant correlated (r=0.141; p<0.05), and explicit result control tightness and unit performance are also positively significant correlated (r=0.159; p<0.01). These correlations are in line with the positively formulated hypotheses. The other significant correlations found via the Pearson correlations are: a significant negative association of firm reputation with work tension (r=-0.185; p<0.01), a significant positive association of firm reputation with unit performance (r=0.412; p<0.01), a significant positive association of firm size with unit performance (r=0.134; p<0.05). significant positive association of firm size with firm reputation (r=0.149; p<0.05), a significant negative association of firm ownership with unit performance 0.153; p<0.05), a significant negative association of firm ownership with firm reputation
(r=-0.156; p<0.01), a significant positive association of firm ownership with firm size (r=0.134; p<0.05).
Table 3: Descriptive statistics
Variable / Items
Factor loadings
Cronbach's
Alpha N Scale range Mean
Standard deviation
Explicit result controls 0,813 2,87 0,859
There is a performacne measure for everything 0,823 276 1 - 5 2,72 1,114 Organization sets a large number of goals/targets that
employee has to meet 0,820 276 1 - 5 2,99 1,065
Constant check of employee goal/target attainment 0,774 276 1 - 5 2,89 1,065 Supervisor checks if employee is meeting targets 0,755 276 1 - 5 2,89 1,057
Work tension 0,819 2,36 0,699
Not having enough time to complete work, the way that
the employee thinks it should be done 0,796 269 1 - 5 2,85 1,143 Due to lack of adequate resources and materials, employee
cannot execute assignments properly 0,785 269 1 - 5 2,12 1,015 Presence of conflict between the work standards and
procedures of the organization and employee's own ability
to act according to professional judgement. 0,762 269 1 - 5 2,24 0,916 The organization hinders the employee from doing work
properly 0,715 269 1 - 5 2,28 1,079
The extent to which employee has to alter their
professional behavior in order to perform their job the way
the organization wants them to. 0,698 269 1 - 5 1,96 0,980
Employee cant perform their job the way they think they
should in this organization 0,565 269 1 - 5 2,50 1,115
Employee could do their job much better without the
conditions imposed by my organization 0,533 269 1 - 5 2,39 1,149 The type and structure of the employment gives employee
the opportunity to fully express themselves as a
professional 0,427 269 1 - 5 2,43 1,051
Unit performance 0,812 3,36 0,763
Compared to other organizations, my organization:
is more competitive 0,827 278 1 - 5 3,41 0,960
has greater market share 0,801 278 1 - 5 3,40 1,092
is growing faster 0,746 278 1 - 5 3,29 1,046 is more profitable 0,682 278 1 - 5 3,21 1,075 is more innovative 0,860 278 1 - 5 3,50 1,040 is larger in size 0,823 278 1 - 5 3,38 1,180 Firm reputation 0,807 My organization:
is well respected in its field 0,863 274 1 - 5 4,32 0,853
is perceived to provide good value for price 0,855 274 1 - 5 3,99 0,858 has a strong reputation for consistent quality and service 0,735 274 1 - 5 4,27 0,832 has strong brand name recognition in its service area 0,691 274 1 - 5 4,22 0,897
Table 4: Correlation matrix Explicit result controls Work tension Unit performance Firm
reputation Firm size
Firm ownership Explicit result controls 0.141* 0.159** 0.069 -0.025 -0.068
(0.018) (0.008) Work tension 0.107 -0.106 -0.202** 0.078 0.095 (0.077) (0.001) Unit performance 0.149* -0.105 0.434** 0.167** -0.119* (0.013) (0.000) (0.005) (0.047) Firm reputation 0.042 -0.185** 0.412** 0.167** -0.128* (0.002) (0.000) (0.005) (0.033) Firm size -0.063 0.063 0.134* 0.149* 0.124* (0.025) (0.013) (0.038) Firm ownership -0.091 0.098 -0.153* -0.156** 0.134* (0.010) (0.009) (0.025)
Pearson correlations appear below the diagonal, non-parametric Spearman correlations appear above the diagonal
(p values between brackets)
** Correlation is significant at the 0.01 level (two-tailed) * Correlation is significant at the 0.05 level (two-tailed)
The other significant correlations found by the Spearman correlation analysis are: a significant negative association of firm reputation with work tension (r=-0.202; p<0.01), a significant positive association of firm reputation with unit performance (r=0.434; p<0.01), a significant positive association of firm size with unit performance (r=0.167; p<0.01). significant positive association of firm size with firm reputation (r=0.167; p<0.01), a significant negative association of firm ownership with unit performance (r=-0.119; p<0.05), a significant negative association of firm ownership with firm reputation (r=-0.128; p<0.05), a significant positive association of firm ownership with firm size (r=0.124; p<0.05).
Both methods give clear correlations as stated previously, very similar to each other, all significantly correlated are the same with the exception that with the Spearman tho’s coefficient give a significant correlation for work tension and explicit result controls whereas the pearson correlation only gives a positive view of the correlation between these two variables but not the significance. To measure the causality of the positvely correlated variables for this research a linear regression is performed.
5.2 Linear regression
Field (2013) defines a linear regression as a statistical process which is for estimating the relationships among variables. It includes many techniques for modeling and analyzing several variables, in this research the focus is on the relationship between a dependent variable and one independent variable. More specifically, regression analysis helps one understand how the typical value of the dependent variable changes when the independent variables are varied, while the other independent variables are held fixed. In this paragraph the we will elaborate on the assumption of linear regression in regards to the framework that is built to measure these variables and also the results of the linear regressions are presented. 5.2.1 Assumptions testing
The assumption of normality has to be done for a linear regression. After a visual inspection of the histograms and Q-Q plots, it showed that explicit result control tightness had an approximate normal distribution, with a skewness of 0.144 (SE=0.146) and a kurtosis of -0.313 (SE=0.291), the dependent variable unit performance also showed an approximate normal distribution, with a skewness of -0.209 (SE=0.146) and a kurtosis of 0.454 (SE=0.290), work tension on the other hand showed that it was a bit skewed to the left 0.317 (SE=0.146) and an approximate normal kurtosis coefficient -0.298 (SE=0.291) (Cramer, 1998; Cramer & Howitt, 2004; Doane & Seward, 2011). They provide the rule that if the absolute value of the skewness is less than three times standard error (SE) the data isn’t significantly skewed if this is indication is however violated than there is a skewness issue present. To bring the skewed data to an acceptable level for parametric testing the work tension variable is transformed and as a result it has a more approximate normal distribution with a skewness of -.0409 (SE=0.146) and a kurtosis of -0.189 (SE=0.291). A Komogorov-Smirnov test was performed and both the independent as the dependent variables data came out as having a deviation from normality (p<0.05). A deviation of normality, as in this case, is not always problematic, as long as it’s not too severe. The deviation of normality is caused by the skewness of the data in this case as reported by the minor skewness of work tension and also by the ordinal nature of the data. A test for outliers was performed in which a calculation was made based on the rule of thumb to delete outliers which were three standard deviations from the mean. A calculation was conducted, results control tightness and work tension did not have values that fall beyond this rule, however unit performance (mean=3,3683; SD=0,76280) had three smaller values (1.00) than the lower value based on this rule of thumb (upper=5,6567; lower=1,0799) which in this case is deemed acceptable
seeing the ordinal nature of the data and also that the deletion of these cases are not urgent, thus it is decided to leave these three cases in. There are no issues regarding multicollinearity, the VIF (table 5 & 6) is for all the variables under the thresholds of 5.0 and 10.0 (O’Brien, 2007 and Field, 2013).
5.2.2 Hypotheses findings
To test the causality of the variables, work tension (Spearman correlation p<0.05) and unit performance (Spearman correlation p<0.01) which had a significant Spearman tho’s correlation with explicit results control tightness a linear regression analysis is performed to assess the ability of work tension and unit performance to predict explicit results control tightness. In table 5 the standardized beta coefficients, t-values and significance levels and collinearity statistics of the variable work tension are showed and in table 6 the standardized beta coefficients, t-values and significance levels and collinearity statistics of the variable unit performance are showed.
H1a predicted that explicit results controls tightness would increase work tension in professional service firms. Firstly, model 1 (table 5) entails the regression of the control variables firm reputation, firm size and firm ownership as independent variables and dependent variable work tension. The variance analysis (Anova) in model 1 is significant (p=0,003), with an explained variance on the dependent variable (work tension) of 4,3% (Adj. R2). We also see that the dependent variable (work tension) and the control variables firm reputation (β= -0,217, p= 0,000) has a negative significant effect on work tension, firm size (β= 0,094, p= 0,119) has a positive effect on work tension but isn’t significant, firm ownership: partnership (β= -0,065, p= 0,496) and firm ownership: shareholders/investors (β= -0,013, p= 0,889) show a negative effect on work tension but isn’t significant. This means that organizations that are a partnerships have less work tension than nonprofit organizations (nonprofit organizations were used as a basis of this construct), and also that organizations that are owned by shareholders or investors have less work tension than nonprofit organizations. Also the significant effect of the degree of firm reputation on work tension is explained as being a decrease of one standard deviation of degree of firm reputation, resulting that the standard deviation of work tension (dependent variable) negatively changes with the standardized Beta value of -0,217, meaning that the higher degree of good firm reputation, the higher the work tension in the organization. Also the insignificant effect of firm size on
work tension is explained as being an increase of one standard deviation of degree of firm size, resulting that the standard deviation of work tension (dependent variable) positively changes with the standardized Beta value of 0,094, meaning that the larger the size of the firm, the higher the degree of work tension in the organization.
Secondly, model 2 entails the regression of the same control variables as model 1 with the insertion of the independent variable in this research which is explicit results control tightness. The variance that this model explains is larger than model 1 at 5,9% (Adj. R2). We also see that the dependent variable (work tension) and the control variables firm reputation (β= -0,219, p= 0,000) still has a negative significant effect on work tension, firm size (β= 0,102, p= 0,091) still has a positive effect on work tension but isn’t significant, firm ownership: partnership (β= -0,096, p= 0,314) and firm ownership: shareholders/investors (β= -0,049, p= 0,606) show a negative effect on work tension but isn’t significant just as in model 1. The effects of the control variables remain unchanged and are the same as in model 1. The independent variable explicit result control tightness shows a positive significant relationship with work tension (β= 0,140, p= 0,019). The linear regression of model 2 (table 5) shows that there is a significant causal relationship between explicit results control tightness and work tension (p<0.05). Preliminary analyses were performed to ensure there was no violation of the assumption of normality and linearity. The standardized Beta of work tension (0.140) shows that for an increase of one standard deviation in explicit results control tightness (independent variable), the standard deviation of work tension (dependent variable) positively changes with 0.140 (table 5). The significance of this positive change is zero-point-zero-one-nine (p<0.05) and is significant, thus H1a is supported that explicit result controls tightness increases work tension in professional service firms.
H2a predicted that explicit results controls tightness would increase unit performance in professional service firms (table 6). Thirdly, the linear regression of model 3 entails the regression of the control variables firm reputation, firm size and firm ownership as independent variables and dependent variable unit performance. The variance analysis (Anova) in model 3 is significant (p=0,000), with an explained variance on the dependent variable (unit performance) of 17,3% (Adj. R2). We also see that the dependent variable (unit performance) and the control variables firm reputation (β= 0,377, p= 0,000) has a positive significant effect on unit performance, firm size (β= 0,093, p= 0,097) has a positive effect on unit performance but isn’t significant, firm ownership: partnership (β= 0,162, p= 0,068) and firm ownership: shareholders/investors (β= 0,087, p= 0,319) show a positive effect on unit
