The effect of organizational reputation on the relationship between organizational performance and management control tightness

(1)

The effect of organizational reputation on the relationship

between organizational performance and management control

tightness

Name: Hélène van Zuijlen Student number: 11361700

Thesis supervisor: Prof. dr. ir. M.J.F. Wouters Date: June 25, 2018

Word count: 18257

MSc Accountancy & Control, specialization Control

(2)

Statement of Originality

This document is written by student Hélène van Zuijlen 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.

(3)

Abstract

In this thesis the effect of organizational performance on management control system (MCS) design in professional service firms (PFSs) is examined. MCS design was explained on the basis four constructs: action control tightness, results control tightness, personnel control tightness and cultural control tightness. Additionally, organizational reputation was examined as a moderator to be positively influencing the effect of organizational performance on management control tightness. The research was conducted with a survey that was created by Helena Kloosterman of the PSF Thesis Survey Project 2017-2018. The survey was completed by service professionals who are working in PSFs. The data, gathered from 519 respondents in the final sample, led to the following results: (1) organizational performance is exponentially positively related to results control tightness, personnel control tightness and cultural control tightness, and (2) no significant evidence has been found to support the theory that organizational reputation is moderating these relationships.

Key words: management control system; control tightness; professional service firms; service professionals; organizational performance; organizational reputation

(4)

1 Introduction

In the field of management control research little research has been done to investigate how professional service firms (PSFs) design their management control system (MCS) and what the determinants and the effects of the design of their MCS are. Previous literature about PSFs concern examining their ownership structures and its relation to firm performance (King and Clarkson, 2015), whether public ownership is bad for PSFs (Von Nordenflycht, 2007) and whether the service process type, amongst other constructs, influenced the design of the MCS (Auzair and Langfield-Smith, 2005).

One of the reasons why management control of PSFs is less examined than management control of other kinds of organizations is the fact that PSFs are more difficult to examine (Suddaby et al., 2008). PFSs highly rely on the complex knowledge of its service professionals (Von Nordenflycht, 2010), who’s tasks are hard to understand for other people (Greenwood et al., 2005), who are treated differently by management because of this (Barley & Tolbert, 1991), and whose characteristics often lead to conflicts with management control (Abernethy & Stoelwinder, 1995; Raelin, 1985; Løwendahl, Revang, & Fosstenløkken, 2001). In order for management to control its service professionals, they design the MCS by making choices about its tightness. Management control tightness is associated with extensive controls, from which only little deviation is allowed. This construct of management control tightness is explained for four kinds of management control: action, results, personnel and cultural control (Merchant and van der Stede, 2017).

The aim of this thesis is to deepen the understanding about the MCS design in PSFs, specifically the degree of management control tightness for each of the four types of management control. Therefore this thesis contributes to the existing literature by further investigating MCS design in PSFs. Two constructs that can influence this design will be examined: organizational performance and organizational reputation. First, the effect of organizational performance on management control tightness is investigated. This has not been researched before. In addition, the effect of organizational reputation on the previously mentioned relationship will be examined, which is therefore also contributing to the existing literature, as this has not been researched before as well. The research question of this thesis is thus as follows:

What is the effect of organizational reputation on the relationship between organizational performance and management control tightness?

This research question is relevant for several reasons. First of all, Greenwood et al. (2005) state that the PSF industry is an important industry to examine because the interest in knowledge as a competitive asset is growing and because PSFs represent a significant part of the economy.

(7)

Second of all, Auzair and Langfield-Smith (2005) found that mass service firms use more bureaucratic MCSs than PSFs. From this we can derive that PSFs use less bureaucratic MCS. It is often suggested that use of less bureaucratic MSC means the focus is more on results controls as opposed to action controls (Abernethy and Stoelwinder, 1991; Chenhall, 2003; Ouchi, 1979). This apparent difference in the use of controls for PSFs makes it relevant to better understand how PSFs design their MCS. Organizational performance is relevant to examine in PSFs, as its output is produced by its hard-to-manage service professionals. Organizational reputation is an interesting construct to examine for PSFs as well, since it is one of the key success drivers for these kind of firms (Kaiser & Ringlstetter, 2010).

The data for this research was gathered by a survey, that was filled out by service professionals. The survey was created by Helena Kloosterman, who is supervising the PSF Thesis Survey Project 2017-2018. This project was joined for the purpose of examining the predicted relationships in this thesis. The results of this thesis show significant exponential positive effects of organizational performance on several types of management control tightness. This means that evidence has been found that the construct of organizational performance is shaping the MCS design in PSFs. High organizational reputation does not have a significant effect on the relationship between organizational performance and management control tightness. Thus, it can be said that the level of organizational reputation is not taken into consideration by management while deciding on tightening management controls when the PSF is climbing the performance ladder.

This thesis is structured as follows: first, an overview of the literature will be given that will explain the constructs that are used to answer the research question. Second, two theories are described, which results in the development of eight hypotheses. Third, the research method that was used to examine the predicted relationships will be discussed and the validity and reliability of the method will be assessed. Fifth, the results of the statistical testing of the hypotheses will be given. Last but not least, the results, limitations and implications for future research will be discussed.

(8)

2 Literature Review

The objective of this literature review is to give an understanding of the different concepts that are used in this thesis. First, the construct of a management control system is explained. Second, the construct of control tightness is explained and what this means for action, results, personnel and cultural control. Third, the construct of the professional service firm is described by analyzing literature about services, service professionals and professional service firms. Last but not least, the constructs of organizational performance and organizational reputation are analyzed.

2.1 Management Control System

In order to understand what a MCS is, it is important to know why it exists. The bigger and more complex an organization becomes, the more ownership and control become separated. A growing organization often goes from several owners that simultaneously manage the organization, to a structure with many owners (shareholders) and a small group that manages the organization for them, a board of directors. For example, Mark Zuckerberg founded Facebook with a few friends and nowadays Facebook has thousands of shareholders and their board runs the company with Mark as CEO.1

This separation causes problems to arise within the organization that need solving: agency

problems. Agency problems are the result of the situation where the agent, the person who is

responsible for managing and governing the organization, is acting in his own interest which is not in line with the interest of the principal, who gave the agent the authority to make decisions on his behalf (Jensen and Meckling, 1976). Managers make short-term investment decisions (Dechow and Sloan, 1991), while shareholders are more interested in the long-term opportunities. Manager’s risk appetite could be different from the risk appetite of the shareholders (Masulis 1988, p.48; Eisenhardt, 1989), and managers are more likely to consume perks when it does not affect their wealth (Jensen and Meckling, 1976).

This misalignment of interests can be reduced by managing the organization with a MCS. This system holds different kinds of controls to ensure that the interests of the employees will be in line with the interests of the organization. Over the years different kinds of frameworks for MCSs are developed, of which a short overview will be given.

(9)

2.1.1 MCS frameworks

As mentioned, the MCS solves agency problems. However, no one type of MSC exists that suits all organizations best. This thought is in line with the contingency theory: no one size fits all. In accordance with this theory organizations with different kinds of sizes, strategies, structures, technology, environments, etc. should use different kinds of MCS, because the better the match, the better the MCS can function (Otley, 1980; Chenhall, 2003). Thus, this section gives an overview of the kinds of MCS frameworks that exist in the literature.

First of all, Ouchi (1979) described three different mechanisms that organizations can use to cope with the agency problem: market control, bureaucratic control and clan control. Market control is a form of control that tackles the agency problem by measuring and rewarding employees to get them in line with the organization. Bureaucratic control means that employees are closely evaluated and that certain rules are designed flowing from these evaluations, whereas clan control is more about a socialization process that eliminates differences in objectives between employees (Ouchi, 1979).

Second of all, Simons (1994) developed a framework called ‘the four levers of control’ that can be linked to Ouchi’s prior established forms of control. This framework contains a beliefs system, a boundary system, a diagnostic system, and he added a fourth lever: an interactive control system. The beliefs system comprises the communication of values, purpose and direction by management, which strongly relates to clan control. The boundary system gives certain boundaries and rules for employees in order to avoid them from acting in a non-desirable way, which relates to bureaucratic control. The diagnostic control system can be linked to market control, as it reports information about the results of employees with the purpose for management to align behavior. Finally, the interactive system concerns how the other three systems can be communicated: by active and frequent dialogue (Simons, 1994).

Another framework that strongly relates to Ouchi’s and Simons’ frameworks is developed by Merchant (1982). The research in this thesis is built around his social control, action control and results control, which will be discussed more extensively in the next paragraph.

Table 1: Management control types

Source Types of management control

Ouchi (1979) Clan control Bureaucratic _control Market control

Simons (1994) Beliefs system Boundary system Diagnostic control _system Interactive control _system Merchant & van der

(10)

2.1.2 Merchant and van der Stede

The aforementioned kinds of control by Merchant (1982) were adjusted later as social control was split up into personnel and cultural control. From that moment on four kinds of controls are distinguished: action control, results control, personnel control, and culture control (Merchant and van der Stede, 2007).

First of all, action controls make sure that the employees perform desired actions and do not perform undesired actions. This means that some behavior is mandatory and other behaviors are prohibited. Action controls can thus only be used when the behavior desired is known and the control can stimulate the desired behavior or restrict the undesired behavior. For this reason action control is often referred to as behavioral control. Examples for action control would be passwords for access, procedures, rules and the monitoring of employees (Merchant and van der Stede, 2017). Second, results control is a form of control where the performance of the employees is measured and rewarded for accordingly. A good performance is compensated with a bonus and a bad performance is punished. Controlling employees’ behavior by measuring their results and rewarding accordingly seems like a very powerful tool to influence the behavior of the employees. Especially, when it is not known which behavior of the employees leads to the desired results, results control can offer the solution. Because results control creates consequences for the employees, they are incentivized to adjust their behavior so they will get the desired results. Examples of results control are setting performance targets, measuring performance and providing rewards (Merchant and van der Stede, 2017).

Third, personnel controls can put employees in the organization who are more likely to have the same interests as the organization. In that way, the risk of misalignment is reduced. With these controls the employees will better understand the objectives of the organization. It ensures that they have the capabilities and resources needed and a certain loyalty and intrinsic motivation is created towards the organization. Personnel controls concern the selection, placement and training of employees (Merchant and van der Stede, 2017).

Last but not least, culture controls are “designed to encourage mutual monitoring; a powerful form of group pressure on individuals who deviate from group norms and values” (Merchant and van der Stede, 2017, p. 97). The employees keep each other in line, because they feel pressure from their colleagues to act in the interest of the organization. Culture control is more difficult to implement, but certain team-building, social events and whistle blow opportunities can help shape this kind of control (Merchant and van der Stede, 2017).

(11)

2.2 Control Tightness

The existing literature is defining control tightness in differently manners. According to Merchant and van der Stede (2017) a MCS can be designed with a certain degree of tightness, which provides assurance about the alignment of interests within the organization. This implies that the tighter the controls in the MCS, the more benefits can flow from the use of the MCS. Auzair and Langfield-Smith (2005) agree with this partly, as they consider this to be true for certain kinds of tasks. They make a distinction between easy defined tasks and tasks that are not easy defined, where the former benefits from tight use of controls and the latter benefits from a loose use of controls, as more creativity may be needed for these kinds of tasks (Auzair and Langfield-Smith, 2005).

Butler et al. (1998) discuss clear and fuzzy structures that could bring two types of organizational error: errors of tightness and errors of looseness. When a structure is clear, it is based upon precise rules and it is clear what kind of decisions one can make, and how these decisions need to be taken. An error of tightness takes place when this structure is too tight: decision making is too restricted by hierarchical procedures and bureaucracy. An error of looseness takes place when the structure is too fuzzy. This structure holds flexible rules, which do not give the right guidance for decision making, which results in inefficient use of the structure (Butler et al., 1998).

Kober, Ng and Paul (2007) considered control tightness as a component that helps understand the design of the MCS in use. When defining control tightness they concentrate on the level of monitoring. The monitoring is extended to the obedience of rules, policies and plans and all other activities that the MCS comprises. This implies that a high level of monitoring in the MCS is associated with tight control. Auzair and Langfield-Smith (2005) also focus on the level of monitoring when defining tight control, namely the frequency. They assign tight control to a more bureaucratic MCS, in which actions and targets of employees are monitored in a more detailed and timely manner (Auzair and Langfield-Smith, 2005).

Management control tightness does not have a single definition yet, so for this thesis the following assumption is made about this construct: a MCS is tight when the system is firmly and extensively defining the controls in the system, only a small amount of deviation from the system is allowed, and a large amount of incentives is tied to the MCS.

2.2.1 Action control tightness

Regarding action control, tightness implies that the MCS contains a large number of rules and procedures and a high frequency of monitoring, which additionally are also used extensively to

(12)

influence behavior. This can also be described as a more bureaucratic, formal use of the controls (Snell, 1992). Furthermore, only little deviation from the rules and procedures is allowed when tight action controls are used in the MCS.

2.2.2 Results control tightness

Regarding results control, tightness implies that the MCS contains a large number of targets, budgets, measures to determine performance, and a high reward for achieving the desired results. Van der Stede (2001) examined tight budget control and found that the tighter the budget control, the lower the tolerance for deviation from the budget. Therefore, when tight results control is used in the MCS only little deviation is allowed from these controls.

2.2.3 Personnel control tightness

Regarding personnel control, tightness implies that the MCS contains an intensive selection procedure, that requires high qualifications and abilities in order for a person to get hired. The training of employees is also done more extensively and more frequently. Deviation from this selection and training procedure is not allowed when tight personnel controls are used in an MCS.

2.2.4 Cultural control tightness

Regarding cultural control, tightness implies that the socialization procedure is intensive: frequent team-building and social events that extensively aim at aligning the norms and values of the employees with the organization are part of the MCS. Triandis (1989) describes heterogeneous and homogeneous culture, where “homogeneous cultures are often rigid in requiring that ingroup members behave according to the ingroup norms”(Triandis, 1989, p. 511). Therefore, for this type of control, only little deviation from the norms and values of the organization is allowed as well. 2.3 Professional Service Firms

This section will illuminate what a PSF is. In order to understand what these sort of firms are, first some background information will be given about the difference between products and services, as services are the main output of these firms. However, this in this thesis professional services are examined as opposed to regular services, so the subsequent paragraphs will describe the characteristics of service professionals and give an overview of what PSFs are according to the existing literature. The distinctive characteristics could give an explanation as to why providing professional services leads to a different approach concerning MCS design.

(13)

2.3.1 Services

In the literature a distinction is made between products and services upon the characteristics of intangibility, inseparability, heterogeneity and perishability (Zeithaml et al., 1985; Wolak et al., 1998). Although this is not specific to professional services, these characteristics will briefly be reviewed as a first step for understanding professional services as well.

Intangibility refers to the fact that services do not have a physical appearance and cannot be touched, and therefore are not tangible. Regan (1963) created this concept, which indicates that services are evaluated differently than products. For instance, a campaign delivered by an advertising agency is not something that is valued in a quantitative manner. The quality is based upon the amount of desired attention the advertisement receives, which is also an intangible matter, whereas the posters that are part of the advertisement can be measured quantitatively and the quality depends on physical appearance. This intangibility characteristic makes services harder to measure and therefore it is harder to assign value to a service as opposed to a product (Regan, 1963; Langaerd et al., 1981).

Inseparability concerns the intertwinement of the delivering and the consumption of the service. This means that consumers are part of the quality of the service and are able to influence the quality of the service (Berry, 1975; Sasser, 1976). For example, a firms gets new software installed: they can be content with the software itself, but if the delivery of the software is done by an IT professional who communicates differently from the firm, the entire service can be perceived of less quality. Therefore the inseparability also adds to the difficulty in measuring services, because another firm could communicate in a similar manner as the IT professional and thus assign a higher value to the quality of the service.

Heterogeneity means that the variability in services is very high. The same kind of service can be different per client, per professional and per occasion (Bell, 1981; Rathmell, 1966). A consult is tailored to the client, the consult differs from time to time, and moreover the same kind of consult can be executed differently per consultant. This characteristic also provides a problem for measuring the value of a service, because services highly involve people, who’s behavior differs per person and also per day.

Perishability refers to the fact that services cannot be stored (Thomas, 1978; Knisely, 1979). For example, when an auditor delivers an audit report this cannot be saved for the next year to sell, because the audit is based on an annual report that varies per year. Obviously this also implies that products are easier to measure than services.

(14)

To conclude, these different characteristics for products versus services will attribute to the different design of a MCS (Suddaby et al., 2008), as the measuring of the quality of services is more difficult as opposed to measuring the quality of a product.

2.3.2 Service professionals

Services can be provided by a hotel, an airline, a teacher or a gardener, however this thesis focuses on organizations that provide professional services. Accountants, architects, lawyers, doctors, engineers and consultants are examples of providers of these professional services. Service professionals can be distinguished from regular services providers in several ways, as the following outline of existing literature will illustrate.

First, service professionals usually underwent a high level of education which allows them to be part of a professional society. Thus, they are required to meet certain special conditions to be a member of this professional group and to be able to perform their occupation (Shapero, 1985; Larson, 1990). Because of these specifications obtained through their education, they often have an expertise that is difficult for other people to understand, that “clients believe is beyond their own competence”(Greenwood et al., 2005, p.663). For this reason it is harder for clients and managers to evaluate professionals themselves as the work of professionals can only be evaluated by other professionals (Mills et al., 1983; Patterson, 2000; Greenwood et al., 2005).

Second, service professionals perform tasks that are complex and they therefore use their specific skills, expertise and knowledge to accomplish these (Von Nordenflycht, 2010). Because of this, service professionals are treated differently from other employees. This means they are also expected to be treated differently from managers (Barley & Tolbert, 1991). Managers delegate work to others, whilst professionals have the ability to identify and solve problems themselves: they have a certain autonomy and thus are more independent in performing their tasks (Goodale, Kuratko, and Hornsby, 2008; Derber & Schwartz, 1991).

Third, service professionals tend to demand this independence and autonomy in their work, because of their socialization process as a member of their professional society (Raelin, 1989). Managers will use the MCS to try to align the values of the organization with its employees, whereas service professionals will want to stick with their own professional values (Abernethy & Stoelwinder, 1995). Raelin (1985) suggests that the favor for autonomy is a characteristic of service professionals that creates this conflict with management control. Specifically, service professionals tend to resist bureaucratic forms of control, as they prefer freedom to operate based on their expertise (Von Nordenflycht, 2010). Another conflict arises from the fact that managers will want

(15)

to rely less on individuals and their knowledge, whereas service professionals will want to shield their individual knowledge as it is their main asset (Løwendahl, Revang, & Fosstenløkken, 2001). To sum up, service professionals are highly skilled and educated people, they perform complex tasks that clients and managers cannot understand, are treated differently than other employees in the organization and their characteristics conflict with management control. These factors also contribute to a different design of MCS, as these service professionals respond differently to management control as other employees.

2.3.3 Professional Service Firms

Service professionals can be employed by an organization, which are called professional service firms. In the existing literature different definitions have been used to describe what an PSF is. As Von Nordenflycht (2010) explains, it is important to determine the characteristics of an PSF in order for future research to examine these specific organizations. He suggests that an PSF can be distinguished from other organizations by the following characteristics: 1) knowledge intensity, 2) low capital intensity and 3) a professionalized workforce.

1) Knowledge intensity is the degree of reliance on complex knowledge of individuals in the organization, which is necessary to produce the output of the organization. In other words, the PSF highly relies on its service professionals, who’s knowledge is the backbone of the organization.

2) Low capital intensity is the degree of reliance on non-human assets. This means that the PSF does not depend on a large amount of inventory or equipment.

3) A professionalized workforce indicates that the profession of the service professionals in the organization stipulates a certain ideology and is self-regulated. Therefore the service professionals in the PSF adhere to several norms or ethical codes and their occupation is often controlled by an professional association that determines who is certified for membership of the professional association (Von Nordenflycht, 2010).

Greenwood et al. (2005) use somewhat the same kind of definition, as they suggest the following two characteristics for PSFs. First, they say that the output of PSFs “are intangible applications of complex knowledge”(Greenwood et al.. 2005, p. 663). Second, they state that PSFs “employ a highly educated (professional) workforce to customize complex knowledge to client situations”. (Greenwood et al., 2005, p. 663).

Another take on the definition of a PSF comes from Auzair and Langfield-Smith (2005). They applied the six dimensions of Silvestro et al. (1992) to describe PSFs as “organizations with

(16)

relatively few transactions, highly customized, process oriented, with relatively long contact time, with most value added in the front-ofﬁce and where considerable judgment is applied in meeting customer needs.” (Auzair & Langfield-Smith 2005, p. 402). This definition seems to be more specifically addressing the way a professional service is delivered by a PSF. However, this thesis focuses on the broader definition laid out by Von Nordenflycht (2010). For this reason the following examples of PSFs are given in Table 2 below.

Table 2: PSF examples

2.4 Organizational Performance

In the existing literature, organizational performance (OP) is often referred to as organizational effectiveness or efficiency, which have been said to have the same meaning (Venkatraman and Ramanujam, 1987; Hancott, 2005). However, the definition of OP is somewhat ambiguous, as will be clear from the following overview of literature.

First of all, Ho (2008) states that “OP is an indicator which measures how well an enterprise achieves their objectives”(Ho, 2008, p. 1238). She implies that the definition of organizational performance is including the fact that this construct is a ‘measure’ in the definition itself. Andersen (2006) also beliefs that performance can be seen as the degree of goal attainment. He focusses on the profitability objectives of an organization when assessing the OP, whilst others use additional goals to determine the effectiveness of an organization (Delaney and Huselid, 1996; Steers, 1975; Ho, 2008).

According to Hoque and James (2000), a ‘Balanced Scorecard’ measures the performance of an organization from four perspectives: the financial, customer, internal-business-processes and learning and growth perspective. The financial perspective includes the measuring of cash flows, operating income and sales growth, whereas in the customer perspective market share and customer satisfaction are the main focus of measurement. For the internal-business-processes, the

Accounting Actuarial services

Advertising Architecture

Biotechnology Consulting IT

Consulting Engineering Consulting Management/Strategic

Consulting HR Engineering

Consulting Technology Financial advising

Fashion design Insurance brokerage

Graphic design Investment banking

Investment management (hedge funds, VC, mutual funds) Law

Marketing/public relations Media production (film, TV, music)

Medicine/Physician practices Pharmaceutical

Project management Real estate

Recruiting - executive Research/R&D

Risk management services Software development

(17)

effectivity is measured as the quality, development and design of the output. The learning and growth perspective measures how employees, systems and procedures can adapt when changes occur, and labels this as the performance of the organization (Hoque and James, 2000).

As the previously mentioned literature is explaining organizational performance to be a combined bundle of measurements, the organizational performance can be expressed as a certain value. To address the problem of assessing what a high or a low value for organizational performance is, the effectiveness of the organization can be compared to other organizations. King and Clarkson (2015), for instance, measured performance as “financial performance relative to peers” (King and Clarkson, 2015, p.24).

In this thesis organizational performance of a PSF is considered to be the perceived effectiveness of the organization by its stakeholders, such as clients and service professionals, in terms of profitability, competition, innovation, growth rate, market share and size, in comparison to other organizations.

2.5 Organizational Reputation

Organizational reputation is also a construct that can be defined in different ways, as can be found in the existing literature (Barnett et al., 2006). When examining the literature about organizational reputation Barnett et al. (2006) came to the conclusion that reputation can be seen as an asset, an assessment and a state of awareness. First, the state of awareness means that the organization is observed by outsiders, who have a certain perception of the organization. Second, reputation as an assessment means the observant stakeholders of the organization act upon their perception of the organization: they will judge the status of the organization. Third, the reputation can be seen as an asset, as an intangible that adds a certain value to the organization (Barnett. et al., 2006).

Reputation can thus be seen in three different ways. But what are the characteristics of a reputation? Argenti & Druckenmiller (2004) have formulated reputation to be “the collective representation of multiple constituencies’ images of a company, built up over time and based on a company’s identity programs, its performance and how constituencies have perceived its behavior” (Argenti & Druckenmiller, 2004, p. 369). Rindova et al. (2005) also focusses the fact that reputation is generated by the performance of the organization, namely on the perceived quality dimension: the degree of quality to which an organization is perceived to produce their output. They also add another dimension that concerns the degree to which an organization is perceived to get a large amount of recognition in its industry: the prominence dimension (Rindova et al., 2005). These

(18)

authors state a definition of reputation with an instigator for reputation as part of the definition: they explain it to be an effect of certain aspects.

However, as Fombrun (2012) discusses, the use of antecedents and consequences in the definition itself could limit the possibility to do research about organizational reputation. Therefore, he proposes a definition that more plainly states what reputations are: “collective assessments of a company’s attractiveness to a defined set of stakeholders relative to a reference group of other companies” (Fombrun, 2012, p. 100). Thus, he focusses on the competitive aspect of reputation.

In line with the previously mentioned authors, Eccles et al. (2007) recognize organizational reputation to be a perception. They state that a variety of stakeholders can assess the reputation of an organization, and that the function of these perceived reputations combined is the actual organizational reputation. Furthermore, they argue that this reputation can be different from the genuine character of the organization, which generates a gap. This gap can in fact create a certain risk for organizations to lose the incremental perceived ‘better’ character of the organization (Eccles et al., 2007).

As mentioned before, in the literature section about PSFs, it is hard for clients to assess the work of a service professional, because their tasks require complex knowledge. According to Kaiser and Ringlstetter (2010) this leads to a certain insecurity amongst the clients, because the quality of the service of the PSF will only be recognized after the service is completed, not before the PSF is hired. To reduce this insecurity when selecting a PSF, clients take the reputation of the organization into account because it “reflects the company’s knowledge and relational competence” (Kaiser and Ringlstetter, 2010, p.5). The reputation can also be used to communicate quality to attract and retain service professionals, which is important for a PSF, since the service professionals are its critical resources (Kaiser and Ringlstetter, 2010).

In this thesis, the construct of organizational reputation of a PSF is considered to be an assessment of the organization’s representation, in comparison to other PSFs, which is perceived by stakeholders such as clients and service professionals. Thus, for this thesis organizational reputation is best explained by the definition of Fombrun (2012).

(19)

3 Theory & hypotheses development

This chapter concerns the theory that is used to describe the predicted relationships in the hypotheses. First, tournament theory will be discussed and the four associated hypotheses (H1a-d) will be given. Second, contingency theory for reputation is explained, which leads to the additional four hypotheses (H2a-d). Figure 3 at the end of this chapter will give a visual representation of the eight hypotheses in this thesis.

3.1 Tournament theory

Lazear and Rosen (1981) examined the competition between employees for job promotions to prove tournament theory. According to this theory, the performance of an employee is assessed when deciding on promoting the employee and this performance increases when the level of job promotion is higher, because the higher job position is associated with higher compensation. Therefore, the effort to reach a higher position is associated with the magnitude of the prize of winning the tournament (Lazear and Rosen, 1981). Moreover, Rosen (1986) stated that this predicted relationship is expected to be a convex relationship, which Baik et al. (2016) confirmed to be true with their research.

Following this line, PSFs are competing for a better market position (the prize): they are striving towards a higher market share, more profit, more innovation and they want to be growing faster and become bigger in size than other organizations. Because the market position is determined by the level of performance of the PSF (the position), the PSF will increase its effort in order to increase its performance. The ‘effort’ can be seen as the degree of control tightness that management can impose upon its service professionals. This is assumed because service professionals are the main asset of a PSF to generate its output (Von Nordenflycht, 2010), and therefore controlling them is crucial for management to influence the performance of the organization.

When a PSF is improving its performance with the same amount each time it climbs the performance ladder, the PSF will tighten management controls more and more to improve its performance. This is assumed, because the better their performance in comparison to other organizations in their field, the less room for improvement in comparison to these other organizations. As top market positions become rarer and rarer at an increasing rate, the extra effort in the form of management control tightness will increase at an increasing rate, as (Baik et al., 2016) established for tournament theory. Therefore the relationship between organizational performance and management control tightness is expected to be exponential, as Figure 1 on the next page will demonstrate.

(20)

Figure 1: Predicted exponential effect on management control tightness (H1a-d)

For the same amount of increase in organizational performance in a PSF (the position), the market position (the prize) will increase at an increasing rate. Because the prize is increasing at an increasing rate, management control tightness (the effort) will increase at an increasing rate.

Regarding action control tightness, effort can be seen as the installation of more rules and more procedures to get the employees more in line to reach a higher market position. Although it is known from the existing literature that service professionals have a certain aversion towards bureaucratic controls (Von Nordenflycht, 2010), management might be

unaware of this and ty them to more rules and procedures, instead of giving them more freedom in performing their tasks. This is assumed because management cannot fully understand the complex tasks that service professionals perform (Mills et al., 1983; Patterson, 2000; Greenwood et al., 2005). The following hypothesis reflects this:

H1a: Organizational performance and action control tightness are exponentially positively related in PSFs

With respect to results control tightness, effort can be seen as information. Organizations use results control usually when they do not know which behavior leads to the desired better performance. So the tighter the results control, the more information an organization can gather, the better the organization can apply this information to get a better performance. Therefore it is expected that the higher the performance of an PSF, the tighter the results control. . This is also in line with the existing literature that predicts firms with a strong market position to use

(21)

controls in the form of budgets more prominently than firms with a weaker market position (Galbraith, 1977). Hence, the following hypothesis:

H1b: Organizational performance and results control tightness are exponentially positively related in PSFs

For personnel control tightness, effort can be seen as stricter application procedures and mandatory training to ensure that the best capable employees are in line with the organization. Thus, management will tighten the personnel controls as predicted by tournament theory, which results in the hypothesis:

H1c: Organizational performance and personnel control tightness are exponentially positively related in PSFs

With respect to cultural control tightness, an organization can aim to change the culture by enhancing the interaction between colleagues with teambuilding assignments and social events. This kind of effort can be used to improve the performance and thus the market position of the organization. In line with tournament theory the following hypothesis is formulated:

H1d: Organizational performance and cultural control tightness are exponentially positively related in PSFs

3.2 Contingency theory: reputation

PSFs differ from other firms such as manufacturing firms, thus should also differ in designing their MCS (Suddaby et al., 2008). According to this contingency theory, no one MCS is the best fit for all organizations (Chenhall, 2003). Reputation is especially important for PSFs (Kaiser and Ringlstetter, 2010), which is why it is expected to be part of the decision-making process about the design of the MCS. Therefore, the MCS design could be different for PSFs with a high reputation from PSFs with a low reputation. Specifically, reputation could have a moderating effect on the previously mentioned relationship between organizational performance and management control tightness for several reasons, which will be discussed in this section.

First, organizational performance and organizational reputation have been associated with one another in the previous literature. It has been examined in the past that a high reputation leads to higher profitability (Roberts and Dowling, 2002). The other way around, the literature also found another connection between reputation and performance: an unexpected loss of reputation can affect the performance of an organization (Mitroff, 1988; Shrivastava and Mitroff, 1987). Because of this connection between performance and reputation management is assumed to be more eager to safeguard the performance of the PSF when its reputation is high. In order to

(22)

accomplish this safeguarding of the performance they will tighten the MCS, which was explained to be the most influential tool for management in the previous section about tournament theory. This is supported by the fact that high reputation organizations are more ‘in the picture’ than low reputation organizations. For example, the data scandal at Facebook in March 2018 led to a 5% drop in stock prizes for Facebook.2_{Organizations with a lower reputation will probably}

be less scorned in the media for the same mistake, since the attention will be more on the organizations with a higher reputation. When examining individual reputation, Yaniv and Kleinberger (2000) found evidence that implied that it is easier to lose a reputation than to gain one. Following this line, organizational reputation is likely to have this characteristic as well. The gap between reputation and the actual character of the organization produces a certain kind of risk that is tied to the reputation of a PSF (Eccles et al., 2007). This gap is assumed to be bigger for high reputation firms: they have more to lose than low reputation firms. High reputation PSFs could be damaged more from a mistake than low reputation PSFs, as their reputation is already more close to their character (Eccles et al., 2007). These mistakes can be attributed to the performance of an organization, as the organization was not effectively achieving its objectives (Ho, 2008). PSFs with a high reputation are thus feeling more pressure not to make mistakes and want to incentivize their service professionals more to behave in the best interest of the organization. In other words, PSFs with a high reputation place more focus on their performance and consequently use a tighter focus on controls.

Third, PSFs are expected to include the level of reputation in their decision to tighten the management controls when the performance is improving, because of anticipation. Wilson (1985) stated that "the essential requirement for a player's reputation to matter for his current choice of action is his anticipation that his later decisions will be conditioned by his later reputation”(Wilson, 1985, p.27). This anticipative feature connects reputation and performance even more, because the applied tournament theory is also expecting this anticipation for performance when management is deciding on controlling its service professionals. For this reason reputation is also expected to be included in the decision-making process about management control tightness when the organizational performance of the PSF is improving.

This theory emphasizes organizational reputation to be a contingency factor, so the predicted relationship by tournament theory is different for PSFs with different levels of

(23)

reputation. The following hypotheses show this moderating effect of reputation on the relationship between performance and control tightness:

H2a: When organizational reputation is high, organizational performance and results control tightness are more

positively related than when reputation is low in PSFs

H2b: When organizational reputation is high, organizational performance and action control tightness are more

positively related than when reputation is low in PSFs

H2c: When organizational reputation is high, organizational performance and personnel control tightness are

more positively related than when reputation is low in PSFs

H2d: When organizational reputation is high, organizational performance and cultural control tightness are

more positively related than when reputation is low in PSFs

Figure 2: Predicted moderating effect of organizational reputation (H2a-d)

This figure shows that the predicted positive effect of organizational performance on management control tightness is more positive for PSFs with a high reputation than for PSFs with a low reputation.

A visual presentation of the eight hypotheses is given in Figure 3 on the next page. This includes size and environmental uncertainty as the predicted relationships will be controlled for by these two constructs. This is explained more extensively in the next chapter.

(24)

(25)

4 Research Method

The research method used to examine the hypotheses in this thesis is described in this chapter. First, the PSF Thesis project that was joined is outlined in the first section, as are the reasons why and how I joined this project. The second section gives an overview of the pre-tests that have been conducted to establish the survey of this project. In the third section, the process of the sample selection and the characteristics of the respondents in the final dataset are described. In the fourth section the variables used in this thesis are given. After this, validity and reliability tests were computed of which an overview is given, and the last section shows the statistical model.

4.1 PSF Thesis Project

To acquire data to test the hypotheses for this thesis I joined the PSF Thesis Project 2017-2018. This project is supervised by Helena Kloosterman, a PhD student at the University of Amsterdam, who is examining (for the Accounting section) how management accounting and control systems are designed in PSFs and what the effects of these systems are on its service professionals. The project started in 2016, and since then students from the Master Accountancy and Control were able to join this project. As the condition for joining the project was to find ten respondents who completed the survey that was used to generate the data for this project, the project gives the opportunity the examine a large sample of respondents. On my own I would not have gotten a sample that contains this many respondents, which is the main advantage for joining the project. Another advantage for joining the project is that the survey concerns a lot of different variables, for which the data is not available publicly, so students can construct research questions that have not been examined before in the existing literature, and actually answer them.

The survey, that was designed by Helena Kloosterman, contains eighteen constructs respondents have to answer one or more questions about, which adds up to a total of 136 questions in the whole survey. The constructs concern PSF characteristics, MCS variables, outcome variables and control variables. For this thesis I used six of the constructs, namely action control, results control, personnel control, cultural control, organizational performance and organizational reputation. The questions for the constructs are based upon existing literature, or questions were developed because some constructs are relatively new and have not been examined before.

The condition for joining the project was to send the link to the survey to at least ten respondents who then had to complete the survey online before the 1st_{of February 2018. The}

respondents had to be working in a PSF. Respondents with certain occupations counted as 2 respondents, as these were underrepresented in the sample. Two of my friends, who are engineers,

(26)

completed the survey for me. Furthermore, I used my network to find a nurse and two doctors to fill in the survey. As their occupations counted as two, my total number of respondents was already ten. Additionally, I found a notary who completed the survey. This led to a total of eleven respondents, which is how I met the condition for joining the project. After completing the survey online, the respondents had to write me an e-mail noting they completed the survey, in order for Helena to establish that I did not use a robot to complete the surveys, which safeguarded the quality of the data.

4.2 Survey pre-testing

In order to examine the quality of the questions in the survey of the PSF Thesis Project, two pre-tests were conducted by Helena Kloosterman. This was done to determine whether the questions related to the different constructs would generate outcomes that enable effectively examining these constructs.

The first pre-test was a sorting task about the MCS constructs, that fourteen service professionals (2 Management Consultants, 1 IT Consultant, 1 Security Consultant, 1 Accountant, 1 Psychologist, 1 Dentist, 1 Architect, 1 Marketing professional, 4 Lawyers, 1 Graphic Designer) completed. Each of them was handed a paper with eight types of management control and a paper with 52 associated questions, that they had to match to the control type which they felt the question related to the most. For each of the eight control types, the four questions that were least mismatched were selected to be included into the survey, which is why the total number of questions to test the MCS constructs is 32.

The second pre-test was conducted to test the quality of the whole survey, by asking 20 additional service professionals to look at the survey online and answer questions about its content, its appearance, the time it took to complete the survey and its clarity. The 20 service professionals (3 Medical doctors, 1 Dentist, 3 Chemists, 1 Advertising executive, 3 Lawyers, 2 Management Consultants, 1 Architect, 1 Graphic Designer, 1 Artist, 2 Marketing professionals, 1 Electrical engineer, 1 Mining engineer) provided written answers and answers by phone, which only led to minor changes to the questions in the definitive version of the survey.

4.3 Sample

The survey was completed by service professionals who had to meet several requirements. Obviously the respondent needed to be working for at a PSF. The information about the project made clear what kinds of organizations would qualify to be a PSF, as the list of the main examples

(27)

was presented. (See Table 2 in the literature review section about PSFs.) Respondents could also fill in their occupation when no example was applicable, as not all existing PSFs were stated as an option to choose from. Next, the respondents needed to meet four other criteria:

1) Respondents need to have at least three years of experience in their professional service field and preferably at most ten years of experience in their professional field. This is because service professionals who have less than three years of experience may not fully understand or know how they are subjected to a MCS, so they may not be able to answer the questions concerning a MCS. Service professionals who have more than ten years of experience are more likely to have a position as discussed in the third criteria.

2) The respondents need to be working at an organization (or an organizational unit) that had at least 50 or more employees. It is assumed that smaller organizations may not have a MCS that controls it employees, as is examined in this thesis.

3) The respondents need to be subjected to the MCS rather than designing it, which is why owners, partners or board members do not meet this criteria.

4) The respondent must speak and understand English at a business level.

5) The respondent must not work at a non-profit organization, NGO or government organization, except for organizations in medicine/physician practices.

The sample that originated from the respondents who completed the survey of the PSF Thesis Project existed of 730 respondents. However, some of the respondents were excluded for this thesis for several reasons. First, 30 respondents did not read the ‘note on terminology’. I excluded these respondents because I consider understanding the terminology an important part for filling out the survey. Second, a lot of respondents did not answer all the questions. As not all 136 questions are relevant to this thesis I only excluded those respondents that did not answer the relevant questions. The relevant questions concern the different types of management control (action, results, personnel and cultural control), organizational performance and reputation. A total of 97 respondents were omitted because of this.

This leaves 603 respondents. However, 84 of them did not meet the requirements for the survey. Some of the respondents did not have enough experience (less than 3 years in the field), so I considered them not to have the required knowledge to answer the relevant questions. Because of this reason, 47 were excluded from the sample. After this, 2 respondents were omitted, because they did not fill in any occupation or formal job title, so there is no indication whether these are service professionals. Moreover, 2 were deleted because they did not say how many employees were working at their organization or organizational unit, which means that one cannot tell

(28)

whether they are subjected to an MCS. Additionally, 16 were removed because the occupation was not professional or not in an PSF (for example: a singer, people in education, self-employed people, people working in a government organization). Last but not least, 17 people who are probably (partly) designing the MCS (executives, directors, CFO, managing partners) were omitted. Therefore, the final dataset sample consists of 519 respondents.

Table 3: Sample selection

Original amount of respondents 730

Respondents who did not read the note on terminology 30

Respondents who did not answer all relevant questions 97

Respondents who did not meet the requirements of the survey 84

Not enough experience (<3 years) 47

No occupation or formal job title 2

No number of employees 2

None-PSF 16

MCS designing positions 17

Total respondents in the final dataset 519

The 519 respondents in the final dataset sample have 34 different nationalities. Most of the respondents have a Dutch nationality (73,4%). Subsequently, 5,8% of the respondents are German (N=518). An overview of nationalities can be found in Appendix H. In the sample, 66% of the respondents is male and 34% is female (N=518). Of the respondents, 40.3% obtained a Bachelor degree or lower, 45,9% obtained a Master degree, and 13,8% obtained a PhD or other professional doctorate degree (i.e. J.D./M.D./D.D.S) (N=516). The youngest respondent is 21 years old and the oldest respondent is 64 years old, with an average respondent age of 35 years (N=514). Regarding the experience of the respondents, the average experience in the field is 7 years (N=519) and the average experience in the organization is 6 years (N=518). The occupation with the highest percentage of respondents(14,8%) is Accounting (see Appendix I).

Table 4: Descriptive statistics of respondents

Percentage N Minimum Maximum Mean Std. deviation

Age 514 21 64 35,42 8,878 Gender* 518 1 2 1,34 0,424 Male 66% Female 34% Education** 516 1 3 1,73 0,686 Bachelor degree 40,3% Master degree 45,9% PhD or other degree 13,8% Experience in field*** 519 3 11 7,44 2,906 Experience in organization*** 518 1 11 6,08 3,172 * Male = 1, Female = 2

** Bachelor degree or lower = 1, Master degree = 2, PhD or other professional degree = 3 *** 11 years = 10 years or more

(29)

4.4 Variables

The constructs in the survey that are relevant for this thesis were used to create the variables for statistical testing. This thesis examines the relationship between organizational performance (independent variable) and the degree of management control tightness for each of the four types of management control (dependent variables). Additionally, it is examined what the effect of organizational reputation (moderator variable) is on the previous relationship. Throughout the examination of these variables, other factors that may influence the relationships between these variables will be controlled for: size and environmental uncertainty (control variables). The unit of analysis is the organization the respondents are employed by, as the informants, the service professionals, are asked to answer the questions about the total organization or the unit in the organization they are working for. The questions in the survey were created by Helena Kloosterman, either by adapting from existing literature or by creating new items.

4.4.1 Independent variable: organizational performance

The independent variable in this thesis is organizational performance (PER). The measurement of this variable was adapted from King and Clarkson (2015), and was called ‘Unit performance’ in the survey. Respondents were asked to score their PSF in comparison to other organizations in their field, on the basis of profitability, competition, innovation, growth rate, market share and size. They needed to score these six items of PER on a 5-point Likert scale, from strongly disagree (1) to strongly agree (5). The Likert scale can be made quantitative: a score of 1 represents a low organizational performance and a score of 5 represents a high organizational performance. The results of the six items were then added up and divided by six, in order to determine the average value of PER. The questions in the survey regarding PER can be found in Appendix E.

4.4.2 Dependent variables: management control tightness

The dependent variables are the four management control tightness types: action, results, personnel and cultural control tightness (ACT, RCT, PCT and CCT). In the survey, eight items are related to each of those four constructs, as already explained in the section about the pre-tests. Each of the 32 items was a statement that had to be rated on a 5-point Likert scale, from strongly disagree (1) to strongly agree (5). The higher the result on the statement, the tighter the use of the concerning type of management control. For each of the four variables, the average of the results of the eight items was calculated, which was used to execute the statistical models. The exact questions that the respondents had to answer about the management control tightness variables are displayed in Appendix A, B, C and D.

(30)

The variable of ACT is labelled ‘behavioral control’ in the survey, which has the same meaning as action control. The measurement of this variable was adapted from Aiken and Hage (1968), Van den Ven and Ferry (1980), Cunningham and Rivera (2001), Van der Stede (2001) and Bodewes (2000). The eight statements for this variable concerned the amount of rules, procedures, the amount of monitoring and the amount of deviation that is allowed from these rules and procedures in the organization by which the respondent is employed. Statements Q4_10, Q4_11, Q4_12 and Q4_13 were asked negatively, so the results for these statements were reverse coded. Regarding RCT, eight questions were asked about the number of performance measures the respondents experienced in their organization and the allowed deviation from these measures. This measurement was adapted from existing literature (Jaworski et al., 1993; Aiken and Hage, 1968; Simons, 1987; Van der Stede, 2001). Helena Kloosterman also added new items, specifically the items related to the extent of results control were original. Statements Q5_9, Q5_10 and Q5_11 were asked negatively, thus the results that arrived from these statements were reverse coded.

The eight statements concerning PCT were about the hiring process in the organization of the respondents: how extensive and how much deviation from the process is allowed by the organization. The measurement of this variable created by Helena Kloosterman was entirely new. Statements Q3_7 and Q3_11 were asked negatively, so these were reverse coded.

The measurement of CCT was partially adapted from O’Reilly and Chatman (1986), but most items were created by Helena Kloosterman as original items. CCT was also examined with eight statements, that included the amount of social activities organized and the degree to which the respondent is in line with the norms, values and objectives of the organization. One statement (Q10_5) was asked negatively, and was therefore reverse coded.

4.4.3 Moderator variable: organizational reputation

The moderator variable is organizational reputation (REP). The measurement of this variable in the survey is adapted from Combs and Ketchen (1999), by giving the respondents four statements about organizational reputation. The respondents had to rate the statements on a 5-point Likert scale from ‘strongly disagree’ to ‘strongly agree’. As the Likert scale enables to compute quantitative results, the higher the number for REP, the higher the organizational reputation is perceived by its service professional(s). The results of the four statements are added up and divided by four, as an average to determine the score on REP for each respondent. The questions in the survey for the measurement of this variable can be found in Appendix F.

(31)

4.4.4 Control variables: size & environmental uncertainty

As mentioned before, other factors that are not of theoretical interest for the present study, may also influence the relationships that are predicted in the hypotheses. These factors are controlled for by adding them to the statistical model. These ‘control variables’ can show effects on the dependent variables, and the effect of the independent variable is therefore “cleaned” for the explanatory effects of the control variables.

The size of an organization could be important because the examined relationships could be significant for big organizations, but not for small. The control variable SIZE was examined by one question: “How many people are employed by your entire company?”, which was a new item in the survey of Helena Kloosterman. The answer was given a 4-point scale, with assigned values varying from less than a 100 to more than 5000.

The control variable environmental uncertainty (UNCERTAINTY) could be important to test for, because high performing firms could be using tighter controls only because of a highly uncertain environment. The measurement of this variable was adapted from Gordon and Narayanan (1984). The respondents were asked to rate three areas of intensity in their industry: bidding for contracts, competition for manpower and price competition. They also had to answer a question about the amount of new products/services brought out in their industry and two other questions about the predictability of market activities and client preferences. All items were measured on a 5-point Likert scale. The results of these six questions were taken as an average for the statistical testing. The questions can be found in Appendix G.

Table 5: Overview variables

Type of variable Variable name Abbreviation

Independent variable Organizational performance PER

Dependent variable Action control tightness ACT

Results control tightness RCT

Personnel control tightness PCT

Cultural control tightness CCT

Moderator variable Org. performance*Org. reputation PER*REP

Control variable Size SIZE

(32)

4.5 Validity & reliability analyses

In this section the validity and the reliability of the survey items will be tested. This leads to the deleting of several items of the variables to improve the validity and reliability of the remaining items. This is done to ensure the variables can effectively be used to test the hypotheses.

4.5.1 Explanatory factor analysis

To test the questions in the survey for validity I conducted an explanatory factor analysis. All the variables with more than 1 question (PER, ACT, RCT, PCT, CCT, REP and UNCERTAINTY) were part of this analysis. The explanatory factor analysis gives insight into which questions can be used to effectively examine the variable. All the items for all the previously mentioned variables are analyzed to see which items load on the same factor, or in other words, which questions load on which variable. For this analysis I used a fixed amount of 10 factors, because the control tightness variables are expected to be loading on 2 factors each, except for CCT. For ACT, RCT and PCT the items seem to concern 2 kinds of topics: questions about the extent of the control and questions about the amount of deviation that is allowed from the control. For CCT this distinction is less clear, so for the control tightness variables in total I expect loadings on 7 factors. Then I added 1 factor for the other 3 variables (PER, REP and UNCERTAINTY), which leads to the fixed amount of 10 factors.

First, to determine whether this explanatory factor analysis is adequate, a Kaiser-Meyer- Olkin test and a Bartlett’s test were executed. This gives a KMO of 0,789. According to Field (2013) the KMO lies between 0 and 1, and the closer the KMO gets to 1, the compacter the patterns of correlations. This means that the analysis can produce reliable factors and thus this factor analysis is appropriate (Field, 2013). A Bartlett’s test is done as well in this factor analysis, which gives a significance of 0,000. This implies that the correlations between the variables are significantly different from 0 (Field, 2013), and therefore the explanatory factor analysis is considered adequate.

Second, the eigenvalues (EV) of the factors are examined, as these give insight in the importance of each factor. All the EV’s of the 10 factors are greater than 1, so this in in line with Kaiser’s criterion (Field, 2013). As displayed in Table 6 and 7, the items have a certain loading onto the factors. Stevens (2002) has given a guideline for the significance of factor loadings, and therefore only loadings that are greater than 0,298 are presented in the table, because this is a proper amount of loading for the sample size in this thesis.

The effect of organizational reputation on the relationship between organizational performance and management control tightness