How to manage men and women, when men are from Mars and woman from Venus?

(1)

How to manage men and women, when men are from Mars and

woman from Venus?

Name: Noelle Papaloukas Student number: 10003301

Thesis supervisor: Helena Kloosterman Date: 25-06-2018

Word count: 17445

MSc Accountancy & Control, Control track

(2)

Statement of Originality

This document is written by student Noelle Papaloukas, 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

This paper started with questioning the use of women’s quota in European countries. The problems of getting women promoted in top positions might be allocated within the management control design. This paper investigates what the effect is of the management control design on the different experience of professional tension and performance by professional men and women. The framework developed by Hackman & Oldham (1981) is used to explore this possible relationship. Specifically, the effect of action controls and result controls on the different experience of professional tension between genders is explored. The different experience of professional tension due to the used controls is explained by, the influence on role conflict and role ambiguity and the typical characteristic of professional men and women. To test the hypotheses, I used data collected from a survey conducted among professionals working within professional service firms. The regression analysis found empirical evidence that professionals experience more professional tension with the use of result controls than with action controls. I did not find evidence supporting the assumption that professional women experience less professional tension with the use of action controls than professional men. Furthermore, the results did not support the prediction that the experience of professional tension decreases the performance of professional women more than professional men. This paper concludes with discussing possible other factors which might influence the promotion difficulties for women and the suggestions for future research.

Keywords: Professional Service Firms, Management Control System, Gender, Professional

(4)

1 Introduction

In 2016, Germany instated a law which required top 100 firms or publicly traded companies to have women hold 30% of the board seats. Germany is not the only country to instate a woman’s’ quota. Norway was the first European country to stipulate a woman’s quote of 40% for public limited companies, soon after, France, Italy and Spain followed. A quota-based selection policy is a form of proportional decision making that ensures that a certain number of jobs go to a particular segment of the population(i.e., 30% of supervisory seats going to women by 2017) (Shaughnessy, Braun, Hentschel, & Peus, 2016). Through quota-based selection policies, social groups that have been disadvantaged by past and present inequalities (e.g., women in management) receive favorable treatment in order to reduce imbalance and ultimately restore balance (Harrison, Kravitz, Mayer, Leslie, & Lev-Arey, 2006).

Reactions to such policies are not always positive and these policies do not only affect social equality but may have unintended consequences (Shaughnessy, Braun, Hentschel, & Peus, 2016). For instance, the degree to which the quota policies affect the decisions of the selection committee is likely to be inversely related to perceptions of justice (Harrison, Kravitz, Mayer, Leslie, & Lev-Arey, 2006). It follows then, that quota-based selection policies are likely to invoke strong injustice perceptions, especially for non-beneﬁciaries, because of the perceived violation of merit-based norms. The favorable treatment of disadvantaged groups makes doubts occur about the fairness of the quota policy and subsequently inﬂuence perceptions of the organizations as well as applicants’ job pursuit intentions (Shaughnessy, Braun, Hentschel, & Peus, 2016).

Why is it that females need a women’s quota to overcome the difficulties for promotions and equality? The percentage of women in board positions and their influence on firm performance has been a research topic of previous studies. Results show that firms with gender-diverse boards perform better due to the unique pool of resources and human capital that women bring to the business (Campbell & Minguez-Vera, 2008). The diversity of board members brings more resources to a firm. The increase in resources will result in higher firm value and performance. This will give a firm more access to resources that help reduce external dependency and diminish uncertainty (Hillman & Dalziel, 2003). But there seems to be a lack of research on the difficulties for women to even get into those board positions. This research

(7)

hopes to contribute to fill this gap of research by investigating whether these difficulties might be constructed within management control design.

Several studies have focused on the definition of a management control system (Ferreiraa & Otley, 2009). Simon (1995) views a management control system as the means used by senior managers to successfully implement their intended strategies. Others defined the management controls system as the use of management accounting in relation with other forms of control such as personal, cultural (Chenhall, 2003), result or action controls to achieve organizational goals (Merchant & Van der Stede, 2007).

Could it be that the use of certain management controls influences the experience of stress differently for male and female employees? An individual experiencing a higher level of stress may spend more time in coping with this stress and the efforts for job performance may be reduced, resulting in lower job performance (Muhammad, 1984). The possible negative influence of stress on performance might be a source that could explain the difficulties for women within business settings. If women experience more stress than men due to the use of certain management controls, women’s performance will probably decrease more than of men. Stress will in the rest of this study be addressed as professional tension, explained as: the degree

to which the employees feel that they cannot perform work in the way they believe it should be done (Rizzo, House, & Lirtzman, 1970).

Within the last couple of years, the overall numbers of burn-outs have risen to 42% of all occupational diseases, with 14% of all employees reporting burn-out symptoms yearly. These employees are on average 180 days unable to work. Absenteeism in the Netherlands is costing employers annually over 11 billion euros (Academisch Medisch Centrum, 2017). Burn-out symptoms like emotional exhaustion closely resemble traditional stress reactions that are studied in occupational stress research, such as fatigue, job-related depression, psychosomatic complaints and anxiety (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001). In 2016 female employees experienced more burn-out symptoms than male employees. Especially, young women between the ages of 25-35 experienced more mental fatigue symptoms due to work pressure (CBS, 2018). The percentage of women employees with a burn-out increased from 9,4% to 15% and male employees from 6% to 9% in the last two years (Witteman, 2017). In the last two years, women in the Netherlands experienced more stress symptoms than men. If women experience a higher level of professional tension with the use of certain management controls, this might explain the higher burn-out percentage among women. For organizations it could be valuable information if management understands better how to decrease the

(8)

experience of professional tension to avoid burn-outs and be able to save cost. This study will explore the influence that the management control features have on the different experience of professional tension between men and women. To be able to investigate if professional tension plays a role in the performance of employees, which could result in the difficulties of getting promoted into top positions.

Professional women working on managerial and professional-level within corporate settings, tend to have limited influence on promotion opportunities, financial rewards and time demands that negatively affected their job satisfaction (Chiu, 1998). Professional women with larger employment gaps were less involved in their jobs and less optimistic about their career prospects. On the other hand, organizations equalize a women’s temporary break in their employment with low commitment to their career (Chiu, 1998; Burke & McKeen, 1996). Despite the increasing visibility of women in professional and managerial positions, management remains male-dominated (Jick & Mitz, 1985). The lack of saying professional women have on these business aspects could be of influence on the difficulties women experience in business settings. This makes professional women an interesting population to investigate for this study.

If professional tension influences the performance of professionals and women experience a higher level of professional tension, this might be an explanation for why professional women experience more difficulties when trying to get promoted. The management control system might affect the experience of professional tension differently for men and women. This study, investigates the influence of the management control design on the different experience of stress between professional men and women and tries to answer the following research question: Should professional men and woman be managed differently, if

examining the influences of professional tension on the performance of professionals within professional service firms?

Literature gives limited evidence to make predictions regarding the control preferences of professional men and women for the design of the management control system. The lack of empirical evidence gives the opportunity to explore predictions related to this group. The study will contribute by filling in the literary gap on how the management control system influences the different performance of genders. This study contributes by explaining if factors of the management control system play a part in the gender inequalities still present in business settings. To eventually contribute, with advice for organizations if the design of the

(9)

management control system does influences the equal opportunities for both genders and could serve as an alternative for the women’s quota.

This paper is structured with the introduction of a framework explaining the possible effect of the management control system on professional tension and performance. Following, the theoretical review examines possible factors explaining the preferred controls by professionals. Next, the review discusses what kind of factors might influence the different experience of stress (professional tension) between men and women, to analyze which controls they may prefer. Chapter 4 describes, the survey used to gather the data for this study and the variables constructed to test the hypotheses. Chapter 5 contains the analysis of the results from the tests. The last chapter concludes with the discussion and the proposed suggestions for future research.

(10)

2 Theoretical review

2.1 Research

The effect of the management control system on the experience of professional tension is explained by the framework of Hackman and Oldham (1981), followed by the construction of the framework for this study. The frameworks are used to analyze whether the problems of getting women into top position lie within the organization.

2.2 Management control system

In scientific studies there occurs a lack of findings which explain, how professional tension is related to the management controls used within organizations. There is a gap in empirical evidence on the direct or indirect effect of those controls on work performance (Shields, Deng, & Kato, 2000). Due to the lack of extensive literature within the management research field on this subject, I will use the psychological framework developed by Hackman and Oldham (1981) as a management control system to examine what kind of organizational structure will influence professional tension and eventually performance.

Hackman and Oldham (1981) describe the following two frameworks in their research. The Attraction-Selection framework gives an explanation of the relationship between the organizational structure and the employee’s reaction, consisting of two steps. The first step suggests that the type of employee attracted and selected by an organization with a certain type of structural properties or conditions is dependent on the particular personal background of the employee. The second step proposes that employees’ reaction to work and the organization, will mostly be determined by the personal attributes of employees. Individuals different work attitudes and behavior are in a large part explained by these personal attributes of employees. Concluding the framework specifies that employees reactions are indirectly related to the properties of the organizational structure, through the attributes of employees.

Another framework explaining the relationship between the organizational structure and employee’s reaction is called the Job-Modification framework. This framework argues that the relationship of the structural properties of organizations influences employee’s reactions by shaping their job. This is again further elaborated by two steps. The first step states that the organizational structure significantly affects the number of challenges and complexity of the employee’s job. The second step explains how these job challenges and complexity directly

(11)

influence employee’s reaction to their work and organization. Concluding that within the Attraction-Selection framework the personal attributes of individuals are mediating the relationship between organizational structure and employee’s reactions and within the Job-Modification framework the job characteristics mediate this relationship. The findings of the research by Hackman & Oldham (1981) state that the relationship strongly is determined by joint characteristics of both models.

A combined framework which includes both the personal and job characteristics as mediators is more effective in explaining the relationship between organizational structures and associated reactions of employees (Oldham & Hackman, 1981). Further research resulted in more elaborated models of these discussed frameworks. These models involve intrinsic motivators and give more details about the influence of single components of job characteristics on psychological states that will lead to certain personal and work outcomes (Sharp, Baddoo, Beecham, Hall, & Robinson, 2009). Figure 1 gives the combined model with personal and job characteristics as mediators. I will use this model in my research and will be addressed as the Hackman & Oldham framework in the remaining of the paper.

Figure 2. Hackman & Oldham Framework (additions to the original Hackman and Oldham model are

underlined).

Goal clarity is an attempt by organizations to increase the probability that employees will behave in the way that leads to attainment of organizational goals (Flamholtz, Das, & Tsui, 1985). A management control system goal is to assist managers to achieve their goals or those of the organization (Chenhall, 2003). The controls that are used to design a management control system are a technique to achieve goal congruence within organizations. Result controls define

(12)

performance measures and rewards for job outcomes. Result controls influence actions of employees because they cause them to think about the consequences of the actions they take. With the use of result controls the organization does not dictate towards their employees which actions they should take. Employees are empowered to take actions they believe result in the desired outcomes. The use of result controls encourages employees to discover and develop their talents and get placed into job positions were they will perform well. Action controls prescribe and monitor actions of employees. These controls involve ensuring that employees perform certain actions in the way that the actions are most beneficial for the organization. Behavioural constraints are a form of action controls, that make it (almost) impossible to do things that should not be done. Action controls are only feasible when the manager knows which actions are desirable and have the power to ensure the right actions are taken (Merchant & Van der Stede, 2007). Action controls give clear instructions for the actions necessary to achieve the organization’s goals. This probably will result in more strict boundaries for employee’s behaviour to ensure the obtainment of goals and results in goal clarity. Result controls are more unclear on which precise actions need to be taken, but clear about which results need to be achieved to obtain the set goals. Therefore, result controls are able to provide goal clarity to employees.

Goal clarity and the use of a management control system could be important to reduce role ambiguity and role conflict, which are of influence on employees professional tension experience. Role ambiguity is defined as the lack of clarity and predictability of one’s behaviour. Role conflict is defined as the degree of incongruity or incompatibility of expectations associated with the role (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964). High role ambiguity within work environments has been hypothesized to lower peoples job-related self-esteem. Such a person might feel incompetent to perform due to the believe that the work environment is in fact orderly and that the experience of ambiguity is only reflecting their own inability to accurately assess and understand the role expectations. The correlation found between role ambiguity and role conflict with tension is categorized as ‘stress’ research (Jackson & Schuler, 1985). Katz and Kahn (1966) originally found similar results and theorized role stress as the results from an employee’s role ambiguity and role conflict. Role ambiguity results when the duties and actions are unclear to the employee. The strain associated in dealing with issues of unclear responsibility is related to job stress (Boles, Wood, & Johnson, 2003). Stress appears to be more closely related to the nature of the management control system than job satisfaction and is implicated in relation to performance (Shields, Deng, & Kato, 2000). Therefore, it is expected that the design of the management control system influences the

(13)

experience of professional tension. Furthermore, the framework of Hackman & Oldham (1981), indicates that not only goal clarity is of importance, but that autonomy also has an influence on professional tension and job performance. Professional organizations in general display distinctive features, such as professionals’ greater autonomy and participation in organizational governance (Von Nordenflycht, 2010). The presence of the preferred autonomy and active participation in organizations governance by professionals, makes professional organizations an interesting research setting and is addressed as professional service firms in the next paragraph.

Professional Service Firms can be distinguished from other organization by the following characteristics, knowledge intensity, low capital intensity and professionalized workforce (Von Nordenflycht, 2010). Knowledge intensity indicates the production of a firm’s output that relies on a substantial body of complex knowledge. Low capital intensity is a firm’s production that does not involve significant amounts of non-human assets, like inventory, factories and equipment, and intangible nonhuman assets like patents and copyrights. Professionalized workforce is the knowledge base of employees and the regulation and control of that knowledge base and its application. Scholars often argue that professionals have a stronger preference for autonomy and a consequent distaste for direction, supervision and formal organizational processes (Von Nordenflycht, 2010). Authority can be more problematic within knowledge intensive firms, this could make it harder to direct members and make them do things they do not want to do. Furthermore, is it hard for nonexperts to evaluate the quality of the expert’s output, even after the output is already produced and delivered, this is called ‘opaque quality’ (Von Nordenflycht, 2010).

The use of action and result controls to manage professionals can give the following problems. First, as professionals have a strong preference for autonomy, they wish to make decisions without external pressure from non-professionals. They can feel superior over others, due to their overspecialized technical skills. This creates settings where professionals feel they can only be evaluated based on professional standards by collegial controls and not by managers based on the rules and regulations of the organization. Secondly, the professionals’ education makes them develop solidarity feelings toward their field and makes them less motivated for monetary rewards. These characteristics might make it difficult to make professionals obey rules and regulations enforced on them and makes them resistant to managerial control. This can lead to conflicts when organizations force professionals to act in line with the firm’s rules and regulations. To avoid these conflicts there need to be the right

(14)

level of controls versus autonomy, which is satisfying for both parties (Hall, 1968; Raelin, 1985).

Consequently, a proper management control design supports goal clarity which will reduce role ambiguity and role conflict, this is expected to reduce professional tension. However, too many controls could have a negative effect on the motivation and performance of professionals due to the feeling of autonomy loss. I constructed the framework in figure 2 representing these expected relations based on the framework of Hackman and Oldham (1981). This framework predicts, how in professional service firms the controls relate to professional tension (psychological states) which influences performance (personal and work outcomes) and what eventually the influences of gender might be on these relations. The next chapter will discuss why it is expected that there is a difference in the displayed effects in Figure 2 between professional men and women.

Figure 2. Relation: MCS - Professional Tension - Performance

Management Controls: Action implicit Action explicit Result Professional Tension Performance: Career Perspectives Job Satisfaction Work Output Gender Control variables: Strategy Firm Size Unit Size

(15)

3 Hypothesis development

3.1 Professional service firms

Characteristics that distinguish work of professionals from other services is that their work outcomes require specialist knowledge and skills developed by lengthy and formal higher education. The autonomy and objectivity of professionals are often the first reason why clients acquire the services of the professionals. Clients approach professionals to benefit from their knowledge, which they cannot easily obtain themselves. This obtained knowledge gives professionals the right to determine the means for the solution of client’s problems. Professionals’ knowledge may also be based on field experience and is also a determinant of the abilities of a professional. Solving a client’s problem will eventually be based upon specialist knowledge, skills and experience. A consequence of the information asymmetry between the professional and their clients is, the unclarity of the client’s need with regard to the service and inability to evaluate the service also after the purchase and consumption (Jaakkola & Halinen, 2006). Scholars frequently list customized output as a key feature of professional service firms, who apply their expertise to each specific client request (Von Nordenflycht, 2010).

The research of this study will be conducted among professional service firms. Where professional tension is expected to be present within the work environment due to the high time pressure professionals work under while following required procedures (Arens, Loebbecke, Lemon, & Splettstoesser, 1987). Looking at the auditing profession concerns arise when too much time pressure will lead to auditing failure. To be able to operate competitively firms need to work effectively and efficiently. Professionals need to gather evidence sufficient to meet professional requirements and audits need to be effective while controlling cost. Many large public accounting firms moved towards a more structured audit approach, because of the need to implement a consistent and effective audit approach (Arens, Loebbecke, Lemon, & Splettstoesser, 1987). To control audit effectiveness, firms develop designed audit programs in such a way that, if all procedures are completed and correctly executed, the target audit risk will be achieved. To achieve this goal, programs consist of detailed procedures or structured checklists, which facilitate compliance with generally accepted auditing standards, thus providing quality control of audit work. Structured programs are also purported to improve audit efficiency (Cushing & Loebbecke, 1986).

(16)

the effectiveness of two accounting groups. The one group worked with structured programs and the other one with unstructured programs. A structured program details the amount of work to be completed and implies a minimum performance standard that is achievable only by completing the program. An unstructured program allows the individual to plan and organize his or her efforts to achieve the implied performance standard. She stated that time budgets are one of the most common means of cost control used within public accounting firms. These budgets have shown to improve effectiveness per unit of time. Meanwhile, this time pressure is an externally imposed restriction on performance and a main source of stress for accounting professionals. Time pressure produces stress if work perceived as necessary cannot be completed. However, her results showed that program structure in combination with high levels of time pressure did not benefit either audit effectiveness or efficiency. This combination of structure and time pressure increases task-related demands, which can result into an additional source of stress. Her empirical results indicate that the overall audit effectiveness declines as time pressure increase. The audit efficiency within groups increased faster when no structure was imposed on the professionals, relative to groups where structure was demanded (McDaniel, 1990). Time pressure combined with the stress from imposed program structure will decrease effectiveness relative to the absence of the demanded structure (Beehr & Newman, 1978).

The Hackman & Oldham framework indicates the importance of goal clarity on the meaningfulness of work and work motivation. I expect that with the use of action controls, role conflict and role ambiguity will be less present than with the use of result controls, as action controls prescribe the actions necessary to achieve goals. I assume that when professionals are forced to work under result controls to meet certain performance outcomes their experience of professional tension will increase. Result controls might make it more difficult for professionals to know which behavior and actions are required to obtain the set goals by the organization and might lead to more role conflict and role ambiguity. Furthermore, with the use of result controls professionals will probably lose the feeling of autonomy to solve their client’s problems as they think is best. It is more difficult for professionals to achieve set result outcomes due to the special attention every new situation requires and the determining role of the clients’ input. The loss of autonomy also influences the responsibility for- and quality of work outcomes and the job satisfaction, as seen in the framework of Hackman & Oldham. However, the results from McDaniel’s (1990) research indicate that imposed structures on professionals, which could be interpreted as action controls did not benefit the performance of

(17)

accountants either. I want to test the following hypothesis to analyze the influence of controls on professional tension:

H1: Result controls will affect the experience of professional tension by professionals.

H1.2: Action controls do not affect the experience of professional tension by professionals.

3.2 Gender Differences

Men and women experience work stress differently due to genetic, structural and social/psychological factors. The meta-analysis study of Jick and Mitz (1985) found evidence that these three factors are of influence of the different experience of stress symptoms between men and women. Male genetics may establish a greater readiness for aggression in males but is potentially also shaped by social learning and experience. Young boys are encouraged to play aggressively but such behavior in young girls is ignored or punished. As long as organizations and institutions reward such behavior primarily in men, more aggressive and competitive behavior will be found in men. Structural aspects explain that work-related stress is primarily a function of the different work roles occupied. Women are seen to be relatively disadvantaged with respect to opportunities for mobility and influence. Employed women are more distressed than working men, due to the absence of mentors or female support groups, lower salaries, career block, and "masculine" job stereotypes. Lastly, the different internal response of men and women to stressors can be explained by social/psychological factors. Men and women have a different cognitive appraisal of stress and strategy for coping with felt stress. This reflects differences in how stress is perceived by men and women.

To my knowledge, there has been little empirical evidence found to explain how a management control design could influence the different experience of work stress between genders. There is reasonable evidence to expect that the management control system might have an impact on this experience. Previous studies have already addressed several different reactions of men and women towards aspects which are under influence of the management control system, such as role conflict, extrinsic motivators and handling risk. The differences between man and women caused by these factors could indicate what might be preferred management controls and this way probably reduces the experience of professional tension. The influence of role conflict, extrinsic motivators and handling risk are further analyzed in this section.

To start, the experience of role conflict is different for men and women. Women tend to experience a higher degree of role conflict than men, because next to the role of

(18)

professional/employee their role of ‘mom’ also interferes with their job. These experiences could be explained by the difference due to cultural and societal expectations. For instance, expectations regarding parental responsibilities are strongest for mothers, they therefore feel more obligated to put family matters first at expense of their job (Gilbert, 1992; Marsiglio, 1993). Women tend to get stressed by the cultural expectations of being a ‘good wife’, ‘good mother’ and ‘good employee’. Part of this stress is the result of carrying more responsibility than their spouses for the maintaining a family life (Crosby & Jaskar, 1993). Men tend to put their job first, which results in a lower perception of work-family conflict than for women. Therefore women probably experience a higher degree of role conflict resulting in role stress (Gilbert, 1992; Marsiglio, 1993). When organizations use compensating pay it will reduce work-family conflicts for men. As men generally in stereo-typed families are expected to conform to the role of breadwinner. Men feel less conflicted about not spending enough time with their family if they are making more money. This tactic however, has little influence on women. Pay satisfaction is not related to female employees perception of working family conflicts. Sometimes, it even can be more effective for organizations to hire an additional employee to reduce the feeling of the high level of work-family conflict for female employees (Boles, Wood, & Johnson, 2003). There are at least two other possible theoretical explanations for different outcomes of role stress for males and females. Firstly, the concept of physio-biological difference leads to different desires for management control systems. Especially women seek work roles that do not interfere with their family roles (Boles, Wood, & Johnson, 2003). The second theoretical perspective predicts that women are more into socialized communal behavior and men more into instrumental values. This difference is explained by the different skills men and women learn by frequently performing different social roles. More women than men stay home as caregiver and therefor probably learn more nurturing skills. Men more than women tend to occupy leadership roles in companies and learn assertive and aggressive skills required for such positions (Eagle, 1987). These different theoretical perspectives suggest men and women will have different expectations of their job conditions (Boles, Wood, & Johnson, 2003).

Hackman & Oldham (1981) state in their study, that different personal attributes influence the reactions of employees. The different personal characteristics which distinguish professional men and professional women from each other indicate that their reactions to certain organizational controls probably will be different. The different reaction by professional men and women might result in experiencing professional tension for the one and not for the other. I would expect that professional women might prefer to work with action controls,

(19)

because action controls prescribe the expected actions and behavior of employees. When women know which precise actions are required and expected from them in their work environment, it probably decreases role ambiguity and role conflict for them as professional/employee. This probably reduces the stress to also fulfill their role as a parent. When working with result controls the way to achieve the set result might be more unclear. It could encourage employees to discover and develop their talents, when they need to make their own decision to achieve desired goals. However, I assume because women tend to have an extra role to fulfil they probably are less willing to put much extra effort into discovering and developing their role as professional/employee. But prefer to have more clear boundaries between these roles to reduce role conflicts. Men on the other hand, probably prefer the possibility to be extra compensated on their result outcomes to reduce work-family role conflicts and fulfil their role as breadwinner better. Also as men tend to have more aggressive and assertive skills, it can be expected that they do not want to be monitored constantly and need to feel the freedom to use these skills to make their own decisions for certain actions. These characteristics suggest that professional women probably experience less professional tension with the use of action controls than professional men. I will test these expectations with the following hypothesis in my study:

H2: Professional women experience less professional tension with the use of action controls

than professional men.

Another reason to assume that women prefer action controls over result controls could be the more risk-averting characteristics of women than men (Croson & Gneezy, 2009). For instance, women tend to prefer a set salary as opposed to variable compensation and are more inequality-averse than men. Men tend to have a higher risk-tolerance than women and would more easily view a risky situation as a challenge than a threat. Men have a different perception than women of the probability distribution underlying a given risk. This is related to the different initial reaction of men and women to risk in general. These reactions affect the evaluation of the probabilities and outcomes of certain risks. Based on this evaluation men and women’s initial reaction will make them take different decisions when it comes to risk (Borrenbergs, Rui, & Georgakopoulos, 2017). The generally higher risk tolerance of men makes them probably experience less professional tension when working with result controls than women who are more risk-averse. This assumption is made because result controls do not dictate which actions employees should take. Employees need to think about their actions and consequences of these

(20)

actions. Men would probably see the risk involved with uncertain outcomes as opportunities and probably feel more comfortable with taking risks with the desired result in mind. Women are expected to have higher preferences to know what action to take and not to worry about the outcomes and being rewarded with a set salary. I expect that the professional tension of women would rise more when these aspects become more uncertain, due to their more risk-averse characteristics than men. To be able to find support for these assumptions I will test the following hypothesis:

H3: Professional women experience more professional tension with the use of result controls

than professional men.

3.3 Influence of professional tension on performance

In this study, the negative influence of professional tension on job performance will be addressed by the following casual effect: when an employee perceives the demand addressed at them directly as threatening or harmful, they will use up time and energy to be able to overcome this stressor and the immediate reaction to this situation is discomfort and or anxiety. Work-related stressors tend to divert effort away from performing job functions towards coping with these stressors and reducing employees ability to perform. Eventually, the stressors can create situations of information overload, which will also negatively affecting job performance. High level of stressors could thus lead to a narrowing of individuals’ perceptual attention and ignoring performance-related information (Gilboa, Shirom, Fried, & Cooper, 2008).

Time pressure, demand for structure and regulated decision-making processes are expected to be work-related stressors for professionals and might increase the experience of professional tension. For instance, the personal resources of professionals such as doctors, lawyers and auditors, rapidly decrease by the required effort in the decision-making processes and leaving the executive function less efficient in performing several tasks. Even though, the need for competent functioning requires greater cognitive resources due to the increase of information processing. If professionals reach the limit of their cognitive capability, their performance can be undermined (Ceschi, Demerouti, Sartori, & Weller, 2017). I expect that the casual effect is also of influence on the performance of professionals and I expect to find support for hypothesis 4:

H4: The experience of professional tension has a negative effect on the performance of professionals.

(21)

Role ambiguity and role conflict are expected to have negative consequences on job performance. The negative relationship can be explained by cognitive and motivational processes. As role ambiguity can be conceptualized as lack of knowledge of the most effective job behaviors and role conflict occurs when conflicting information makes individuals unable to do everything that is expected. From the cognitive perspective, both role ambiguity and role conflict result in lower level of performance, since they either represent lack of information or information overload. From the motivational perspective, role ambiguity and role conflict weakens the effort to perform and performance to reward expectancies. Employees whose job performance largely depend on interaction with others are more likely to experience role ambiguity than employees where performance depends on specific tasks (Jackson & Schuler, 1985). Empirical evidence supports the negative relation between role ambiguity and job performance, but does not support a meaningful relationship between role conflict and job performance (Jackson & Schuler, 1985; Tubre & Collins, 2000).

Within more complex jobs, the lack of formalization of work activities might lead to increased levels of role ambiguity. Job complexity increases with more decision-making authority, therefore it is expected that greater level of role ambiguity than role conflict occurs in these jobs (Naylor, Pritchard, & Ilgen, 1980).

The correlation found between role ambiguity and role conflict with tension was categorized as ‘stress’ by Jackson & Schuler (1985) and Katz & Kahn (1966). Professionals are expected to experience a higher level of role ambiguity since their work is dependent on others like clients. Professionals generally prefer to work with autonomy and the authority to make decisions, which thus can be categorized as complex jobs. Section 3.2 suggested, that women probably experience a higher level of role conflict due to the conflicting roles of professional and mother. Therefore, it is expected that professional women are more sensitive to a higher correlation between role conflict and role ambiguity leading to stress. The higher correlation for professional women is expected to have more consequences for their job performance than for professional men. The previous sections suggested that role ambiguity and role conflict probably occurs more when working with result controls than working with action controls. The use of certain management controls therefore might affect the correlation between role ambiguity and role conflict and influence on the performance. I will test this assumption with the following hypothesis:

(22)

H5: The experience of professional tension decreases the performance of professional women more than of professional men.

(23)

4 Research Method

4.1 PSF: Professional Service Firm Thesis Survey Project 2017-2018

This chapter describes the research method used to examine the hypotheses developed in chapter 3. For this study, I joined the research project at the University of Amsterdam: the PSF Professional Service Firm Thesis Survey Project 2017-2018. This project investigates the way professional service firms design their management control system. It examines how these management control systems affect the professionals working in professional service firms. The data for this research was gathered by a pre-developed online-survey. Data is collected by students who joined the project. To control the reliability of the respondents, participants provide students with an email stating that they completed the survey. The questionnaire consists of questions related to control measures, professional tension and performance. In the next section, I will further elaborate how questions in the survey are constructed into variables for my research.

4.2 Sample

The participants need to meet certain criteria to be eligible as a respondent for the project. Only employees of mid-level to large firms are able to participate and the professional service firms include occupations like lawyers, engineers, bankers, pharmaceuticals, financial advisor, IT/HR/Technology consultants and software/graphic/fashion designers. To specify the sample no occupations within non-profit, publicly owned organizations such as universities (education/teaching), government organizations, NGOs and social work agencies were included. With the exception of medical doctors who might be employed at public or not-for-profit hospitals. Lastly, the respondents and the organizational units they manage must have the following characteristics:

• The respondent has worked the field for more than 3 years (but no more than 10 years). • The respondent is not an owner/partner or board member of the organization. In other words, the respondent needs to be subject to the management control system rather than design it.

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

A total of 730 responses were received. 92 respondents were deleted because they did not finish the survey. From the remaining 638 respondents, 247 respondents had less than 3 years or more

(24)

than 10 years of work experience and were removed from the sample. Two respondents who did not fill in their gender were removed because it is essential for my study to know the gender of the respondent to be able to compare the experience of the used controls. Finally from the remaining 391, the last 10 respondents who did not read the remarks regarding the survey were removed. Leaving a sample of 379 respondents for my research. Generally, this is a good sample size, as I want to examine a medium effect which requires a sample size of 200 to be sufficient (up to 20 predictors). With six or fewer predictors, which is the case in my study, a sample size of 100 would also have been sufficient for a regression (Field, 2009). The sample consist of the following male – female distribution:

Table 4.1 – Gender Distribution

Frequency Percent Male 245 64,6% Female 134 35,4% Total 379 100,0% 4.3 Variable constructs 4.3.1 Control variables

To compute the variables which can be used as the management controls for this study, I conducted a factor analysis. The questions asked in the survey to measure control responses are based on existing literature basic definitions of types of control from Ouchi (1979) and similar conceptualizations are also found in Merchant (1982), Perrow (1970) and Snell (1992). To compute these constructs I used the questions addressed in the survey under: ‘Behaviour Control – These questions address the use of rules and procedures in your organization’ and the questions addressed in the survey under: ‘Result Controls – These questions address measurement of performance’. These questions can be found in Appendix A. containing the survey or in Appendix B. containing the questions used for each variable. The questions 4.10 – 4.13 and 5.9 – 5.11 are recoded before running the factor analysis. In the questionnaire these questions needed to be answered on a 5-point Likert-scale, which also applies for the constructed variables in the next paragraphs. For the factor analysis, I used Varimax and suppressed all values below 0,5. For the controls, the questions 4.8, 4.11, 5.10 and 5.12 were eventually removed because they did not load with any other questions in the rotation matrix. The computed constructs are shown in table 4.2 – Constructs Controls.

(25)

Table 4.2 – Constructs Controls

Constructs Controls Item 1 2 3 4 Cronbach's

Alpha Number of Items Missing Values

Explicit Action Control Q4.1 0,806 0,793 3 1,8%

Q4.2 0,749

Q4.4 0,746

Explicit Result Control Q5.1 0,696 0.794 3 1,8%

Q5.4 0,753

Q5.6 0,760

Q5.7 0,754

Implicit Result Control Q5.9 0,713 0,350 2 1,0%

Q5.11 0,714

Implicit Action Control Q4.12 0,820 0,753 3 1,3%

Q4.13 0,793

Q4.10 0,803

Table 4.2 shows that 4 constructs result from the factor analysis. In this study, the four controls will be addressed as mentioned in table 4.2. I interpreted the questions in the survey based upon the literature of Ouchi (1979) Merchant (1982), Perrow (1970) and Snell (1992) and named the variables implicit and explicit controls. I interpreted explicit controls mostly as established rules and procedures imposed upon an employee by the organization. Implicit controls are more flexible procedures which can be adjusted by the employee when it suits specific situations or tasks better. The precise questions of the survey which are taken into account for the construct, are found in Appendix B. The Cronbach’s Alpha values are acceptable surrounding 0,7 and higher (Field, 2009). Implicit result control is the only construct without a reliable Cronbach Alpha result from the factor analysis. The implicit result control will not be used for the rest of my research. The missing values are all acceptable, generally missing values for a missing specific variable is 10% (Field, 2009). The Kaiser-Meyer-Olkin value for the control constructs was 0,803 and the Bartlett test was significant (χ2 = 1476.931, p-value = 0.000). The total variance explained is 58,99%. These outcomes give strong support to assume correlation between the variables (Field, 2009).

(26)

4.3.2 Professional Tension

The questions to measures professional tension in the survey among professionals is based on the definition by the literature of Rizzo et al. (1970), the literature of Aranya and Ferris (1984) and new items. To test the hypothesis the variable professional tension is constructed by a factor analysis (principal components analysis). To compute this variable I used the questions 33.1 – 33.8, the questions in the survey addressed under: ‘Professional Tension - These questions address some additional aspects of your work’. I selected Varimax rotation and suppressed loadings below 0,5 to improve the readability. The reliability test showed that 6 respondents were excluded, this is 1,6% of the total sample which is acceptable. The result of the Cronbach Alpha test showed a result of 0,817. The total variance explained by the variables was 45,58%. The Kaiser-Meyer-Olkin value was 0,892 and the Bartlett test was significant (χ2 = 841.280, p-value = 0.000), these result shown in table 4.3 assume that there is correlation between the variables (Field, 2009).

Table 4.3 – Professional Tension Construct

Professional Tension Item 1

Cronbach's Alpha Number of Items Missing Values Professional Tension Q33.5 0,798 0,817 5 1,6% Q33.8 0,769 Q33.4 0,760 Q33.7 0,710 Q33.6 0,630 Q33.2 0,625 4.3.3 Performance

The performance measure is based upon the existing measures developed by the studies of Podsakoff and Mackenzie (1989) and Welbourne et al. (1989). The factor analysis is done the same way as the previously explained factor analysis. To compute this variable I used the questions 36.1 – 36.12, the questions in the survey addressed under: ‘Individual Performance - How would you rate your own job performance in term of the following:’. Question 36.4 was deleted as it did not correlate with any of the other questions. The factor analysis gave the following results, shown in table 4.4.

(27)

Table 4.4 – Construct Performance

Construct Performance Item 1 2 3 Cronbach's

Alpha Number of Items Missing Values Performance - Career Q36.5 0,83 0,838 4 1,3% Q36.7 0,819 Q36.8 0,789 Q36.6 0,689 Performance - Innovation Q36.10 0,803 0,822 4 2,1% Q36.9 0,778 Q36.11 0,773 Q36.12 0,662 Performance - Output Q36.3 0,799 0,697 3 0,8% Q36.2 0,755 Q36.1 0,648

The total variance explained by the variables was 62,01%. The Kaiser-Meyer-Olkin value was 0,863 and the Bartlett test was significant (χ2 = 1664.378, p-value = 0.000), the results show we can assume there is correlation between the variables (Field, 2009). The Cronbach alpha value for Performance output of 0,697 is just below the acceptance level of 0,7, but Field (2009) also discusses additional literature which accepts values below 0.7. For instance, Kline (1999) states that when dealing with psychological construct values below 0.7 can be realistic because of the diversity of the construct being measured. The performance constructs are self-reported measure, where professionals rate their performance on diverse job aspect. The low Cronbach Alpha could be explained because the performance output of professionals is more difficult to measure. In chapter 2 this is explained as the concept of opaque quality. In general, professionals performance output are intangible products (software, patents, legal rights) or services that result in tangible output transferred to customers (blueprints, reports, contracts). For non-professionals, even after receiving the product/service it is still hard to evaluate the performance outcome (Von Nordenflycht, 2010). In this case, the Cronbach Alpha is so close to the 0.7 cut off, I choose to use the constructed variable in the rest of my study.

The three conducted variables will be addressed as presented in table 4.4. To name the variables, I interpreted the questions in the survey as follows. Performance output is closest related with the quantity and quality of the performance output of professionals. Performance career is related to the career perspectives like developing and obtaining career goals.

(28)

Performance innovation is mostly related to improving and innovating of business processes. Once again the specific questions of the survey for the three constructs are displayed in Appendix B, to give more examples for these explanations.

4.3.4 Control Variables: strategy, firm size and unit size

4.3.4.1 Strategy: Cost and Differentiation

The conventional contingency approach assumes that management control systems are adopted to assist management to achieve the desired objectives. The right design to achieve these objectives is influenced by the context of the firm’s operations. Management control systems attempt to align strategy with operations by translating official goals into operative goals, this can be different in similar organizations. Evidence from recent studies suggests that strategies characterized by conservative perspectives, defenders and cost leadership are best served by centralized control systems. Strategies with an entrepreneurial orientation and product differentiation better fit with decentralized, flexible structures and result orientated control systems. It is useful when evaluating characteristics of management control systems on goals to provide links with the strategy of the firm (Chenhall, 2003). The strategy is expected to have an influence on the used management control system, but is not of essence in this study. I expect that there will be two strategies computed from the survey data: strategy cost and strategy differentiation.

The questions in the survey to measure the organization’s strategy focus on the cost or differentiation aspects are based on the literature of Auzair and Langfield-Smith (2005). To compute the control variables for strategy, I ran a factor analysis the same way as the previous tests, on the questions 40.1 – 40.11 addressed in the survey under: ‘Strategy – In your organization, how much emphasis is placed on each of the following activities?’. Table 4.5 shows the results of the factor analysis.

(29)

Table 4.5 – Construct Strategy

Construct Strategy Item 1 2 Cronbach's Alpha Number of Items Missing Values Strategy - Differentiation Q40.6 0,725 0,716 6 1,3% Q40.11 0,632 Q40.9 0,619 Q40.7 0,597 Q40.5 0,575 Q40.10 0,550 Strategy - Cost Q40.8 0,785 0,697 5 1,3% Q40.2 0,759 Q40.1 0,694 Q40.3 0,539 Q40.4 0,455

The total variance explained by the variables was 44,53%. The Kaiser-Meyer-Olkin value was 0,794 and the Bartlett test was significant (χ2 = 911.622, p-value = 0.000), these result show we can assume there is correlation between the variables. The Cronbach Alpha shows a result for strategy- cost of 0,697, which is just below the acceptance level of 0,7 (Field, 2009). This could be explained because cost strategy within PSF is more difficult to determine and measure. As mentioned in chapter 3 scholars frequently listed customized output as a key feature of professional service firms, who apply their expertise to each client specific request (Von Nordenflycht, 2010), which could explain that only differentiation strategy occurs within Professional Service Firms. For future research, it might be relevant to investigate, what other kinds of measured strategies might occur within PSF. For this study I will accept the construct of strategy-cost even though it is just below the acceptance level.

4.3.4.2 Firm Size

Large firms employ more formal management control systems, with the possibility that different formal control types are appropriate for these large firms depending on the precise size. The actual measure of the firm size could be an element for the context and dimension of the management control system placed. Unfortunately, there is still little research done on small and medium-size firms, even though it is likely that formal and interpersonal controls differ depending on size (Chenhall, 2003). Also, the study of Auzair and Langfield-Smith (2005) showed a significant relation of the management control design based on life cycle, differentiation strategy and service process type within larger firm samples. The used

(30)

measurement item regarding the firm size is new. Expecting that firm size could have an influence on the model, I will use firm size as the next control variable. I created a dummy variable with the firm size < 500 coded as 0 and the firm size ≥ 500 coded as 1.

4.3.4.3 Unit Size

Montagna (1968) study showed that the units size is related to the used controls. Professionalization of organizations is expected to be positively related to centralization and size, with the administrative components spread throughout the organization. Montagna (1968), investigated the eight largest accounting firms in America. Generally speaking, professionals are assumed to have preferences which conflict with bureaucracy related to the professionalization of organizations (Von Nordenflycht, 2010). However, this was not the case in the study of Montagna. The administrative components were spread throughout the organization, but the accountants worked in smaller teams. There was no need for bureaucratic control mechanism in the smaller units. The size of units could be of potential influence on the experience of professional tension, but I won’t further elaborate on this relationship. To control for the influence of unit size I created a dummy variable for unit size < 50 codes as 0 and unit ≥ 50 as 1.

4.4 Analyses

In my study, I want to test the assumption that the independent variables result controls, implicit and explicit action controls used to design the management control system influences the depended variable: the experience of professional tension of employees. Next, I want to test whether there is a difference between the experience of professional tension by professional women and men when certain controls are being used within the management control system. To test the hypotheses stated in chapter 3, I will use the following models in a regression analysis:

M0: Professional Tension= β0 + strategy + firm size + unit size + ε

M1: Professional Tension= β0 + result explicit + action explicit + action implicit + gender +

strategy + firm size + unit size + ε

M2: Professional Tension= β0 + result explicit + action explicit + action implicit + gender

+ Gender × Result explicit strategy + firm size + unit size + ε

M3: Professional Tension= β0 + result explicit + action explicit + action implicit + gender

(31)

M4: Professional Tension= β0 + result explicit + action explicit + action implicit + gender +

Gender × Action implicit strategy + firm size + unit size + ε

To test the assumption of the negative effect of professional tension on performance, I will use the same construct for professional tension only this time as the independent variable. The three different constructs of performances will be used as depended variables. To be able to test the assumption if professional tension has a more negative impact on the performance of female professionals than of male professionals, the following models are used.

M5-8-11: Performance – Career/Innovation/Output = β0 + strategy differentiation+ strategy

cost + firm size + unit size + ε

M6-9-12: Performance – Career/Innovation/Output = β0 + professional tension + gender +

strategy differentiation+ strategy cost leadership+ firm size + unit size + ε

M7-10-13: Performance – Career/Innovation/Output = β0 + professional tension + gender +

professional tension × gender + strategy differentiation+ strategy cost + firm size + unit size + ε

(32)

5 Results

5.1 Descriptive and Correlation Tables

Table 5.1 presents the descriptive table for the constructed variables discussed in chapter 4. The table shows that for the Professional Services Firms explicit action controls have the highest mean. This might be explained by the possible preferences for the use of action controls within Professional Service Firms because performance results are more difficult to measure for professionals. Interesting to see that professional rate their performance outcomes the highest and their career possibilities the lowest.

Table 5.1 – Descriptive Variables

N Minimum Maximum Mean

Std. Deviation

Professional Tension 379 1,00 4,60 2,3245 0,82401

Strategy Differentiation 379 1,00 5,00 3,7288 0,64952

Strategy Cost 379 1,00 5,00 3,5540 0,70048

Explicit Action Control 379 1,00 5,00 3,1856 0,91674

Implicit Action Control 379 1,00 5,00 2,8325 0,87318

Explicit Result Control 379 1,00 5,00 2,8630 0,86705

Performance - Career 379 1,00 5,00 3,4908 0,84935

Performance - Innovation 379 1,00 5,00 3,5706 0,76146

Performance - Output 379 1,67 5,00 3,9446 0,61352

Valid N 379

Table 5.2 gives the results for the correlation between the constructed variables. The tables show a significant correlation between the constructed control variables. The values are below 0,4, multicollinearity probably will not cause any problems. The correlation between the different performance variables is above 0,4, which could indicate problems with multicollinearity. It is expected that the performance variables correlate. However, the three variables are functioning as dependent variables in separate regression analysis, therefore the correlation will not cause problems.

(33)

Table 5.2 – Correlation Variables

Pearson Correlation

1 2 3 4 5 6 7 8

1.Professional Tension

2. Explicit Action Control -0,010

0,853

3. Implicit Action Control 0,049 ,170**

0,344 0,001

4. Explicit Result Control ,114* _,349** _-,109*

0,027 0,000 0,034 5.Strategy Differentiation -,106* _0,038 _-,207** _,137** 0,039 0,459 0,000 0,008 6.Strategy Cost -0,022 ,106* _-,146** _,208** _,454** 0,677 0,040 0,004 0,000 0,000 7.Performance Career -,218** _,148** _-,140** _,151** _,143** _,127* 0,000 0,004 0,006 0,003 0,005 0,013 8.Performance Innovation -,110* _0,043 _-,181** _0,081 _,206** _,183** _,469** 0,032 0,400 0,000 0,116 0,000 0,000 0,000 9.Performance Output -,220** _,115* _-,151** _,110* _,191** _0,065 _,421** _,462** 0,000 0,025 0,003 0,032 0,000 0,207 0,000 0,000

(34)

5.2 Testing the influences of management controls on professional tension

In this section, the results of the regression analysis are discussed. The four models together examine whether professional tension is related to the use of result or action controls and how gender has an influence on this relationship.

Table 5.3 – Regression Analysis Professional Tension

Model: 0 Model: 1 Model: 2 Model: 3 Model: 4

Dependent Variable Professional Tension Control variables Strategy differentiation -0,114*** -0,112*** 0,107*** -0,112*** -0,112*** Strategy cost 0,038 0,021 0,014 0,021 0,021 Firm size -0,045 -0,035 -0,033 -0,035 -0,035 Unit size 0,043 0,032 0,021 0,032 0,032 Independent Variables

Explicit Action control -0,067 -0,064 -0,067 -0,068

Implicit Action control 0,059 0,048 0,059 0,051

Explicit Result control 0,145** -0,013 0,145** 0,145**

Gender = Female(0) vs. Male(1) 0,021 0,024 0,021 0,001

Interaction Variables

Explicit Result control*Gender 0,200**

Explicit Action control*Gender -0,001

Implicit Action control*Gender 0,022

R2 0,013 0,033 0,047 0,033 0,033

R2 Adjusted 0,002 0,011 0,024 0,009 0,009

F-statistic 1,212 1,539 2,012** 1,365 1,365

P(sig) 0,305 0,142 0,037 0,203 0,201

* Correlation is significant at the 0.01 level (2-tailed). ** Correlation is significant at the 0.05 level (2-tailed). *** Correlation is significant at the 0.10 level (2-tailed).

In hypothesis 1 and hypothesis 1.2, I predicted the following:

H1: Result controls will increase the experience of professional tension by professionals.

How to manage men and women, when men are from Mars and woman from Venus?