Amsterdam Business School
The effect of individual, group-based or hybrid performance
metrics on job performance in industry types
Name: Eda Turker
Student number: 10838929
First thesis supervisor: ir. S. van der Heide
Second thesis supervisor: mw. dr. ir. B.A.C. Groen Date: 20 June, 2016
Word count: 14.929
MSc Accountancy & Control, specialization Control
Statement of Originality
This document is written by student Eda Turker who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
ABSTRACT
The aim of the current study is to investigate: a) the impact of use of individual, group-based, and hybrid performance indicators on the employee job performance outcomes; and b) the
moderation effect of production firms, mass service firms, and professional service firms on the employee job performance outcomes with the use of the previous performance indicators. The lack of knowledge of specific relationship within type of performance indicators and industry types has contributed to investigate the variance within the subjects. Based on a survey among 106 operational employees and their direct supervisor, I find the following results: the use of individual and hybrid performance indicators do not have any impact on the employee job performance. In addition, the use of group-based performance indicators negatively impacts the employee job performance. Moreover, the moderation analysis do not prove any positive or negative impact on employee job performance when individual, group-based, and hybrid performance indicators are used. Contrary, the job tenure seemed to have a positive direct relationship with employee job performance. However, the moderator effect of job tenure on employee job performance seemed to change into an insignificant relationship.
Keywords: Individual, group, hybrid, performance measurement, production, mass service,
Contents
1 Introduction ... 5
2 Literature review and hypotheses ... 8
2.1 Performance Measurement Systems ... 8
2.2 Individual/ Group Measures & job performance ... 11
2.2.1 Individual performance measurement ... 12
2.2.2 Group-based performance measures ... 13
2.2.3 Hybrid performance measures ... 15
2.3 Difference in industry types ... 17
3 Research methodology ... 20
3.1 Data collection & respondents ... 20
3.2 Variable measurement & Analysis ... 21
3.3. Statistical analyses ... 23
4 Results ... 25
Additional Analysis ... 31
5 Concluding Discussion... 34
Limitations & Future research ... 36
References ... 38
1 Introduction
Performance measurement systems are used in organizations to translate their strategies and missions into operational actions (Henri, 2006). The operational actions are mainly based on the type of worker who perform in line positions. These line-workers influence the organizations’ value by their participation in all kind of organization, like the workers in manufacturing
organizations, mass service firms or professional service firms. So the performance of firms are related to their employees’ effort. Some research (McKinnon & Bruns Jr, 1992) has concluded that operational employees are firm specific and that there could be a relation between the difference in work property and activities of different employees. So this difference is taken into account when a firm is considering to develop performance measures (Burney & Matherly, 2007; Chenhall & Brownell, 1988). Also, according to a case study (Groen, Wilderom, & Wouters, 2015)the operational performance measures are mainly designed to obtain a more detailed view of the work that is performed. This study explained these operational performance measures as information that provides information about activities that are meaningful to the organization, because it gives insight to the defects, customer satisfaction, and timeliness of the activities. Groen, Wilderom and Wouters (2015) also describes in their study that the development of performance measures should be determined by looking at the core operational performance that can only be influenced by themselves. Differently, it would not be ethical to hold an employee accountable for something that cannot be influenced by him or her.
Another point of view is investigated by Burney and Matherly (2007) that performance measurement systems are implemented by looking at the organizational design. Afterwards, these specifications are transformed into measures that operational employees are accountable for. So these systems represents a relationship between the management accounting and information systems. Burney and Matherley (2007) also stated that management accounting literature recognize the effectiveness of performance measurement systems, because this depends on the environment in which they are used. Some other studies only focused on contingency-based investigation in combination with performance measurement systems, like the balanced scorecard (Chenhall R. H., 2003) but did not focus on the zoomed version of performance measurement systems. MCkeen and Guimares (1997)stated that contingency factors help to explain what is expected of the employee participation. Because industries differ in operations and management accounting, the performance measurement systems and dependent details also differ. This means that the diversification in production firms and service firms could be a moderator effect that explains the relationship with the types of performance measurements on job performance.
operational employees and its effect on incentive purposes and the job performance (Groen, Wilderom, & Wouters, 2015; Derfuss, 2009; Lau & Tan, 2006) while Burney and Matherly (2007) investigated the group-based performance measurement on the incentive purposes of managers by comparing this measurement with the individual effort of an employee. They also recommend in their research to investigate the contingency-based approach and link this to the use of
performance measurement systems (2007, p. 64). So this research is based on the effect of usage of performance indicators on employee job performance in different types of industries.
In addition Wageman (1995) has investigated whether task design has an effect on group functioning and whether the effectiveness of group-based performance indicators is related to the individual job performance of an employee. His report stated that groups performed best if the performance indicators were measured on individual-level or group-level. A plausible explanation for such effect is that the design of work and performance measurement systems determines the effort and motivation for work groups. Also, the literature about performance measures of individuals or groups gives attention to the evaluation of performance and to measure the performance of a firm or employee (Dumond, 1994). All kind of organizations are responsible for their operational activities and their organizational perspectives. For example, they have to serve internal control functions, internal professionals or suppliers, the determination of targets, and organizational goals. These so-called procurement areas of an organization measures the organizational performance, which includes department costs and productivity. Therefore, the chosen and implemented core measures determines and evaluates the achievement of goals and are used to motivate and improve an individual’s decision making. The motivation of
measurement is to improve employee job performance, providing feedback and give evaluation to individuals regarding to how well they are in line with the organizations’ goal. As a result, this will motivate the individuals in the adequacy of their job performance (1994).
The existing literature explains the development, fit and use of the performance measurement system, but it does not explain the relationship and variance between the type of performance indicators and type of industries and their effect on employee job performance (Skaerbaek & Tryggestad, 2010). This study aims to explore the relationship between the use of performance indicators on the employee job performance. The contribution for this study is by separating the kind of performance indicators, whereby I will make a distinction between the use of individual, group-based or hybrid (mix) performance indicators. Secondly, this study aims to explore the moderator effect of industry types between the use of performance indicators and employee job performance. To address these subjects, I use a sample size that consists of managerial and operational employee pairs.
This study consists of five sections. The relevant literature review is discussed in the next section, followed by the hypothesis development. Section three presents the research
methodology and explanation of the survey project of mw. dr. ir. B.A.C. Groen. This section gives insight to the project and hypotheses testing method. Afterwards, section four gives an overview of the analysis results and main findings. To sum up, section five gives the concluding discussion, as well as the limitations of the research and recommendation for future research.
2 Literature review and hypotheses
Given the motivation in the introduction, the following research question will be answered in this study:
Do the use of individual, group-based, and hybrid performance indicators have a positive influence on job performance in different industry types?
In the first part of this section I will give an overview of the different use of PMSs and performance measures, use of performance indicators for incentives purposes and what the meaning and link is to the group-based performance measurement. After the theoretical setting I will present my hypothesis in the second part of this section.
2.1 Performance Measurement Systems
Performance measurement is used to measure the job performance of employees (Groen, Wilderom, & Wouters, 2012, p. 44). As stated before in the introduction, performance management is also used to align the interests of employees or managers and is created to improve their working field. Performance of employees are settled with management controls, which is about making sure that an organization reaches its objectives. The performance measurement systems (PMSs) captures a broad approach to the management and the controls and performance of organizations. By including all the control aspects of an organization the systems become management control systems (Ferreira & Otley, 2009, p. 264). To obtain the effect of using the measures in a performance measurement system, it has to be developed within organizations. This means that in each organization design the performance measurement
systems differs and is not contingent. So, a contingency approach proposes that performance of a firm is related to the fit between different kind of factors, e.g., structure, culture, technology, strategy and employees. Contingency approach therefore focuses only on the relationship between context and organizational appearances (Deng & Smyth, 2013; Volberda, van der Weerdt, Verwaal, Stienstra, & Verdu, 2012). Notwithstanding the high presence of service organization, most research on management control systems and design is based on production firms (Auzair & Langfield-Smith, 2005, p. 400). According to Auzair and Langfield-Smith (2005, p. 401) pure service is characterized in intangibility of services, inseparability of production from consumption where the customers are also involved in the production of services. Also the service process type seems an important variable for the MCS studies, because these variables might be the determinants of the service industries. This means that within the service
Fitzgerald, Johnston and Voss (1992, p. 70) states that service types are divided into three archetypes:
1. Professional services: which are the organizations with few transactions and are mostly process orientated. Examples are management consultancies, accounting firms and corporate banking.
2. Mass services: which are the organizations that have many customer transactions but with little customization. Examples are newspaper retailers and transportation firms.
3. Service shop: which are organizations that are categorized between the professional and mass services. Examples are hotels and rental services.
They also conclude that that there is a difference within the three service types, because there is also difference in the management control systems and the performance measurement systems.
Another difference is also visible in manufacturing industries. This industry is separated into archetypes that explain and distinguish the key values and are clustered as follow (Lillis & van Veen-Dirks, 2008, p. 27):
1. Low-cost manufacturing/production: which means that firms are focused on efficiency and productivity. These firms also rely less on customer-focused measures than
differentiation firms. It is stated that low-cost firms will not invest in refined
measurement systems to track the performance on customer responsiveness and the rate of new product introduction.
2. Differentiation manufacturing/production: which is the opposite of the low-cost firms and focus more on customers and there is no place for cost controls.
So even if the overall category- manufacturing industries- are the same, with further investigation it is clear that there is certainly a difference in strategy and therefore also in performance
management systems between those archetypes (Dekker, Groot, & Schoute, 2013, p. 79). The differences between those archetypes also rely on the incentive systems, evaluation and also in result and action controls.
According to Peljhan and Tekavcic (2008, p. 179) the management starts to realize the importance of the PMSs when the strategic performance management will be implemented. This will lead to information about what is critical to the company’s success and what will be evaluated and rewarded by using the performance indicators. These indicators are typically used by
management to monitor the performance of the organizations and by taking new decisions. According to Groen, Wilderom and Wouters (2015) the use of performance measures implicates the extent to which managers find it important to use the measures for monetary compensation, non-monetary rewards, and evaluation purposes. They also state that if a company involves their
employees to compose performance indicators the participation and the quality of the
performance will increase. Afterwards, these composed performance indicators can also be used as a base for the evaluation of employees. A limitation is that not all industries use performance indicators with this purpose.
2.2. Individual/ Group Measures & job performance
Interdependence of employees in all kinds of organizations comes from the operational tasks. For example, Wageman (1995) has stated in his report that task inputs like distribution of skills and resources including the technology define the work. This latter means that individuals could work on composing products or whereby teambuilding is needed to build a whole product. Another important aspect is the that organizational goals are outlined and achieved. To measure if the strategy is achieved, is by simply looking at the individual-level performance measurement versus group-based and hybrid performance measurement. Because with this information a manager is able to reward the achievement of the goals. The following content will give an overview of the performance measurement:
Figure 1 Hypotheses structure
Construct short name Construct definition
Use of individual PI The use of individual performance indicators for employees used by their manager
Use of group PI The use of group-based performance indicators for employees by their manager
Use of hybrid PI The use of hybrid performance indicators for employees by their manager.
Industry type The employee works in a production firm, mass service firm or professional service firm.
Employee job
performance The extent to which employees meet their job requirements according to their manager. H2c H2b Use of PIs H2a Individual PIs Group PIs H1B Hybrid PIs Employee job performance H1A H1C Industry type
2.2.1 Individual performance measurement
Measuring the performance of employees has some relationship to culture. This actually means that “culture of high performance” shows what an organization looks like (Robson, 2005, p. 138). In such high performance culture employees tend to behave differently and behave to their own individual observation of situations. Also in such a culture employees feel that their tasks are helpful in performance improvement. In fact, employees perceive the feeling that they could affect or control the indicators of their performance. This latter means that if an
organization develops such perception the employees would be inspired to a high performance culture.
Measuring is a way of assessing a situation. For example, when the actions of an employee is not controlled by the manager, this could lead to a high performance or a low performance. Because when the work behavior of this employee is not measured there is no comparison available that gives the employee feedback about how well he/she has performed. To give insight into this aspect Robson (2005, p. 139) proposed two interconnecting systems. The first one is called the control system, which is focused on setting objectives, measure the performance and afterwards take the important actions in order to improve the performance when needed (Futrell, Swan, & Todd, 1976, p. 25) (Robson, 2005, p. 139). Secondly, Robson (2005, p. 140) states that the process systems are used by the employees to determine and asses themselves how well they will perform and where they want to be. By using these two systems the employees would perceive themselves as being in control of their performance.
Conversely, when the previous situation is switched from taking actions individually in mandatory performance, the employees feel they have to fulfill instructions of others- their managers or group- which means they are now responsible for the consequences of their actions. This latter is also applicable when they are not aware of those consequences. So the culture of an organization is determined by who is taking the actions and by who is controlling the
performance. As stated before, to achieve a high performance culture, the control system and the process system should be aligned to one person. When there is a separation between the two systems this could lead to inefficiencies in the performance of employees. So the behavior of the employee will change from being “in-control” of performance to “being controlled” by the manager or group (Robson, 2005, p. 142).
In situations whereby individual employees are being forced to be responsible and accountable for a specific task they are “being controlled”. This means there is misalignment between the two control systems. The accountability of these individual employees is being measured by individual performance indicators. According to Robson this will lead to
competition between the employees to achieve the best result of the department. The way there is no perfect performance culture in the organization, is not mentioned here.
There is a distinction between the managers and the work of its operational employees. This means that their working performance is designed to be measured on an individual-level, or it can be described as highly interdependent which means that the required input needs more than one employee to accomplish the task. It should be kept in mind that the required work is completely different for each investigated organization. A team based working group might include responsible tasks whose contribution is necessary for the end product or service (Wageman, 1995, p. 147). Therefore each responsible member is accountable for the quality of the result and also the strategy. On a more individual-level where the job is performed totally interdependent and the measurement- which is linked to a reward system- will be based on an individual scale. An example of such individual-level performance is the responsibility for sales in a specific working environment whereby the commission payment is also based on an individual sales performance (Futrell, Swan, & Todd, 1976). Specifically, if an employee is aware of the job task responsibility and also believes in his or her ability to perform a given task, this will lead to a higher commitment and increase in the job performance (Stajkovic & Luthans, 1998, p. 248). This is linked with the fact that an individual is also dependent on the goal-setting of the
department. Because this clarifies the performance expectations from the manager, so the actions of an employee has a direct performance outcome. In such situations, the employee has personal control over his or her performance behavior (Gibson, 2001, p. 792).
H1a: The use of individual performance indicators leads to a higher job performance.
2.2.2 Group-based performance measures
Another difference in the performance of an individual and a group is the fact that an individual performs differently. This is usually seen in situations whereby the manager is looking forward to meeting a favorable high performance, but the operational employees doesn’t feel alliance with the same goal. When this happens, there is a lack of motivation that determines the job performance. In situations whereby the manager is not focused on individual performance measures it is usually because of the group-level performance measures (Breugst, Patzelt, Shepherd, & Aguinis, 2012, p. 190). A positive effect of such a method on job performance is that individual employees think that they are being controlled and they are being put on actions. This means that group-based performance could have a positive effect on job performance of each individual (Robson, 2005; Wageman, 1995, p. 148).
Interdependence is also called an essential work characteristic and is equal to task interdependence and goal interdependence (Chen, Tang, & Wang, 2009, p. 626).This interdependence is also described as an association with the outcomes for individual group members that are affected by each other’s actions. In such highly interdependent situations employees are motivated to build close relationships with each other to complete tasks and achieve high job performance (Van der Vegt, Van de Vliert, & Oosterhof, 2003; Van der Vegt & Van de Vliert, 2000). If an organization uses group-based performance measures it will increase employees’ effort and job satisfaction. Conversely, this can also lead to a negative effect on the job performance, because groups are helping each other in situations where there is a negative information or a negative environment, which later turns into a decrease in own performance (Campion, Medsker, & Higgs, 1993, p. 824). The need for group-based performance measures are important in organizational settings, because these measures are providing feedback on job performance and the financial effectiveness. But according to Schmidt and Kleinbeck (1997, p. 304) it is almost impossible to make a distinction between the involvement of individual employees. The reason for such limitation is because the interaction of tasks among group members is high. The creation of procedures how such group based performance measures should be made is not found and is specific to each organizational department.
According to Chen, Tang and Wang (2009, p. 628) the lack of cooperation behavior among employees leads to failure of their tasks and collective operational goals. In addition, goal interdependence is influenced by collective goals and rewards for collective performance.
Previous studies (Saavedra, Earley, & Van Dyne, 1993; Comeau & Griffith, 2005) also concluded that the cooperation among employees- as group members- improves the group performance because they want to achieve collective goals. So when an individual employee ignores self-interest to achieve a cooperated goal, this will lead to cohesion and the completion of collective tasks. Individual employees without a collective goal are integrating each individual’s involvement and impact into collective efforts to achieve a positive group performance. Thus, individual employees that are working as a group and share a collective effort to achieve a collective job performance means that these individual members create a group cohesion to achieve higher performance.
Necessarily, work of group members in organizations has shown a high increase and is therefore an interesting investigation area for researchers to understand group member’s work dynamics (Cohen & Bailey, 1997; Campion, Papper, & Medsker, 1996). But individual employees in group context didn’t get much attention for investigating what the effectiveness is of their performance in such a group context (Shaw, Duffy, & Stark, 2000, p. 260). The latter leads to a
gap of knowledge, because the positive sides that indicate the individual satisfaction and
performance level in group context is related to the identification of individual employees. Cohen and baily (1997) state that group effectiveness is important to determine cooperation and positive job performance, but group performance measures could also have implication for the individual group member attitude and his or her behavior. Drawing from these early findings, I proposed the following hypothesis:
H1b: The use of group-based performance indicators leads to a lower job performance.
2.2.3 Hybrid performance measures
A combination of the interdependent and the dependent measurement design is called a “hybrid” design. This means that the group-based and individual-level performance measurement design enforces employees to work on their own project as an independent, but at the same time they collaborate on a shared enterprise. Such employees therefore operate completely
independent or as a team. Early studies concluded that group-based performance measures have more problematic aspects like group thinking, free-riding and social loafing. Also early studies were more focused on performance and management in organizations and its productivity of individuals rather than groups (Schmidt & Kleinbeck, 1997). Goals settings, incentives and feedback are often measured on individual level. But potential benefits of groups should be seen as productive units (Locke & Latham, 1990).
The combination of individual and group-based performance measurement could also have an effect on the behaviors when group members can and cannot interact with each other. Within a group, individuals affect the decision makings and the collective resulting outcome (Kocher & Sutter, 2005). According to Besedes, Deck, Quintanar, Sarangi, and Shor (2014, p. 297) individuals also exert effort when they need to make decisions as part of a group versus for themselves. They will be more accurate and think twice when performing a certain task. Possible reasons for such behavior comes from the responsibility feeling. In group settings where the individuals are only responsible for group outcomes, individuals may free ride or engage in social loafing. But when they are also responsible for their own performance, it will affect their social responsibility.
The social responsibility for individuals remains the same for hard tasks. However, if the purpose of such hybrid performance measure design is to avoid or reduce free riding it could be a costly method (Luhan, Kocher, & Sutter, 2009). In addition, the hybrid performance
created by the collaboration between individuals. Also, with the use of hybrid performance measures free riding by individuals will decrease because of their individual and social responsibility. Based on the early findings, I expect the following hypothesis:
H1c: The use of hybrid performance indicators leads to a higher job performance.
In all kind of organizations the operational method involves mass production of
standardized products or services (Banker, Potter, & Schroeder, 1993). Some organizations also moves towards flexible and interdependent production environments whereby the self-managed work-teams are working on customized products and the workers perform different tasks (Ellemers, Gilder, & Haslam, 2004). It is given that such teams are able to quickly take decisions and are appropriate to deal with global competition (Dunphy & Bryant, 1996). But Abernethy and Lillis (1995) stated in their research that in practice many firms also face difference in experience and the implementation of work settings. They think the main reason for this difference might be the lack of fit of performance measurement systems that rewards the team performance for coordination, motivation and work commitment. The use of incentive systems gives techniques to managers so they can assure the appropriate worker behaviors. The worker behavior is linked to the productivity and the organizational performance. So incentives systems have been used by managers to reward or punish specific behaviors (Chow, Kato, & Merchant, 1996). Nowadays, more cooperative structures are challenging incentive systems to promote coordination and interactive behaviors by the operational employees (Libby & Thome, 2009).
When looking at team-based structures, the literature on organizational behavior stated that group-based incentives can increase the cooperation of team members and also affects their motivation and effort. To obtain this commitment the employees should understand why they are incentivized to perform well. This means that the understanding of this system is mainly through the release of accurate information and suitable incentives. However, not all employees are willing to collaborate with the team and therefore put little effort in the team performance (Naranjo-Gill, Cuevas-Rodriguez, Lopez-Cabrales, & Sanchez, 2012). This means that their performance is measured on a collaborate scale rather than measuring their performance on an individual basis. In such a case, employees receive the same incentive as their group members. This is the so called agency theory that shows how group-based perceived incentives are likely to affect the group productivity.
According to Bonner, Hastie, Sprinkle and Young (2000)individual performance measures have a strong relationship between the compensation and the use of performance measures- which is called by Groen (2015) the extent to which managers find it important to use performance measures for incentive purposes. The reason is that these incentives purposes
depends on individual effort instead of team productivity. An additional effect of this occurrence is that when incentives are used after measuring the performance of individual workers, the performance appears to be higher than rewarding group-based performance, because the individual pay and his or her performance seems to be tighter (Bucklin & Dickinson, 2001).
Despite these facts, some studies show that individual workers perform better when the manager incentivize them based on their individual performance measures (Honeywell-Johson &
Dickinson, 1999), while others state that performance levels are equivalent (Stoneman & Dickinson, 1989).
2.3 Difference in industry types
The study of Groen, Wilderom and Wouters(2015) investigates the managers’ observation of the quality of operational performance measures after the participation of his operational employees by designing the performance measures. They define the performance measures as a quantitative expression of the operational employees’ work, which makes it important that the defined performance measures are of a suitable quality. When the manager finds the performance measures to be sensitive to the actions of his operational employees, the quality of these
measures verify their job performance. While Groen et al. (2015) investigates the participation of employees by developing the performance measures, this research is focused on the existing performance measurement system and the interactive use of these measures. The interactive use of performance measures refers to how managers and operational employees use an existing performance measurement system in their communication (Ferreira & Otley, 2009). An
involvement of operational employees by developing performance measures result in high quality of performance measures. This high performance measure quality shows that information
derived from these measures are valuable and therefore managers use these measures by their incentives purposes (Banker & Datar, 1989). But when combining the facts that performance measurement systems differ in each type of industry and some firms are already working with pre-defined performance measures, the quality of these measures are mainly observed by looking at the individual-level and group-level (Naranjo-Gill, Cuevas-Rodriguez, Lopez-Cabrales, & Sanchez, 2012). So, a combination of two investigations- one is looking at developing performance measures and the other one is looking at the relation between individual- and group-based performance measures- this research expects some difference in individual level and group-based existing performance measures and the extent to which difference in industry type gives a significant deviation in the job performance of employees.
Several authors have stated that the fit between strategy, management control systems and performance is very important, but it is not totally understood by organizations (Skaerbaek &
Tryggestad, 2010). Also, Ittner, Larcker and randall (2003) investigated in their report that the alignment between management control systems and firm strategy leads to a higher firm
performance. As result they concluded that in service firms there was no significant deviation in fit between a contingent management control system approach and the firm performance. In another investigation on the impact of alignment between strategy and types of management control systems (Govindarajan & Fisher, 1990, p. 280) found a significant result. The result of this investigation was that a cost leader gained higher performance when there is a fit between the management control system and resource sharing. For production organizations with product differentiation strategy the fit between behavior control and high resource sharing resulted in better performance. The essence of distinction between a service firm and a production organization relies on the fit between its strategy and chosen management control system. The performance measures of individual employees and group members could be a moderator in the explanation of such measure design on the job performance of employees. Drawing from these combined findings, I proposed the following hypotheses:
H2a: The use of individual performance indicators leads to a higher job performance in production firms, mass service firms, and professional service firms.
Koopmans, Bernaards, Hildebrandt, de Vet and van der Beek (2014, p. 230) states in their report that in professional service firms the individual performance indicators are often used as measure for the work quality instead of work quantity. An explanation for this is that literature captures an indicator that is often used for assessing task performance, like quantity of work. But Koopmans et al. (2014, p. 235) concluded that employees working in a professional service firm did not select work quantity as the most important performance indicator. Instead, they were more interested in the quality of their work and being result-oriented. According to Homburg, Art and Wieseke (2012, p. 59) service firms adapt performance indicators that are related to the organizations’ strategy. For example, these indicators looks at the achievement of customer satisfaction with service. Additionally, service firms integrate performance measures in a way that it can track the changes in employee- related performance measures, like employee qualification and employees’ satisfaction scores. So the measurement is not only looking at the customer perspective, but is also looking to the employees’ perspective. Secondly, the latter performance indicators could also have an effect on the financial performance outcomes. So, these
performance measures are looking at the overall action of a company and are therefore measured on individual and group-based level. If the organizations’ strategy fits with the performance measures, it will have a positive effect on the job performance of the employees. But despite the
fact that this research has not made any distinction in the right strategy per industry, I want to propose the following hypotheses:
H2b: The use of group-based performance indicators leads to a lower job performance in production firms, mass service firms, and professional service firms.
H2c: The use of hybrid performance indicators leads to a higher job in performance in production firms, mass service firms, and professional service firms.
3 Research methodology
To test the hypotheses I used two online surveys to collect the data. The first online survey is used for managers and the other one for his or her operational employees.
3.1 Data collection & respondents
This research design explains in what way the hypotheses will be tested. This testing will be employed by an online survey for both the managers and their operational employees of random firms. Firstly, I will join the survey project of mw. dr. ir. B.A.C. Groen which will be supported by the University of Amsterdam. To have access to the database a minimum of six respondents was required. With the help of my contacts I reached the minimum of six
respondents, which means I will have access to the database. The reason for joining this research project is the availability of data and the interesting research field. Because of the investigation of relationship between the manager and the operational employee, I think this project will bring much results to light. Also the survey research will be suitable to examine the effects of
difference in group or individual performance indicators used in industries on job performance. So this survey project provides much information that affects the theory about performance indicators and employee job performance.
The respondents in this study were kind of specific in a way that they had to be pairs of employees and managers. Also for this study the pairs of managers and employees should have met three criteria: (1) the manager and his operational employee must have worked in their current function for at least one year, (2) the employee has to be in the working floor/operational member, whereas the manager has to be in charge of these operational employees, (3) the
manager must use performance metrics to measure and evaluate the operational employee’s actual performance. Additionally, there was a fourth requirement hiding the details of the
investigation (Groen, Wouters, & Wilderom, 2012, p. 12; Groen, Wilderom, & Wouters, 2015, p. 7). So the participants were being told the overall subject, but did not know any specific
information. This way of doing deemed to reduce creating biases.
For reaching the minimum target, I asked my connections to help me with this research by participating the online survey. Groen et al. (2015, p. 7) called the collection of respondents a snowball sampling, which means that every potential participant was asked to join the project or might ask for contact details of other potential participants (Salganik & Heckathorn, 2004). My starting point was to ask my own network for respondents. If my connections were from a managerial background, I asked them to involve their operational employee to join this survey. If one of my connections was not from a managerial background, I asked them to involve their manager to participate. Unfortunately, some of my connections wanted to participate but could
not convince his or her manager to take place of this research. In some situations the respondents did not know what this research was about. But it was my task to inform the respondents about this research by explaining what performance measures are and why they are used. I also sent them a link to a site where they could find examples of performance measures.
As participant of Bianca Groen’s survey research, it was required to inform the contact details of the participants in a excel format. This file contains the name, phone number, email address, function and their relationship with the contactor. After this file was completed, a survey request was sent to all the participants. In this survey research the participants were not able to see their personal scores were. For this survey project Bianca Groen wanted to use a whole new dataset with the new obtained pair responses. But to realize this goal, it was necessary to have more than 100 responses. If this requirement could not be reached, the old dataset of Bianca Groen had to be combined with the new responses.
3.2 Variable measurement & Analysis
The dependent variable in this research is the job performance of employees. The measurement of this variable will be based on the prior research by Douglas Jr., Mitra, Grupta, Shaw (1998), Groen, Wouters, Wilderom (2012) and Moers (2006) and will be measured with the responses of the manager and the operational employee, which will conduct several questions based on a seven-point Likert scale: (1) Totally disagree; (2) Disagree; (3) Moderately disagree; (4) Neutral; (5) Moderately agree; (6) Agree; and (7) Totally agree. The survey is composed in Dutch and English and based on the background of the respondents, the survey was sent in their prefered language. According to the research of Bianca Groen, Wouters and Wilderom (2012, p. 17) employees completed the questions regarding the Use of PM and the performance indicators that are based on individual-level, group-level or both (hybrid). The managers gave answers regarding to the job performance of their operational employees and so did the employees about themselves. Also, Bianca et al. (2012, p. 17) stated that their survey design reduced common bias, because they had guidelines to reduce such common source bias explained by Podsakoff, MacKenzie and Lee (2003). Before the respondents were asked to fill in the survey they began with a brief
introduction to help the respondents understand what they are asked for. This method improves the quality of the data. A view of the survey is added in the appendix.
Moreover, the independent variable in this research is the type of usage of performance indicators. The independent variables exists of the use of individual, group-based and hybrid performance indicators. This new variable is created in the survey by firstly asking them what type of
have to fill in if their performance is measured on individual-level, group-level or hybrid-level (Strobach, Frensch, Soutschek, & Schubert, 2012).
Employee job performance was also used in the investigation of Groen, Wouters and
Wilderom (2012) whereas the well-known scale for in-role job performance was first developed by Williams and Anderson (1991) and revised by Podsakoff and Mackenzie (1989). This variable measures the extent to which managers perceive employees as meeting their job requirements. So this variable sees job performance from the perspective of the manager. Goen, Wouters and Wilderom (2012) also states that latter scale is important because increase in job performance means the performance metrics used are helping in stimulating the employees. According to Burney, Henle and Widener (2009) an important detail regarding to the measurement of job performance is that the scale correlates with objective measures of performance. If the objective performance measures itself would be used as scale, this could lead to incomparable performance measures. A large diversity in job is the biggest reason for the latter. Therefore, Groen et al. (2012) used a wide range of applicable scales to compare the employee job performance in jobs and industries.
The employee job performance was measured by Groen, Wilderom and Wouters (2015, p. 18) based on the scale for job performance. This variable measures the extent to which managers perceive employees as meeting their job requirements, but also the extent to which operational employees meeting their job requirements. So this variable could be used in general, but from the perspective of the operational employee. For this research this scale is important because the increase in use of performance indicators would stimulate the job performance (Williams and Anderson, 1991). This variable is measured by multiple questions and it is therefore required to look at the reliability of this variable. The coefficient of the reliability is therefore used to measure the scale which is comprised for a number of items. This is the Cronbach’s’ alpha of 0,870 for the seven question items for job performance.
Industry type is used to explain which category their organization belongs to. The manager and the operational employee will answer the question whether their organization can be
specified as a production organization, mass service firm or professional service firm. I use industry type as the moderating variable in this research. In my regression analysis this moderator will be taken as two dummy variables, because the response to which industry type the employee is working in will be measured as reference groups and as values 1, 2 and 3. For a more accurate result, I use the following control variables: gender, age, and job tenure of employees.
Table 1 Respondent characteristics
Characteristics Employees Managers
Gender 51% male 69% male 49% female 31% female Education
16% Low-level high school 15% Low-level high school
37% High-level high school 24% High-level high school
47% MSc or higher 61% MSc or higher Age Mean = 33 (SD = 9.5) Mean = 39 (SD = 9.6) Job tenure Mean = 5.4 (SD = 5.8) Mean = 6.9 (SD= 6,8) Size of the
organization (# staff) Mean = 2,366 (SD = 8,606 ) Same as for employees, as manager-employee pairs belong to the same organization
min 1, max 70,000
Industry type employee 16% production firms
28% mass service firms
55% professional service firms
3.3. Statistical analyses
The valid sample size of this survey project exists of 106 respondents in the first analysis. The composition of the sample size is the combination of one manager and one operational
employee, which are seen as one pair participating in the survey. All of the respondents/pairs are from organizations located in the Netherlands and use a performance measurement system in their lower hierarchical level (Groen, 2012). The data was analyzed with multiple regression and ANOVA. The reason for using two methods, is because the nature of my variables can be used by both methods. In both methods the same correlation matrix is expected to be produced for the control variables, independent variable, moderator, and dependent variable. The regression model is based on an equation that keeps the fundamental idea that an outcome for an employee can be predicted from a model. To test the reliability of employee job performance Cronbach’s alpha is used to show the value. Furthermore, to test the hypotheses with the regression model, some assumptions should be executed. The first one was the multicollinearity by looking at the VIF value <10 or <5. The individual variables and the moderators were recoded into dummy variables and have to be centralized before looking at the multicollinearity. The reason for centering is based on the interpretation of the outcome of a regression that can be facilitated by centering the original variables before calculating the interaction term. After centralizing, all test variables should show a VIF value between the 1 and 2, which is acceptable. The second step for the assumption was the normality of the residuals. To test this, scatter plots and histograms were executed and showing an acceptable normal distribution. Also, there were no outliners found.
The third step for testing assumptions is the homoscedasticity model, which is the Levene’s test with an insignificant p-value of 0.726. This means that the variances of the population are equal. And finally, the interdependency of the variables is measured with the Durbin-Watson scale. This value resulted in the descriptive statistics with a value of 1.636, which is between the 1.5 and 3 and therefore is acceptable.
Once the measurement model was adequate, the structural model was analyzed. I began with a model containing only the control variables (age, gender, working tenure) against the
dependent variable (employee job performance). This was not to test a hypothesis, but only to look whether there is some relationship between the chosen control variables and the dependent variable. Furthermore, I executed a second model by including the independent dummy variables (individual performance indicators, group-based performance indicators, hybrid performance indicators). This model was testing the first hypothesis, including control variables. The third model was the same as the second model, but the moderator dummies (production firms, mass service firms, professional service firms) were included against the dependent variable. The fourth model was testing the interactions between the independent variables and the moderators against the dependent variable.
4 Results
The first step to achieve the results for the hypotheses testing was the storing and coding of the data. Each chosen variable is coded into nominal, ordinal or scale values. Before the hypothesis testing, attention had to be paid to items in the questionnaire that measured the exact opposite of the chosen variable. But the questionnaires about the job performance of an
employee had no negatively worded items that had to be reversed. Furthermore, the missing values are derived from SPSS when running the preliminary analysis. For the employee job performance the number of respondents was 106, which means there were 4 missing values. When testing the regression analysis, the sample size became 102.
The reliability test is focused on the question “to what extent are the items of employee job performance supposed to measure a single construct”. This means that the items related to the variable should be aligned with each other. In this sample the reliability of the variables is tested on the employee job performance in SPSS. Because the employee job performance is the only scale variable that has 7 items and can be found in the Appendix. The measurement for the reliability is shown with the Cronbach’s alpha in table 2. Since the reliability analysis shows high levels for all the items of 0.870, it is not needed to remove variables to improve the Cronbach’s Alpha. The Cronbach’s alpha is only used for the dependent variable employee job performance and it is not used for the independent variable, moderator, and control variables. These variables are not measured on question items, so there is no need and also no possibility to test for their reliability.
In order to detect item retention, I used a principal component analysis with varimax rotation. Because I rely on self-reported measurement, I conserve only strong factor loadings above 0.5 on the construct. These components represent the factor loadings of the dependent variable and explain that each item (question) examines the highest loading to determine which factor affects the employee job performance the most. So, all items in the measurement model are loading on their intended factor and the factors are allowed to correlate with each other. Table 2 shows that 58% of the variance is explained by one factor and it is also supported by the factor loadings. In this analysis no questions were excluded from the model. Additionally, in table 2 the means, standard deviations, number of items in the scale and the reliability coefficients are presented. The means of the items conclude that the employee job performance is rated very high. This is also confirmed by the normality distribution of the dependent variable (employee job performance) given in the Appendix. I also looked at the Eigenvalues, which shows the total variance explained in job performance by one factor. The Eigenvalues of all items demonstrates values below 10, which is acceptable.
Table 2Descriptive statistics and factor loadings of the measurement model
Variables α N M SD Min Max Standardized factor loadings R²
Employee job performance 0.870 Item 1 106 5.73 1.047 1 7 0.845 0.579 Item 2 106 5.84 1.015 1 7 0.855 0.135 Item 3 106 5.91 0.900 1 7 0.801 0.097 Item 4 106 5.34 1.294 1 7 0.728 0.076 Item 5 106 5.63 1.132 1 7 0.768 0.054 Item 6 106 5.73 1.019 1 7 0.699 0.037 Item 7 106 5.57 1.219 1 7 0.599 0.022
Furthermore, the next step of the analysis is the dummy coding. The independent variable and the moderator are in this case categorical. The predictors that are included in the testing model have more than two categories. Therefore it is useful to convert the predictors to several variables each of which has two categories. I used the dummy coding for individual, group, and hybrid performance indicators, whereby the reference group is the hybrid performance indicators category. I did the same construct with the moderator variable, whereby the reference group is the professional firm. So in total, I had four dummy variables that could be placed in the linear regression model.
Table 3 represents the correlation matrix of all the variables. Consistent with my expectation, there is a weak negative association between the use of group-based PIs and employee job performance (r= -0.188; p<0.05). But the individual and hybrid PIs are not consistent with my expectation because of their weak negative association with employee job performance (r=-0.029; r=-,078). The control variable “job tenure” of employees are negative associated with job performance (r= -0.107;p<0.01). Furthermore, there is a strong negative association within the firms (p<0.01), but a weak negative association between professional service firms and job tenure (r=0.261; p<0.01). Hence, this table reveals the most important expectation that only the use of group-based performance measure is weak negative associated with employee job performance and the individual and hybrid PIs are not consistent with my expectation. Remarkable is that job tenure shows significance and negative associations with group-based and hybrid PIs, and production and professional service firms.
Table 3 Correlation matrix Variables 1 2 3 4 5 6 7 8 9 10 1. Job performance 0,026 0,062 -,163 0,001 0,0192 -0,112 -0,008 -0,079 0,079 0,80 0,53 0,10 0,99 0,053 ,261 0,936 0,428 ,433 2. Age ,031 ,634** -,009 0,19 ,139 -,261** 0,051 ,205* -,225* ,757 ,000 ,926 0,05 ,164 ,008 0,612 0,038 ,023 3. Job tenure -,107 ,637** -,001 0,04 0,17 -0,136 ,284** 0,074 -,270** ,281 ,000 ,992 0,69 0,09 ,174 0,004 0,462 ,006 4. Genderᴬ -,171 -,005 -,075 0,04 -,311** ,140 0,138 -0,043 -,059 ,082 ,961 ,456 0,67 0,001 ,161 0,165 0,668 ,554 5. Individual PIs -,029 ,160 ,097 ,049 -,210* -,825** -0,04 0,074 -0,038 ,764 ,103 ,325 ,622 0,03 0,00 0,67 0,462 0,708 6. Group PIs -,188* ,074 ,195* -,314** -,200* -,379** -0,03 0,103 -0,072 ,049 ,451 ,047 ,001 ,037 0,00 0,75 0,305 0,474 7. Hybrid PIs -,078 -,194* -,204* ,133 -,835** -,373** 0,06 -0,129 0,077 ,416 ,047 ,038 ,177 ,000 ,000 0,56 0,196 0,442 8. Production firms -,042 ,178 ,258** ,129 -,030 -,042 ,052 -,268** -,467** ,663 ,069 ,008 ,192 ,756 ,660 ,588 0,006 0,000 9. Mass service firms -,030 ,185 ,076 -,055 ,062 ,108 -,119 -,277 ** -,726** ,757 ,059 ,445 ,581 ,520 ,262 ,214 ,003 0,000 10. Professional service firms ,058 -,300 ** -,261** -,041 -,034 -,066 ,069 -,494** -,699** ,545 ,002 ,008 ,676 ,726 ,492 ,473 ,000 ,000
Notes: n= 102 (all variables together, including their missing values); **p<0.01; *p<0.05 p-values appear with italic; front; Pearson value are below the diagonal; non-parametric Spearman correlations appear above the diagonal; ᴬ1 = male, 0 = female
After the correlation table is presented, the assumptions to test the hypotheses are met. It seemed that the multiple linear regression analysis was the most suitable testing method, because it shows in steps (models) what the effect of different independent variables is on the dependent variable. In the previous chapter I explained that I will use the ANOVA and regression analysis, but I have chosen only to use the regression model. The reason for this choice is because of the model fit and ease. Also, when running the ANOVA and ANCOVA (including the control variables) the direct relationships and the moderation effect did not result in a significant value. Thus, there were no relationships between the variables using ANOVA. But after the multiple linear regression analysis and the interaction models some relationships seemed to show significance. Therefore I have chosen to test my hypotheses with the regression analysis.
To do so, the dummy coded independent variables are recoded into dummy variables because of their categorical nature. The multiple linear regression model was only applicable if the assumptions were met. For example, the multicollinearity and the normality of the residuals must show whether the variables are suitable for the testing model. These assumptions could
only be met when the variables are centralized. This means that the regression model used the centering option to cover multicollinearity between the predictor variables. After centering predictors, the mean of the independent variables and the moderator variables must be equal to zero. This assumption is met and does not show any multicollinearity because the VIF is below 5 for all the variables. Table 4 shows the model fit indices of R² of 0.040 (4%) in model 1, but it increases in model 4 to 0.144. (14.4%). This means that 14.4% of the variance in employee job performance is explained by the use of performance indicators in industry types (including control variables). The adjusted R² of the four models is divergent as compared to the R². This measure shows the loss of predictive power in the regression model. In this case the adjusted R² is improved in the fourth model, but it differs a lot from the R². A plausible explanation for this difference is because of the high variability of the dataset and the low number of the chosen control variables. Still, the multiple regression shows a higher R² and adjusted R² than the
ANOVA analysis. Thereafter, model 1 has an F-ratio of 1.345, model 2 is 1.159, model 3 is 1.399, and model 4 is 1.380. These F-ratios are the average variability in the data that the multiple regression model explains to the average variability unexplained by the same model. So, the F-ratio- in combination with the p-value- tests the fit of the multiple linear regression model and do not show any significance of the model. To improve the adjusted R², I ran a hierarchical
regression model in SPSS. I tried the stepwise and the remove method, but these two methods did still not improve the adjusted R².
Model 1 in table 4 contains only the control variables (age, gender, job tenure) which analyzes the direct relationships with employee job performance. The unstandardized B shows both positive and negative coefficients, but there is no significant value between the dependent variable and control variables (p>0.05). Model 2 of the regression model tests the independent variables and the control variables against the dependent variable. This model provides support for H1b, which means that the use of group-based performance indicators leads to a negative employee job performance. Model 2 also concludes that H1a, the use of individual performance indicators leads to a positive employee job performance, and H1c, the use of hybrid performance indicators leads to a higher employee job performance, are rejected. Model 3 is almost the same as model 2, but in this model the industry dummies are added in the effect on the employee job performance, including the control
variables. The use of group-based performance indicators in the table still turn out to be significant (p<.05), but the moderators do not show any significant values. So there is no relationship between the use of performance indicators on employee job performance in the three industry types. Model 4 tests the moderation effect between the independent variable and the dependent variable. However, this model does not show any significant effect in the
interaction terms. So, H2a,b,c are not supported, which means there is no relationship between the use of performance indicators and employee job performance in the three industry types. There is no appearance of an interaction between the variables, the interaction graph is shown below to support these insignificant results. There is no visible effect in the interaction graph and the differences for the industry types also differs from each other.
Table 4 Standardized regression weights and model fit of the structural model, with control variables
Variables Dependent variable Model 1 Model 2 Model 3 Model 4
Independent variables
Dummy Individual Employee job performance 0,025 0,037 0,041
Dummy Group Employee job performance 0,202* 0,219** 0,184**
Moderators
Dummy Production firms Employee job performance 0,022 0,021
Dummy Mass service firms Employee job performance -,142 -,157
Interactions
Ind x Prod Employee job performance -,167
Ind x Mass Employee job performance -,011
Group x Prod Employee job performance 0,175
Group x Mass Employee job performance 0,027
Control variables
Age employee Employee job performance 0,133 0,139 0,180 ,218
Job tenure Employee job performance -0,192 -,235 -0,259* -0,318**
Gender employee Employee job performance -0,150 -,091 -,098 -,117
Model fit indices
R² 0,040 0,073 0,094 0,144
Adjusted R² 0,100 0,025 0,027 0,040
F 1,345 1,519 1,399 1,380
p 0,264 0,161 0,215 0,166
Figure 1 The moderating effect of industry type (insignificant effect)
To sum up, the correlation matrix table resulted in a strong negative correlations between the use of individual, group-based and hybrid performance indicators. The correlation matrix shows in this case whether the variables are correlating with each other without predicting a causal link between the variables. The strong negative correlation values between the use of individual, group, and hybrid performance indicators implicates that a high value of one is correlated with a low value of the other. Finally, the regression analysis shows the exact relation with the variables by testing the hypotheses. In addition, the results above, the second, third and fourth models support H1b “the use of group-based performance indicators. The use of
individual or hybrid performance indicators does not show any relationship with employee job performance in the regression model, so H1a and H1c are rejected. Furthermore, the use of individual, group-based or hybrid performance indicators does not show any relationship with employee job performance in the industry types. Thus, H2a,b,c are not supported and therefore rejected in this regression model. Remarkable, the control variable “job tenure” seems to have a significant effect on job performance with the use of group-based performance indicators, but this is not the main research. To give more insight to this finding, an additional analysis test will be performed in the following subparagraph.
Additional Analysis
Since I discovered a significant effect of job tenure as control variable on the relationship between the use of performance indicators and employee job performance of the study, I
performed a moderation effect for the relationship between the use of performance indicators and employee job performance. The additional analysis is tested by running the multiple linear regression model. The main moderation effect- industry type- is not eliminated from this model, because of the main research subject. So I added job tenure as a second moderator variable. The assumptions for the regression analysis including job tenure, were consistent with the previous assumptions. Also, the subsample correlation matrix shows strong negative correlations with employee job performance. In order to facilitate the results of the regression analysis, I also centered the moderator variable before calculating the interaction term.
Table 5 represents the additional regression analysis. In the first model the job tenure is excluded as a control variable. Next, the job tenure is added in the third model to analyze the direct relationship with employee job performance. The fourth model is testing the moderation effect as an interaction term between the use of performance indicators and employee job performance. The results of this analysis are consistent with the findings of Ali and Davies (2003). Their analysis of job tenure on employee job performance resulted in a positive effect, but when testing the interaction between tenure and job performance, it turned out to be an insignificant result.
Table 5 Regression results testing effect moderation of job tenure with employee job performance
Variables Dependent variable Model 1 Model 2 Model 3 Model 4
Independent variables
Dummy Individual Employee job performance 0,019 0,063 0,07
Dummy Group Employee job performance 0,457 0,603* 0,45
Moderators
Dummy Production firms Employee job performance 0,049 0,021
Dummy Mass service firms Employee job performance -0,244 -0,157
Job tenure Employee job performance -0,035* -0,039
Interactions
Ind x Prod Employee job performance -0,759
Ind x Mass Employee job performance -0,026
Group x Prod Employee job performance 0,175
Group x Mass Employee job performance 0,803
Tenure x Ind Employee job performance -0,017
Tenure x Group Employee job performance 0,009
Control variables
Age employee Employee job performance 0,001 -0,001 0,015 0,018
Gender employeeᴬ Employee job performance -0,21 -0,13 -0,153 -0,179
Model fit indices
R² 0,018 0,042 0,094 0,149 Adjusted R² 0,002 0,003 0,027 0,023 F 0,915 1,067 1,399 1,184 p 0,404 0,377 0,215 0,304
Notes: Unstandardized regression coefficients are shown, n=102; **p<0.01; *p<0.05; ᴬ1 = male, 0 = female Model 1 in table 5 contains two insignificant control variables. Model 2 shows that there is no significant effect between the use of group-based performance indicators, while it was significant in the main regression model (p<0.05). The use of group-based performance indicator and the job tenure of employees are significant (p<0.05) in model 3. This model is testing the direct relationship of job tenure on employee job performance and shows a significant effect. This means that when adding job tenure as an independent variable, the two variables become important. In model 4 the moderator effect of job tenure and employee job performance was tested as an interaction term, but this test result was not significant, therefore no relationship. Despite this result, adding the job tenure as a moderator effect in model 4 improves the
explained variance of employee job performance (R²= 14.9%). So, the additional analysis results in a much lower adjusted R² than the main regression model in table 4. This variance between the two regression models is due to decrease of control variables. There were three control variables in the main regression model and after the additional analysis, the amount reduced to two control variables. Which means the two control variables in table 5 do not show a high
relationship with the dependent variable. Hence, there is no effect in the additional analysis that support a moderator effect of job tenure on the relationship between the use of performance indicators and employee job performance.
5. Concluding Discussion
Do the use of individual, group-based and hybrid performance indicators have a positive influence on job performance in different industry types? This study seems to have variance in answering the research question, but the main response to this question is “no”. However, for a small part of the research (use of group-based performance indicators), the relationship is direct and significant with employee job performance. Nonetheless, I found no relationship between the individual and hybrid use of performance indicators on job performance. The use of performance indicators in three kind of industry types seemed to be challenging because these measures are functioning as leading and managing the activities of employees. Therefore, a moderator effect of industry types on employee job performance is analyzed, but did not show any relationship with employee job performance. So, the use of group-based performance indicators seems to lead to a negative employee job performance, while the use of individual or hybrid performance indicators do not show any effect at all. Secondly, the industry types (production firms, mass service firms and professional service firms) which the respondents are working in, do not have any moderator effect between the use of performance indicators and employee job performance. The analysis of the research question has not contributed to a new scientific finding. Job tenure was added as a control variable, but seemed to have a relationship with the employee job performance.
Even though, past research has shown that the use of group-based performance measures could contribute to a higher job performance of employees (Van der Vegt & Van de Vliert, 2000), the current study found a negative significant effect of group-based performance measures. This is consistent with a report (Campion, Medsker, & Higgs, 1993, p. 824) that has stated the use of group-based performance measures could face serious problems like group thinking, free riding and social loafing. According to Schnake (1991, p. 42) the social loafing effect indicates that individuals have the ability to perform well but choose not to do so because they believe in their indirect contribution that affect others to perform better. So this is
consistent with the results that group-based performance measures lead to a lower employee job performance. But this result has no meaning in the industry types.
Gibson (2001) provides an review of research linking effort to job performance. He concluded that the performance of an individual has a direct performance outcome. The reason for this is due the dependency on the goal setting of the organization which leads to a personal control over the individual’s performance behavior. Futrell, swan and Todd (1976) also confirm the latter, because they believe that the ability to perform a given task will encourage the
