The effect of employee performance measures participation on reported employee job performance

Amsterdam)Business)School)

The effect of employee performance measures participation on

reported employee job performance

Name: Linrui Xiang Student number: 11025514 Thesis supervisor: Qi Yang Date: 15 August 2016 Word count: 12682

MSc Accountancy & Control, specialization Control

2

Statement of Originality

This document is written by student Linrui Xiang 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

I measure employee participation in designing performance measures and performance indicators in two aspects: how he could be heard and how he could influence performance indicators. I focus on lower level of managers that make incentives use of performance measures on their sub ordinate employees. When organization has a flexible culture to let the employees involved in designing performance indicators, the more managers attach importance to incentive use of performance measures, the more influence employees could have on choice of data and ongoing modification. When organization has a flexible culture to let the employees involved in designing performance indicators, the less managers attach importance to incentive use of performance measures, the more influence employees could have on implementation and maintenance of performance indicators. Using these high quality performance measures for incentive purpose, quality of performance measures and employee performance measures participation, will formulate a closed loop. These three variables will have combined influence on employee job performance.

Keywords:

Performance measures; performance indicators; performance measures participation; self reported performance

4 Contents

1! Introduction ... 6!

2! Theory and hypothesis ... 10!

2.1! Employee performance measures participation ... 10!

2.2! Performance measures quality ... 11!

2.3! Employee performance measures participation and job performance ... 13!

2.4! Performance measures quality and norm ... 13!

2.5! Control variable ... 15!

2.6! Variables and relations ... 17!

2.7! Hypothesis ... 19! 3! Methodology ... 21! 3.1! Participants ... 21! 3.2! Data collection ... 21! 3.3! Survey instrument ... 22! 3.4! Variable construct ... 25! 3.5! Research model ... 31! 4! Result ... 37! 4.1! Correlations ... 37! 4.2! Hypothesis testing ... 41!

4.3! Moderation of enabling control environment ... 41!

5! Conclusion and discussion ... 42!

5.1! Practical implications ... 42!

5.2! Limitations ... 42!

References ... 43!

6

1) Introduction

In this paper, I examine the effect of using performance measures for incentive purpose on the co-development of performance measures. There is a great many of literature on the participation of employees in management accounting system design (Eldenburg, Soderstrom, Willis, & Wud, 2010). It is an innovative topic in management control research. However, most studies focus on participation in setting unit’s budgets, specially on determining unit’s budget by line managers. Derfuss (2009) finds the negative influence on budgeting efficiency if more participation of employees is involved. Although these studies provide valuable insights into co-development of performance measures, none of these studies connect using performance measures for incentive purpose with co-development of performance measures. This omission probably comes from the stereotyped thinking based on following three findings. Robert (2002) find that participation in designing performance measures could lead to better quality performance measures. Moer (2006) investigates that good incentive measures could help improve performance of employees in short term. Spector (2006) convince that using performance measures for incentive use could positively affect the performance of employees. Employees participations in performance measures design could improve performance measures quality, moreover, high quality performance measures could enhance the incentive use of performance measures by managers and improve performance of employees. It is neglected that more incentive use of high quality performance measures could also attract or encourage employees to participate in designing performance measures. Three independent variables influence each other and form a closed loop in one direction. Employees participation in designing performance measures, performance measures quality and using performance measures for incentive purpose all separately contribute to better performance of employees.

Empirical studies that investigate relationship between performance measurement system and organizational performance are abundant (Grafton, Lillis, & Widener, 2010; Lee & Yang, 2010), but studies that investigate relationship between quality of performance measures and using performance measures for incentives are rare. Groen (2015) finds no evidence to support the hypothesis on using performance measures for monetary compensation to increase employee performance. Other study investigates that monetary incentives could negatively affect employee job performance. Using performance measures for incentive purpose includes using performance measures for both monetary compensation and nonmonetary rewards. Monetary compensation includes potential salary increase and potential bonus. Nonmonetary reward

7 includes chance of promotion and authority within organization. I examine both monetary compensation and nonmonetary rewards to discuss potential implication in management accounting research, specialized on designing performance measures through participation of employees.

In this paper, designing performance measures through participation of employees refers to performance measures participation. It means that employees would participate in designing performance measures indicators rather than simply participating in budgeting and goal setting. Budgeting and goal setting are management accounting techniques within organizations. Their aims are to set goals and determine how difficult to achieve the goals. Performance measures participation has a broaden definition, that employees should participate and be responsible for transferring goals to good performance (Bourne, Mills, Wilcox, Neely, & Platts, 2000). This suggests that employees should have influence on total performance measures development process: design, implementations and maintenance. Employees influence in this process is the beginning of managerial control over developing sustainable performance measures (Gravesteijn, Evers, Wilderom, & Molenveld, 2011).

I measure employee participation in designing performance measures and performance indicators in two aspects: how he could be heard and how he could influence performance indicators. My research based on survey designed by Bianca A.C. Groen. The survey focus on lower level of managers that make incentives use of performance measures on their sub ordinate employees. The employees should be professional or at the lowest hierarchical level of organization and directly contact the frontline. Surveys are designed in three areas: how employees could be heard; how employees could influence performance indicators; how managers use performance measures for incentive purpose. Pairs of managers and employees should answer same questions separately. Same questions answered by managers and employees are two different variables. Answers from managers about how employees could be heard can be mediator variable to select organizations that has enabling environment for employees.

This paper contributes to literature in following ways. First, the result provides an explanation for why previous research could not connect monetary compensation and nonmonetary rewards to employee participation in designing performance measures, since these studies ignore their

8 mutual improvement caused by “reasoned action model” (Ajzen, 2012). Previous study focus on linking performance measures participations to employee job performance; linking performance measures quality to using performance measures for incentive use to enhance job performance. Second, it initiates more and more management accounting system designers to take low level performance measures users into consideration. Implementation of using performance measures for incentive use could help encourage employees better understand that designing performance measures is beneficial to themselves and organizations. Last but not the least, it has practical contribution to organisation on improving employee job performance by increasing their involvement in performance measures system.

The data I use in this study based on survey built by Bianca A.C. Groen. The survey contains 20 questions with topics such as performance measures participation, employee capability to take initiative, incentive use of performance measures and delegation rights. The survey has two specific versions to employees and managers and participants should be in pair meeting strict requirements. Each topic of question has more than one sub question. For each sub question, it has a 7 points scale for participants to select from totally disagree to totally agree. Questions are all subjective and each participant was given a dedicated survey link on the internet. Subjective questionnaires are designed according to previously research from Abernethy, M.A., & Bouwens, J. (2005). Answers of 7 points scale initiated from Francis et al. (2004). Basic information about participants such as sex, industry and working years are also included. They could address further discussions on performance measures participations within this survey.

In this paper, I control the managers answer to performance measures participation. Whether organization has a flexible culture and management control process to let the employees involved is critical to determine whether employees has real influence on performance measure system design rather than employees excessively speculate that they have influence on designing process. Based on this setting, I find that when organization has a flexible culture to let the employees involved in designing performance indicators, the more managers attach importance to incentive use of performance measures, the more influence employees could have on choice of data and ongoing modification. When organization has a flexible culture to let the employees involved in designing performance indicators, the less managers attach importance to incentive use of performance measures, the more influence employees could have on implementation and maintenance of performance indicators. This is consistent with previous study that autonomous

9 motivation of employees will largely influence employees on how significant others, who are admirable to employees, want employees to to perform better (Hagger & Chatzisarantis, 2009). When line managers attach importance to use performance measures for incentive use, employees are more willing to participate in the performance measures designing process if they could be heard in organization.

The main research question of this paper is to figure out the relationship between how manager use performance measures and how employee could participate in performance measures design. Whether the relationship will differ if the organization is flexible and practical to let employees involved in management control process. In the following sections, I will present the management accounting theory and cognitive psychology theory that related to my research question. The research method will be detailed explained in section three, with more practical verification. The statistics results will be presented in section four and conclusions and limitations will be discussed in section five.

10

2) Theory and hypothesis

In this section, the hypothesis will be formed based on previous management accounting research. First, the variables and their relations will be presented. Second, the variables and existing findings will be discussed. Third, the mediator variable, innovation of this paper, will be described. The hypothesis will be developed at last.

2.1! Employee performance measures participation

Performance measures are designed to measure employee job performance. As for organization, performance measures are quantifiable indicators aimed to assess performance and position of its business. Managers may routinely check employee performance through pre determined goals according to performance measures. Types of performance measures different depends on who the performance measures will assess. This study will focus on performance measures that assess non managerial employee rather than managerial employee. There are more and more literatures quantitatively link efficiency and effectiveness of performance measures to employees that was assessed. Performance measures could motivate and control managers and employees by affecting their decision making and authority level in organization (Sprinkle, 2003). On the contrast, Wouters and Wilderom (2008) argue that performance measures could help enhance managers and employees understanding of organization objectives and improve employees rather than control them strictly to act as performance measures required. Performance measures have its name and purpose at least. As the maturity of developing performance measures, source of original data, calculation formula, accountability and frequency of assessment are also required (Neely et al., 2002).

Employee participation in performance measures has a meaning of letting employees be involved in performance measures designing process. Moreover, it has a broaden meaning of being sustainable with performance measures system. Employee participation in performance measures refer to let employee participate in designing, implementation and maintenance (cf. Bourne et al., 2000). Employees are involved in performance measures setting process, but line managers could participate as well. Managers have significant influence in guiding employees to understand performance measures system process and effectiveness of performance measures development (Groen, Evers, et al., 2011).

11 Abernethy and Bouwens (2005) explored employee performance measures participation with “influence on the system design” scale. They address the extent of influence that employees thought they could have on designing performance measures that will be used to assess themselves. They focus on employee cognitive recognition of their role and responsibility in employee performance measures participation. Employees will have opinions on both questions: how employees could be heard; how employees could influence performance indicators. Their answers to these questions could be original data of employee performance measures participation. This procedure could also reflect their routine performance assessment and their co development with their line managers. Employees will be more willing to accept the assessment result if they are discussed about performance indicators with their managers (Wouters, 2009). Employees reporting about assessment of such performance indicators are also beneficial in developing sustainable performance measures.

2.2! Performance measures quality

Performance measures quality refer to how employees link their actions to performance measures. Whether the performance measures are relevant to reflect their job performance and accomplishment of goals; whether the result of assessment could be verified and gain the same result if performance the evaluating process again (Moer, 2006). There are lots of literature that studies performance measures quality on behalf of managers. They assert that high quality of performance measures should be a tool to help managers control and to supervise job performance of employees. They also address that good performance measures should guide employee find solutions to get better performance and get close to operational budgetary goals. However, employees need precise performance measures that based on convincing designing process with experience (Keeping & Levy, 2000). They prefer performance measures that could well describe their accountability and what they are expected to give out to organization while their daily work (Hall, 2008).

Performance measures quality explain the extent that employees make performance measures precise in measuring their performance in their relevant working criteria. When employees are convincing that performance measures could perfectly reflect what they have done, they will believe in the correctness of assessment that attach importance on evaluating their job performance. They will be motivated to improve their job performance and get close to

12 determined goals. Targets that pre determined by performance measures could be reached easily because employees are willing to put more efforts to think about solutions to get the result (Fried & Ferris, 1987). If employees get good response from their managers about achieving goals set by performance measures, they will be happy when discussing next period goals set by performance measures. The more accurate of the response from their managers about achieving goals set by performance measures, the more employees will feel positive effects from the result of assessment (Kinicki, Prussia et al., 2004). Employees will trust the fairness of high quality performance measure system and feel motivated by each goals set by high quality performance measure system.

Formulating high quality performance measures should design in employees eyes rather than in managers strong control freak. However, high quality performance measures will have a specific feature, that is control. Employees should feel different level of possibility to achieve different job performance or job requirements. Performance measures could give a reference system to employees on average and reasonable efforts that should put on to meet job requirement. Setting a high quality goal within high quality performance measures system could affect employee decision making on allocating efforts and skills (Hall, 2008). Employees will also be motivated to get new skills according to learning curve if they desire to make better decisions and get closer to goals next time. In this way, employees will align strategy of organisations with objective of their job requirements and perform better. This is the nature of motivations and could reduce agency problems if managerial employees are evaluated with high quality performance measures system.

High quality performance measure could give out favourable indications if employees meet all job requirements and take actions to achieve. This suggests that evaluation of job performance will not be affected by factors that out of employees and managers control. This increase the verifiability and fairness of performance measure system. Both managers and employees will be motivated by high quality performance measure. If managers act according to procedures that determine employee goals of performance, employees will be motivated to achieve the goal. If managers attach more importance to achieve the organization goals, employees may desire to measure their performance by themselves according to high quality performance measures. In this way, employees achieved goal may not in the same directions of organization goals. This could increase internal cost and decrease precision.

13 When managers discuss performance measures designing and implementation with employees, they will put their own attitude and experience into discussion. Employees will also think about setting goals on behalf of maximising his own benefits. Although employees have practical skills and may be more professional on skills than his line manger, they may reveal private information or hide key performance indicators to make him easily achieve the goals set by performance measure system (Christensen, 1982; Penno, 1984). Employees could only decide content of performance indicators that his manager do not have enough skills and knowledge about. However, Brown, Evans and Moser (2009) denied the agency theory effects in employee participation in performance measure designing process. Empirical study illustrates that involvement in designing performance measures that will use to assess employees themselves will show great honesty of managers and flexible culture of organization. In this way, employees are spontaneous to show same honesty to his managers in turn. Employees share practical knowledge and convince that he will be beneficial from this performance measures. High quality performance measure will bring positive outcomes to both managers and employees in this way.

2.3! Employee performance measures participation and job performance

Explore the relationship between employee performance measure participation and employee job performance is important because they throw light on involving employees in managerial control by developing performance measures and performance indicators together. Groen (2015) argues that only when employee performance measure participation could enhance the quality of performance measures, can their participation contribute to employee better performance. Managers could use co-developed performance measures for evaluating purpose to increase employee job performance significantly. However, as for using co-developed performance measures for incentive purpose, there remains further investigations on relevant and reliable database. Using co-developed performance measures for control purpose could also improve employee job performance, depending on the degree of control perceptions. Both employees and managers could get benefit from using the co-developed performance measures.

2.4! Performance measures quality and norm

Quality of performance measures will be improved if performance measures system follows strict phrase to build up: design, implementation and maintenance. In this way, high quality performance measures could also help facilitate employees to perform well according to pre

14 determined job performance requirement. Managers will be motivated to discuss performance measures with employees to increase trust between managers and employees. Employees will gain self satisfactions and intrinsic motives when they feel that they could be heard in organization flexible culture (Hagger & Chatzisarantis, 2009).

Performance measures that improved by participation of employees and the adorable managers of employees could facilitate employees working passion. Employees will feel that better performance and reasonable performance assessments should come from better communications with his managers. During the participation and discussions, employees will get better understanding about what their managers focus on and what organization requires. They are expected to learn more about the implementation process and job required knowledge (Hall, 2008). Injunctive norms require employees and managers understand what should be done in order to get acceptable behavior. Descriptive norm means that managers should accept employees behaviour that not applied with performance measures system (Fishbein & Ajzen, 2010).

Quality of performance measures and norm could be explained by the theory of planned behaviour. The theory of planned behaviour address that people will behave according to their attitude and perceived norm. Employees will make decisions after feeling the social normative pressure and relevant other’s belief. They will form a habit on deciding whether one thing should be performed or not. In situation of participation of performance measures system, the relevant other’s belief equals what their managers expected from employees. This could be applied to determine performance measures using purpose. The other important item in the theory of planned behaviour is the subjective norm. The more important the managers in employee heart, the more judgement of behaviour towards performance measures design will be affected by his manager.

Performance measures will be used for evaluate and control purpose if the quality of performance measures are highly linked to subjective norm. Manager want to evaluate whether employees are doing the right thing according to budgetary goals. Management control depends on measurement and management accounting is also a kind of control. Managers will check whether employees have in fact take actions according to job performance requirement. In this

15 way, managers will rely on performance measures to set fair and sustainable standards. Managers will find specified solutions and spend time and money in the specified ways to improve the performance of employees and their participation in performance measures.

2.5! Control variable

Information asymmetry arises when employees have more knowledge related to their work rather than their managers. When there is information asymmetry, the performance measure system gets meanings because managers need to make sure that employees are acting on behalf of organization interest. If there is no information asymmetry, managers will not give monetary compensation and non monetary rewards according to performance measures. Greon (2015) asserts that when there is more information asymmetry, using performance measures for incentive purpose will decrease employee job performance. Monetary compensation and non monetary rewards will motivate employees when there is less information asymmetry and employees believe that performance measures could correctly reflect what they have done in their routines. Managers use performance measures to check employees work and give out incentives. If employee feel his manager do not have enough professional knowledge related to employee front line work, judge of monetary compensation and non monetary rewards made by managers could not convince employees. This effect is more obvious in manufacture company rather than in service company because manufacture company is supposed to have less information asymmetry (Stajkovic & Luthans,1997). Employees will treat job complexity as a means to expect information asymmetry between employees and their managers. Line employees are supposed to have more professional and practical knowledge about the job rather than their managers supposed to. Employees will choose to not perform well to save cost of being professional rather than perform well to get the monetary compensation and non monetary rewards that are assumed to be unmatched.

Whether organization has a flexible culture and management control process to let the employees involved is critical to determine whether employees has real influence on performance measure system design rather than employees excessively speculate that they have influence on designing process. This could be treated as the information asymmetry between organization objective and every employees of the organization. Employees could only be heard in enabling control system (Jordan & Messner, 2012). Formal control system will be positively accepted by

16 managers that prefer to use coercive standards to assess employee work. In this way, formal control system will act as a negative role to coerce managers purpose on using performance measures. How management control system is designed and how performance measures implement will decide whether the management control system is enabling or coercive. Enabling environment within organization management control system has fowling features. These features are all aiming to make employees have same attention from organization as their managers have. First, employees that are involved in management control should have right to repair the control system if there are some emergencies. Both employees and managers are equal users that are involved in the control system. They are permitted to modify existent performance indicators and performance measures to keep update to changing environment (Wouters & Wilderom, 2008). Second, enabling control system requires high quality of transparency within internal control system. Both managers and employees should be heard and be aware the communication hierarchy within organization. This suggests that targets set according to performance measures and assessment according to performance measures should echo each other. Both input and output of the management control system could be performed again and give out the same result as before. Third, flexible environment should be highlighted in organization culture. Managers are possible to consider whether their work could be done according to formal system. Employees are supposed to help managers get better understanding of employee self development requirement. More flexibility could make managers subjectively enable formal system to facilitate their managerial control.

Control is a tool that could help link managers evaluating role to employee action role. It does not matter whether managers accept enabling control system and apply flexibility to managerial work. Employees could give out their voice through well designed enabling control system. New practice and relevant control improvement will be integrated into management control system and managers will be affected by this relevant updates. Evaluation process could also be modified within such continuously self updates. Performance indicators could act as an enabling control tool if employees are accepted in participation of performance measures system design.

Flexible use and continuously modification use of performance indicators should be deeply developed in top managers heart. Top managers could order managers to focus on specified performance indicators and report routine work. When managers feel difficult to add flexibility and modification into managerial control system, they are likely to maximise the coercive role of

17 managerial control system and relate their uselessness within organization to incompleteness of managerial control system (Jordan & Messner, 2012). Both manager and employees should feel flexibility and modification possibility within managerial control system.

Attitude of direct managers could represent the level and acceptance of enabling managerial control system within organization. Attitude of direct managers towards whether employees have influence on performance indicators could be mediator variable in research. Managers are familiar with performance measures setting process and the situations of employees in organization. Appearance of employee participation could be any one of the following situations. Whether employees could give out opinions without being asked. Whether employees were asked to give out personal opinion. Whether employees gave out opinions in group meetings or privately. Whether employees could design performance indicators by themselves. Different situations will make employees have different level of desire to participate in performance measures design, implementation and maintenance. To what extent the managers using performance measures for incentive purpose could affect employee participation in performance measures will also be different in these different situations.

2.6! Variables and relations

This study will focus on the relationship between employee performance measures participation and manager using performance measures for incentive purpose. Performance measures quality will be improved if employee participated in designing performance indicators. Quality of performance measures is measured in three properties: sensitivity, verifiability and precision. If quality of performance measures is high, low level manager will use these high quality performance measures for incentive purpose more than before. This will increase employee authority in organization. Question address: will using these high quality performance measures for incentive purpose increase more employee performance measures participation in turn?

Independent variable is using these high quality performance measures for incentive purpose and dependent variable is employee performance measures participation. This relationship is examined by dividing both independent variable and dependent variable into detailed variable. The independent variable is divided into managers attach importance to using performance

18 measures for monetary compensation and for non monetary rewards. The first includes potential salary increase and potential bonus. The latter includes potential promotion and potential authority in organization. The dependent variable is divided into the way and the content, i.e. how employees could be heard; how employees could influence performance indicators. These variables are analysed from employee aspects. However, to control the situations that employees overestimate their influence or employees collude with manager, same questions answered by managers and employees are two different variables. Answers from managers about how employees could be heard can be mediator variable to select organisations that has enabling environment for employees.

The dependent variable is expected to have positive or negative influence on independent variable according to different answers from managers about how employees could be heard. The two variables, using these high quality performance measures for incentive purpose and employee performance measures participation, are positive related to employee job performance. Therefore, any changes of these two variables contribute to changes of employee job performance, which has practice contribution to managerial control of organization. Moreover, if these two variable has significant relationship, the three variables, using these high quality performance measures for incentive purpose, quality of performance measures and employee performance measures participation, will formulate a closed loop. These three variables will have combined influence on employee job performance. Linkage effects will be strengthened if one of the three variables is changed. Independent variable, dependent variable and mediator variable are demonstrated in Figure 2.1.

19

2.7! Hypothesis

Managers will be motivated to discuss performance measures with employees to increase trust between managers and employees. Employees will gain self satisfactions and intrinsic motives when they feel that they could be heard in organization flexible culture (Hagger & Chatzisarantis, 2009). The more important the managers in employee heart, the more judgement of behaviour towards performance measures design will be affected by his manager. Employees will also be motivated to get new skills according to learning curve if they desire to make better decisions and get closer to goals next time. In this way, employees will align strategy of organisations with objective of their job requirements and perform better. This is the nature of motivations and could reduce agency problems if managerial employees are evaluated with high quality performance measures system. As for employee performance measure participation, employees will feel fairness and trusts, which are intrinsic motives. Employees will also feel extrinsic motives, if their participation brings high quality performance measures and give reasonable financial or non financial rewards to themselves. This research will select participants that pair of manager and employee have same

Hypothesis 1: Employees participation in performance indicators is positively related to manager reported performance.

Hypothesis 2a: Employees participation in performance indicators is positively related to manager reported performance, when he could give performance indicators opinion without being asked.

Hypothesis 2b: Employees participation in performance indicators is positively related to manager reported performance, when he could be asked to give performance indicators opinion.

Hypothesis 2c: Employees participation in performance indicators is positively related to manager reported performance, when he could give performance indicators opinion in a group meeting.

Hypothesis 2d: Employees participation in performance indicators is positively related to manager reported performance, when he could design the indicators together with others.

20

Hypothesis 2e: Employees participation in performance indicators is positively related to manager reported performance, when he could design the indicators by himself.

The main research question of this paper is to figure out the relationship between how manager use performance measures and how employee could participate in performance measures design. Whether the relationship will differ if the organization is flexible and practical to let employees involved in management control process.

Hypothesis 2a: When employees could give out opinions without being asked, using performance measures for incentive purpose is positively related to employee influence in participation of performance measure design.

Hypothesis 2b: When employees could be asked to give out opinions, using performance measures for incentive purpose is positively related to employee influence in participation of performance measure design.

Hypothesis 2c: When employees could give out opinions in group meeting, using performance measures for incentive purpose is positively related to employee influence in participation of performance measure design.

Hypothesis 2d: When employees could design performance indicators with others, using performance measures for incentive purpose is positively related to employee influence in participation of performance measure design.

Hypothesis 2e: When employees could design performance indicators by themselves, using performance measures for incentive purpose is positively related to employee influence in participation of performance measure design.

21

3) Methodology

To test the hypothesis, I join the survey project provided by Bianca A.C. Groen. Her topic is using performance measures to help instead of to control employees. First, pairs of employees and their managers who answered several questions related to performance indicators from their perspectives. Do survey on pairs will reduce the risk that identify wrong possibility of involvement from one point of view subjectively. Second, the pairs must meet criteria including “the employees have to carry out operational activities of their organisation”, which is the same operational level as I want to investigate. Third, survey research is suitable to investigate performance measures system and operational employee participation. Last but not the least, I couldn't obtain sufficient response on my own within one month. The larger database the better internal validity.

3.1! Participants

This survey is designed by Bianca A.C. Groen. I joined her thesis program and help translate English version of original survey into Chinese version. This make Chinese participants are available to provide information to database. I got 99 pairs of employees and their managers to answer the survey properly. Employees in this survey refer to employees that carry out operational activities of their organizations. Managers in this survey refer to managers that direct assess employee performance by using performance measures. Employees should meet three conditions. First, employees have to carry out operational activities of their organization. Second, employees should be familiar with the operational activities and have worked in their current functions for at least one year. Third, the organization must have performance metrics about employee’s performance.

3.2! Data collection

Pairs of managers and employees should answer same questions separately. Same questions answered by managers and employees are two different variables. Answers from managers about how employees could be heard can be mediator variable to select organizations that has enabling environment for employees. Managers and employees that agreed to participate in survey research would receive an email contained specialized survey of manager version or employee version. Each survey has only one serial number and could only be answered by email receivers. Employees and managers in pair will have similar trial of number to make them easily signaled together while dealing with data. The survey would take about 15 minutes to complete. System

22 would automatically send a reminder two weeks later to participants that have not complete all required questions. If any one of the pair could not finish all questions in time, this pair would be deleted from the database. There is no need to distinguish early and late respondents in my study. First, only the pairs of manager and employee that fulfil required criteria will be given specialized survey of manager version or employee version. Second, participants will receive a reminder to finish the survey two weeks after they received survey. The expiration date of survey is two weeks after the reminder. Participants are supposed to finish specialized survey within four weeks. Conditions of missing values or expired values will be automatically deleted by the system. These settings could make sure that all data collected are strictly fulfil the requirements. Moreover, this could increase the internal validity of this survey research.

3.3! Survey instrument

The dependent variable in this research is employee job performance reported by managers. As for operating employees, they focus on employee output to determine employee job performance. As for employees, they focus on whether their efforts could be totally expressed through performance measures, which indicates that employees will determine performance indicators by linking their efforts to performance. I select results from managers’ opinions rather than employees’ opinions on employee reported performance because I want to study whether employee participations in performance indicators could affect the employee output assessed by managers according to performance indicators.

The independent variable in this research is employee performance measures participation. The relationship is examined by dividing both independent variable and dependent variable into detailed variable. The independent variable is divided into six items. The first is “to what extent the employee always performs all tasks that are expected of him/her”. The second is “to what extent the employee always performs all essential duties”. The third is “to what extent the employee always fulfills all responsibilities required by job”. The forth is “to what extent the employee always engages in all activities that will directly affect his/her performance evaluation”. The fifth is “to what extent the employee always meets all formal performance requirements of the job”. The sixth is “to what extent the employee always completes all duties specified in his/her job description. The dependent variable is divided into the way and the content, i.e. how employees could be heard; how employees could influence performance indicators. These variables are analysed from employee aspects.

23 However, to control the situations that employees overestimate their influence or employees collude with manager, same questions answered by managers and employees are two different variables. Answers from managers about how employees could be heard can be mediator variable to select organisations that has enabling environment for employees. The item of mediator variable is to choose one answer to the question “If you have had influence on the performance indicators, how did that influence come about”. Five possible situations are given: give opinions without being asked, asked to give personal opinion, design indicators by myself, design indicators with others and give personal opinion in a group meeting. First three selections are concluded as forms by oneself, and the last two selections are concluded as forms with others. These detailed statements about situations that employees give opinions help participants quickly wake up their memories and make good categories.

Survey asks about influence content in following three ways and asks both employees and managers select to what extent they agree with the statements. Whether employees have influence on how the performance indicators are designed, whether employees have influence on the choice of which data are used as input into the performance indicators and whether employees have influence on ongoing modifications to the design of the performance indicators. Both employees and managers could select from totally disagree to totally agree in 7 grades. It is called a 7-point fully anchored Likert scale: (1) Totally disagree, (2) Disagree, (3) Moderately disagree, (4) Neutral, (5) Moderately agree, (6) Agree, (7) Totally agree. Each item of dependent variable and independent variable are assessed according to this scale. The mediator variable refers to a nominal variable because participants are required to choose one answer of the only question, and the answers are without an inherent order.

In order to allow for inferences with a high degree of internal validity, operating employee and his manager should answer same questions in their point of view respectively. Although they answer questions separately and independently, I compare both of their answers to one question to identify whether there is someone trying to hide facts or deceptions. This suggests that possibility of high involvement is important in studying on employees’ influence on performance indicators. As mentioned before, if managers have intention to let employees be involved in

24 design performance indicators and employees are willing to improve performance indicators, the possibility of employees’ involvement is high.

As for the influence on performance indicators, take employee’s answer into final result, but exclude the answer when it is largely different from what his manager’s answer. Employees may overestimate their potential influence on performance indicators and performance measurements system may have some objective bias when making changes. As for the employee reported performance, it is subjective purpose from managers and different items of this variable has similar indications which could be used for mutual authentication Take manager’s answer into final result, but exclude the answer only when it is largely different from what his employee’s answer.

25

3.4! Variable construct

3.4.1 reliability

Whether the divided items could perfectly reflect the variable could be tested easily by Cronbach alpha. Reliability reveals to what extent the items that supposed to measure a single construct could actually measure the only one and the same object. Cronbach alpha measures inter-item correlation by considering the number of items in a construct. The value of Cronbach alpha could be at the range of 0 and 1. If the construct is acceptable, the Cronbach alpha should be above 0.7. Cronbach alpha could be improved by deleting some irrelative items from construct. Considering this function, I calculated all five items of employee participation and all six items of reported employee performance with “descriptive if item deleted”. If the Cronbach alpha is under 0.7 I could easily select irrelative items and leave them out.

The Cronbach alpha of five items of employee participation is 0.957 and the separated Cronbach alpha of the five items are all above 0.9. This result indicates that the five items could be perfectly reflect the independent variable employee participation in performance measures. Besides, the Cronbach alpha of six items of reported employee performance is 0.915 and the separated Cronbach alpha of the six items are all above 0.8. This result indicates that the five items could be perfectly reflect the dependent variable manager reported employee performance. The result is provided in the table 1:

Table 1 Item-Total Statistics

Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted Independent variable --item 1 14.01 48.316 .920 .940 Independent variable --item 2 14.00 49.388 .880 .947 Independent variable --item 3 13.94 49.853 .903 .943

26 Independent variable --item 4 13.81 49.544 .856 .951 Independent variable --item 5 13.78 50.807 .840 .953 dependent variable --item 6 29.47 18.497 .743 .903 dependent variable --item 7 29.44 18.372 .771 .899 dependent variable --item 8 29.37 17.522 .836 .889 dependent variable --item 9 29.49 18.906 .681 .911 dependent variable --item 10 29.40 17.917 .789 .896 dependent variable --item 11 29.42 18.594 .748 .902 3.4.2 Convergent validity

Convergent validity could be measured by factor analysis. Factor analysis provides the communality in item variance. This means that there is correlation between different items of construct and this variance is in part caused by an unmeasured variable that has an effect on all the items. In this study, I use five relative items to represent the independent variable employee participation in performance indicators. However, I am not sure whether the five items could really represent the independent variable rather than something else. If the five items could represent the independent variable, each of the five items play a special role in this construct. Meanwhile, any one of the five items alone could not represent the independent variable. Convergent validity will be checked if there is no more underlying factor that could affect all five items related to employee participation in performance indicators. I first use factor analysis to

27 test whether there is a common factor that could explain the variance in total five items related to employee participation in performance indicators.

Both the component matrix of independent variable and the component matrix of dependent variable indicate that all items load on one factor. It is proved by only one column of component matrix. See the result in table 2 and table 3. According to total variance explained, the most important underlying factor explains 85.389% variance of employee participation in performance indicators. Meanwhile, the most important underlying factor explains 70.346% variance of manager reported employee performance. They confirm the convergent validity of employee participation in performance indicators and manager reported employee performance separately. See the result in table 4 and table 5.

Table 3 Component Matrixa

Component 1 dependent variable --item 6 .825 dependent variable --item 7 .844 dependent variable --item 8 .896 dependent variable --item 9 .773 dependent variable --item 10 .860 dependent variable --item 11 .829

Table 2 Component Matrixa

Component 1 Independent variable --item 1 .951 Independent variable --item 2 .924 Independent variable --item 3 .940 Independent variable --item 4 .907 Independent variable --item 5 .897

Extraction Method: Principal Component Analysis.a

28

Table 4 Independent Variable Total Variance Explained

Component Initial Eigenvalues Extraction Sums of Squared Loadings Total % of Variance Cumulative % Total % of Variance Cumulative %

1 4.269 85.389 85.389 4.269 85.389 85.389

2 .308 6.166 91.554

3 .224 4.475 96.029

4 .132 2.645 98.675

5 .066 1.325 100.000

Extraction Method: Principal Component Analysis.

Table 5 Dependent Variable Total Variance Explained

Component Initial Eigenvalues Extraction Sums of Squared Loadings Total % of Variance Cumulative % Total % of Variance Cumulative %

1 4.221 70.346 70.346 4.221 70.346 70.346 2 .598 9.969 80.314 3 .487 8.115 88.429 4 .314 5.228 93.657 5 .228 3.793 97.450 6 .153 2.550 100.000

Extraction Method: Principal Component Analysis.

3.4.3 Discriminant validity

Discriminant validity could be measured by factor analysis. After verification of convergent validity, whether the different items could indeed reflect the variable construct remains to be tested. Factor analysis could interpret that there is only one factor could explain the item scores. In this study, it is essential to verify all five items really belong to employee participation in performance indicators and all six items really belong to manager reported employee

29 performance. These eleven items are different and they will not fit better with other potential variable. Take all the eleven items together and repeat the Principal Component Analysis. All figures in rotated component matrix that above 0.5 together could determine which items belong to which factor. Component 1 refers to the five items of independent variable, while component 2 refers to the six items of dependent variable. There is no more component because employee participation in performance indicators and manager reported employee performance do not have factors in common. This is consistent with the convergent validity verification. See the result in table 6.

30

Table 6 Rotated Component Matrixa

Component 1 2 dependent variable --item 6 -.130 .826 dependent variable --item 7 .030 .844 dependent variable --item 8 .042 .896 dependent variable --item 9 .018 .773 dependent variable --item 10 -.001 .860 dependent variable --item 11 .047 .829 Independent variable --item 1 .951 .020 Independent variable --item 2 .924 -.011 Independent variable --item 3 .940 -.003 Independent variable --item 4 .909 -.005 Independent variable --item 5 .895 .011

31

3.5! Research model

When testing about the relations between independent variable and dependent variable, it is essential to categorize both independent variable and dependent variable and examine demographics and descriptive statistics. This preliminary analyses could help form a complete overview of different type of variables. Moreover, the table of correlations could be pre determined in this stage.

Descriptive statistics originally are the study of size, structure and distribution of populations and climate changes in response to education, ages and death. Descriptive analyse characteristics of respondents in selected sample including but not age, gender and education. Examining the range, mean, median and standard deviation according to characteristics of items of each construct. Independent variable employee participation in performance indicators and dependent variable manager reported employee performance are both classified as demographics. As for five items of independent variable employee participation in performance indicators, I take the average score of the five items as the score for the construct “PI participation”. As for six items of dependent variable manager reported employee performance, I take the average score of the six items as the score for the construct “employee job performance”. As for the mediator variable, there are five options to the questions that “If you have had influence on the performance indicators, how did that influence come about”. Each item represents one possibility and each item acts as a control variable. If one item is selected, it will get a figure 1; if one item is not selected, it will get a figure 0. The characteristics of items of mediator variable is the same as characteristics of gender, which gives out figure 1 if the gender is female. I count the number of selected items to mediator questions and take the number of items as the new construct “ways of influence”. As the mediator variable will be used later in statistical analyses, the five items and new construct “ways of influence” should all be put into the descriptive statistics table.

The results illustrate that scale variable PI participation and items of PI participation are typical scale variables. The minimum statistic is the lowest grade of 7-point scale and the maximum grade is the highest grade of 7-point scale. Employee job performance and items of employee job performance are typical scale variables too. However, the minimum statistic is grade 2 of 7-point scale and the maximum grade is the highest grade of 7-7-point scale. This result is consistent

32 with their result in mean. Employee job performance and items of employee job performance have a higher mean (5.8872) than PI participation and items of PI participation have (3.4768). This suggests that participants are more agreed with PI participation and items of PI participation than with employee job performance and items of employee job performance. Considering these are all subjective psychology questions, it also suggests that participants are more willing to choose they are agreed with items of PI participation than choose they are agreed with items of employee job performance. Further analyses would be given in next session.

As for the mediator variable, 69 participants choose one or more optimal answers, and the other 30 participants choose to specify their opinions by themselves. In this study, I delete the 30 participants because the subjective descriptions of ways to influence performance measures are difficult to measure and regression. According to the result of the new construct “ways of influence”, participants choose four options of ways to influence the performance indicators at most. Item 12, item 13 and item 14 have more supporters and the percentage of supporters are 26%, 29%, 35% respectively. Only 3% choose the item 16, which suggests it is a rare influence way that adopted in organisation between managers and employees. See results in table 8. All items are explained in the following table 7:

33

Table&7&Descriptions&of&items&

&& Descriptions& Measure&

PI&participation& Independent&variable:&employee&think&they&have&influence&on…& average&score&of&item&1&to&item&5& Item&1& how&the&performance&indicators&are&designed& scale&1&to&7&

Item&2& the&choice&of&which&data&are&used&as&input&into&the&performance&indicators& scale&1&to&7& Item&3& ongoing&modifications&to&the&design&of&the&performance&indicators& scale&1&to&7& Item&4& the&implementation&of&the&performance&indicators& scale&1&to&7& Item&5& the&maintenance&of&the&performance&indicators& scale&1&to&7&

Employee&job&performance& Dependent&variable:&manager&think&his/her&employee&job&performance…& average&score&of&item&6&to&item&11& Item&6& He/she&always&performs&all&tasks&that&are&expected&of&him/her& scale&1&to&7&

Item&7& He/she&always&performs&all&essential&duties& scale&1&to&7& Item&8& He/she&always&fulfills&all&responsibilities&required&by&his/her&job& scale&1&to&7& Item&9& He/she&always&engages&in&all&activities&that&will&directly&affect&his/her&performance&evaluation& scale&1&to&7& Item&10& He/she&always&meets&all&formal&performance&requirements&of&the&job& scale&1&to&7& Item&11& He/she&always&completes&all&duties&specified&in&his/her&job&description& scale&1&to&7&

ways&of&influence& Control&variable:&employee&think&they&could&influence&by…& sum&of&selected&items&from&item&12&to&item&16& Item&12& I&gave&my&opinion&without&being&asked& scale&0&or&1&

Item&13& I&was&asked&to&give&my&personal&opinion& scale&0&or&2& Item&14& I&gave&my&opinion&in&a&group&meeting& scale&0&or&3& Item&15& I&have&designed&the&indicators&together&with&others& scale&0&or&4& Item&16& I&have&designed&the&indicators&myself& scale&0&or&5&

34

Table&8&Descriptive&Statistics!

N! Minimum! Maximum! Mean! Std.!Deviation!

Statistic! Statistic! Statistic! Statistic! Std.!Error! Statistic!

PI!participation! 99! 1.00! 7.00! 3.4768! .17598! 1.75097! Independent!variable! DDitem!1! 99! 1! 7! 3.37! .193! 1.925! Independent!variable! DDitem!2! 99! 1! 7! 3.38! .192! 1.910! Independent!variable! DDitem!3! 99! 1! 7! 3.44! .185! 1.836! Independent!variable! DDitem!4! 99! 1! 7! 3.58! .195! 1.938! Independent!variable! DDitem!5! 99! 1! 7! 3.61! .188! 1.867! Employee! job! performance! 99! 2.00! 7.00! 5.8872! .08527! .84844! dependent!variable! DDitem!6! 99! 2! 7! 5.85! .101! 1.004! dependent!variable! DDitem!7! 99! 2! 7! 5.88! .100! .993! dependent!variable! DDitem!8! 99! 2! 7! 5.95! .105! 1.044! dependent!variable! DDitem!9! 99! 2! 7! 5.83! .102! 1.011! !dependent!variable! DDitem!10! 99! 2! 7! 5.92! .104! 1.037! dependent!variable! DDitem!11! 99! 2! 7! 5.90! .099! .985!

Ways! of! influence!

(control)! 69! 1! 4! 1.58! .098! .812! Mediator!variable! DDitem!12! 99! 0! 1! .26! .044! .442! Mediator!variable! DDitem!13! 99! 0! 1! .29! .046! .457! Mediator!variable! DDitem!14! 99! 0! 1! .35! .048! .480! Mediator!variable! DDitem!15! 99! 0! 1! .16! .037! .370! Mediator!variable! DDitem!16! 99! 0! 1! .03! .017! .172! Valid!N!(listwise)! 69!

37

4& Result

Present a table with all the correlations between the variables could help get an overview of the bivariate relations between variables. As shown above, all variables in this study are scale variable so there is no need to present both parametric Pearson correlations and non-parametric Spearman rank-order correlations. It is important to distinguish between correlations that are significant at the 0.1, 0.05 and 0.01 level.

Table 9 illustrates the Pearson correlations for all variables. There is an effect between PI participation and control item 12 (r=0.264, p <0.01), and there is an effect between PI participation and control item 15 (r=0.2356, p <0.01). These two significant relationship could indicate the control items separately correlated could have more possibility to get relationship than control items integrated together as construct “ways of influence”. The moderation function of control items cannot be verified in this stage.

These results provide some preliminary evidence of a relation between construct “ways of influence” and its divided control items. More specifically, there is a significant effect between control item 12 and ways of influence (r=0.368, p <0.01). This suggests that if the employee feels that he could gave his opinion without being asked, the total number of ways of influence could increase. Further, there is a significant effect between control item 13 and ways of influence (r=0.408, p <0.01). This suggests that if the employee feels that he was asked to give his personal opinion, the total number of ways of influence could increase. Further, there is a significant effect between control item 14 and ways of influence (r=0.421, p <0.01). This suggests that if the employee feels that he could give his opinion in a group meeting, the total number of ways of influence could increase. Further, there is a significant effect between control item 15 and ways of influence (r=0.500, p <0.01). This suggests that if the employee feels that he could have designed the indicators together with others, the total number of ways of influence could increase. This implication could together prove that employees feel they could have many ways to influence the performance indicators during their work time. Additionally, more enabling the environment in organisation, the more ways employees could be heard by their line managers. Different ways of influence on performance indicators have mutual positive effects.

38 Due to the inspiration of the separated items of constructs, I also do Pearson correlations between five items of PI participations and six items of employee job performance. See the results in table 10. The result suggests that the five items of PI participation have correlations with each other and the six items of employee job performance have correlations with each other. As for the relationship between five items of PI participation and six items of employee job performance, there is no clearly suggestions. Further results will be show in the next session.

39 **.#Correlation#is#significant#at#the#0.01#level#(27tailed).##*.#Correlation#is#significant#at#the#0.05#level#(27tailed).

Table&9&Pearson&Correlations&Between&Construct&Variables&

&& && && 1& 2& 3& 4& 5& 6& 7& 8&

1& _{PI&participation&}

& 1&

&

2& Employee&job&_performance&

&

0.004&

1&

& Sig.&(2Itailed)& 0.968& & 3& ways&of&influence&

&

0.135& 0.097&

1& & Sig.&(2Itailed)& 0.269& 0.43& & 4& control&item&12&

&

.264**& I0.024& .368**&

1& & Sig.&(2Itailed)& 0.008& 0.81& 0.002& & 5& control&item&13&

&

.254*& I0.181& .408**& 0.019&

1& & Sig.&(2Itailed)& 0.011& 0.073& 0.001& 0.849& & 6& control&item&14&

&

.210*& 0.09& .421**& 0.039& 0.174&

1& & Sig.&(2Itailed)& 0.037& 0.373& 0& 0.703& 0.085& & 7& control&item&15&

&

.356**& 0.053& .500**& 0.112& 0.079& 0.135&

1& &

Sig.&(2Itailed)& 0& 0.601& 0& 0.269& 0.436& 0.184& &

8& control&item&16&

&

0.175& I.232*& 0.023& 0.028& I0.114& I0.131& 0.082&

1& && && Sig.&(2Itailed)& 0.083& 0.021& 0.851& 0.78& 0.262& 0.197& 0.417&

40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 Item/1 1 2 Item/2 .902** 1 Sig./(26tailed) 0 3 Item/3 .876** .818** 1 Sig./(26tailed) 0 0 4 Item/4 .844** .797** .782** 1 Sig./(26tailed) 0 0 0 5 Item/5 .765** .750** .864** .771** 1 Sig./(26tailed) 0 0 0 0 6 Item/6 60.097 60.129 60.113 60.159 60.043 1 Sig./(26tailed) 0.339 0.203 0.267 0.115 0.672 7 Item/7 0.061 0.009 0.007 0.021 0.051 .688** 1 Sig./(26tailed) 0.546 0.932 0.942 0.839 0.616 0 8 Item/8 0.055 0.03 0.038 0.004 0.073 .743** .674** 1 Sig./(26tailed) 0.587 0.766 0.706 0.965 0.47 0 0 9 Item/9 0.028 0.024 0.064 60.006 60.025 .558** .681** .553** 1 Sig./(26tailed) 0.783 0.814 0.532 0.951 0.803 0 0 0 10 Item/10 0.01 60.015 60.035 0.028 60.001 .547** .655** .769** .610** 1 Sig./(26tailed) 0.92 0.882 0.734 0.78 0.994 0 0 0 0 11 Item/11 0.047 0.037 0.031 0.095 0 .614** .551** .740** .526** .732** 1 Sig./(26tailed) 0.644 0.715 0.763 0.35 0.997 0 0 0 0 0 12 Item/12 .279** .230* .282** 0.155 .275** 0.137 60.02 0.007 0.011 60.154 60.102 1 Sig./(26tailed) 0.005 0.022 0.005 0.125 0.006 0.178 0.846 0.946 0.917 0.129 0.313 13 Item/13 .210* .220* .220* .303** .220* 6.236* 60.123 60.161 60.155 60.143 60.092 0.019 1 Sig./(26tailed) 0.037 0.028 0.029 0.002 0.029 0.019 0.224 0.111 0.126 0.157 0.364 0.849 14 Item/14 .198* 0.151 0.167 .272** 0.18 60.057 0.069 0.097 0.063 0.181 0.098 0.039 0.174 1 Sig./(26tailed) 0.05 0.136 0.098 0.006 0.075 0.575 0.495 0.339 0.534 0.073 0.335 0.703 0.085 15 Item/15 .358** .345** .329** .310** .300** 60.071 0.165 60.005 0.102 0.088 60.011 0.112 0.079 0.135 1 Sig./(26tailed) 0 0 0.001 0.002 0.003 0.486 0.103 0.96 0.314 0.389 0.916 0.269 0.436 0.184 16 Item/16 0.119 0.181 0.151 0.131 .228* 60.15 6.336** 60.105 6.263** 6.215* 60.102 0.028 60.114 60.131 0.082 1 Sig./(26tailed) 0.239 0.072 0.137 0.198 0.023 0.138 0.001 0.301 0.009 0.033 0.315 0.78 0.262 0.197 0.417 Table/10/Pearson/Correlations/Between/All/Items