Amsterdam Business School
The effect of loyalty and job-hopping in
subjective performance evaluation
Name: Kevin Alsemgeest Student number: 11425121
Thesis supervisor: dr. ir. S.B.M. Morssinkhof Date: August 14, 2018
Word count: 16.148
MSc Accountancy & Control, specialization Control
Statement of Originality
This document is written by student Kevin Alsemgeest 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
Over the past decades, the labor market has been growing with job-hoppers. A lifetime employment is becoming less interesting and the new generation employees tend to choose for the boundaryless career path. This path is characterized by a career that is driven by the individual which consists of inter-organizational career changes, personal responsibility for the career and the development of a social network that supports this career. As inter-organizational career changes and the number of job-hoppers is growing, little is known about the evaluation process of these employees. This study analyzes the effect of job-hopping on the subjective performance evaluation process. This thesis tests if there is a difference between the evaluation process of a job-hopper and a loyal employee. This thesis investigates the research question with a case-based experiment, using accounting professionals from a Dutch accounting firm. The results show no indication that there is a difference between these two groups in the main experiment. When comparison of the two employees is possible there is an indication that these groups are differently evaluated, which is in favor of the job-hopper. This could be the effect of supervisors risk aversion.
Keywords:
Performance evaluation, boundaryless career, subjectivity, employee turnover, growth opportunities
1 Contents
1 Introduction ... 2
2 Literature review and hypotheses ... 4
2.1 Prior literature ... 4 2.2 Job-hopping ... 6 2.3 Growth opportunities ... 8 3 Method ... 10 3.1 Participants ... 10 3.2 Experimental design ... 10 3.3 Variables ... 11 3.4 Experimental procedures ... 11
3.5 Testing the experiment ... 13
3.6 Validity and considerations ... 13
4 Results ... 15
4.1 Description of the sample ... 15
4.2 Testing the hypotheses of the overall models ... 16
4.3 Hypotheses tests within the model ... 20
4.4 Testing reversed results ... 24
4.5 Supplemental analyses ... 25
4.6 Evaluation of the experiment ... 26
5 Conclusion ... 27
6 Limitations and further research ... 29
References ... 31
Appendices ... 33
Appendix I: experimental case setting ... 33
Appendix II: statistical information about the sample ... 37
Appendix III: supplemental analysis for the overall model ... 39
Appendix IV: statistical results of the reversed performance evaluation scores ... 42
2
1 Introduction
In many organizations supervisors can use some discretion for the performance evaluation of their employees. The subjectivity in the performance evaluation gives managers the opportunity to use information about employees that has not been captured by objective performance measures (Bol, 2008). Using more information for the performance evaluation process, gives a better understanding of employee performance which reduces the market risk for employees and gives managers the opportunity to create incentives for their employees (Bol & Smith, 2011; Longenecker et al., 1987). While it’s important to capture al the dimensions of the performance of an employee, theories suggests that subjective judgements can be influenced by the behavior of employees. This means that supervisors unknowingly process information about the performance of employees (Bond et al. 2007).
Using more information about employees performance can be beneficial for employees who have more years of service for the same organization and the same supervisor. They will most likely be compensated when they have some bad luck (Bol & Smith, 2011). Unfortunately managers don’t have a background for all their employees since some employees are new to the company. Over the past decade there have been changes on the job market, people stay for a shorter time at the same organization and leave after a few years, to work for another company. This process is better known as job-hopping (Steenacker and Guerry, 2016). If an employee is job-hopping, managers have less background information about these employees, which means there is less information available for the performance evaluation. This raises the question if more loyal employees are evaluated differently on their performance than employees who tend to be job-hopping. Therefore this master thesis will research the following question: Is there a
difference in the performance evaluation between an employee who just started working for the organization (new) and an employee who works years for the same organization (experienced)?
In this experiment, participants assume the role of a sales manager with the task to evaluate one of their employees. For this evaluation the participants receive some case based information such as some objective performance measures, some background information and how many years an employee is working at the organization. This experiment is an 2 x 2 between subject design. In the four conditions of this experiment the years of experience within the organization (new or experienced) and the performance of the employee on the objective performance measures (high and low) are manipulated. The participants in this experiment are asked to give a rating for the performance of the employee and the growth opportunity of that employee.
Shahnawaz and Jafri (2009) argue that job attitude and job satisfaction are good predictors for employees leaving the organization. Because Bond et al. (2007) finds that all available information shall be processed by the supervisor, we expect that job attitude and job satisfaction will influence the supervisors expectation of an employee, which will influence the
3 performance evaluation. When supervisors use this indirect information we predict that when the overall objective performance of an employee is low, managers shall reward the experienced employee above the new employee. Because there is no information available about the maximum and minimum results of the new employee. Furthermore, we expect that when the overall performance is high, the manager shall give a higher rating to the new employee above the experienced employee to create a greater incentive for the new employee. It is expected that these predictions are not the same for the growth opportunity of an employee, because it is in the managers best interest to motivate all these high performing employees to reduce the risk that high performing employees leave the organization. Therefore, there is no difference expected between the growth opportunity of a new or experienced employee. It is expected that high performing employees have better growth opportunities than low performing employees, indifferent of the years of service.
This thesis makes two contributions to the literature. First of all, while many papers state that subjective performance ratings are often biased (Bol, 2008; Bol and Smith, 2011; Bol et al., 2016; Longenecker et al., 1987), less is known about the effect of loyalty and job-hopping on the evaluation process. This thesis contributes to the literature by examining if loyalty and job-hopping has an effect on the evaluation. Second, despite there is a growing number of papers that are trying to explain why job-hopping is a growing phenomenon in the past decades (Arthur et al., 2005; Dougherty et al., 1993; Steenackers and Guerry, 2016), there still is little information about the effect of this phenomenon on the performance evaluation. This thesis shall contribute to the literature by analyzing if there is an effect of job-hopping and loyalty on the evaluation of employees.
This master thesis is organized as follows. Chapter 2 provides a brief literature review of subjectivity in performance evaluation, job hopping and growth opportunities. Throughout this chapter the hypothesis are developed. Chapter 3 describes the method, with attention for participants selection, experimental design, experimental procedure and validity threats. Chapter 4 presents the results of the experiment, which were analyzed using SPSS. Chapter 5 provides the conclusion of this research and chapter 6 explores the limitations and directions for further research.
4
2 Literature review and hypotheses
The first part of this chapter gives an overview of the recent literature on subjectivity in performance evaluation. The next section provides a definition of job hopping and describes it’s effects. The last part is shows the effect of growth opportunities for an employee. Throughout this chapter some hypothesis are developed regarding the outcome of the experiment.
2.1 Prior literature
There has been a lot of research on the effects and the possible (dis)advantages of subjectivity in performance evaluations. Where objective performance evaluation is independent of the supervisors preferences and biases, the subjective part of the performance evaluation is not. “An objective measure exists as a quantity in and of itself; in contrast, subjective evaluations are based on attitudes, beliefs and perceptions” (Woods, 2012, p. 403). This means that the outcome of a subjective evaluation process will be different, depending on many factors.
There are several advantages for using subjectivity in the performance evaluation. An important advantage when using subjectivity is that objective performance measurements can add a level of noise to the evaluation. Based on the agency theory and risk aversion of employees, greater noise in the objective measurements can lead to more risk for employees (lower performance evaluation due to the wrong or incomplete measures) which can results in less effort by the employees. Subjectivity in the performance evaluation process can compensate for the noise of the objective targets (Kunz, 2015). Another advantage of subjectivity is that it captures all the dimensions of the performance of an employee, in contrast to the objective measurements which tend to discount some activities of an employee. This can lead to an employee neglecting some of the activities, because there is no payoff for those activities (Bol, 2008; Feltham and Xie, 1994). In line with this, Golman and Bhatia (2012) argues that companies should use performance measures that captures all of the employees activities. This means that opinions, preferences and other subjective aspects have to be incorporated for an optimal evaluation process.
The above described studies explain that subjectivity can compensate for the noise in objective performance measures. However, studies have found that scenarios in which supervisors have to give a subjective evaluation can result in a number of different problems for the evaluation. Topics of the problems with subjectivity can be the limited information that the supervisor has of its employee, the incentive role of the evaluation or leniency and compression by the supervisor (Bol, 2008). The problems of subjectivity are related to the biases of the supervisor, which arise from with impressions and history of the employee, personal preferences and organization goals. These biases of the supervisor can be expressed in the performance evaluation in different ways. For example, the supervisor has the ability to give different valuations to the different objective performance measures and the supervisor can adjust the
5 evaluation to correct for external effects which were not yet incorporated in the evaluation (Bol & Smith, 2011). One of these examples was found by Longenecker et al. (1987), their findings indicate that executives often prioritize the incentive goals above the accuracy of an evaluation. In these cases the performance evaluation becomes more focused on the appraisal of employees for a better motivation and encouraging the poor performers (with the result that any hope for objectivity and accuracy walks right out the door). Bol (2011) finds that supervisors tend to give employees a rating more directed to the average performance. This indicates some level of leniency and compression of supervisors, which is associated with evaluations that are focused on incentives and avoiding difficult decisions.
Since evaluators don’t always have enough information about their employees, there will always be some agency costs involved. Bol et al. (2016) argues that it is not always in the best interest of the manager to give the most accurate evaluation possible. A more accurate evaluation will cost the manager more time for collecting information and analyzing the data that is informative for the performance of the employee. These opportunity costs (spending time and effort on performance evaluation) will not always be in the managers interest for providing accurate evaluations. Instead, it is more likely that supervisors will give ratings that will cost only a little extra work, but are accepted for higher management and employees. These ratings will be easy to explain and will cause little stress to all participants (Bol, 2011).
Bond et al. (2007) finds that the knowledge of supervisors about performance of employees on a specific task will likely be affected by the performance on another separate task by making use of cognitive information distortion. This means that supervisors unknowingly process information about the performance of employees, this will favor their belief about those employees and will result in an evaluation with a desired outcome. In line with these findings Bol and Smith (2011) finds that supervisors tend to give ratings that are higher and more compressed when their employees suffer from uncontrollable outcomes. This is consistent with the line that managers tend to give employees ratings which will have the lowest personal and opportunity cost. Maas & Verdoorn (2017) finds that even when there is a lot of information about the performance of an employee, analyzing the information in a structured way doesn’t give a higher performance rating. This gives us reasons to believe that the input of objective information has less impact on the ratings of employees.
As discussed above, there are cost of leniency (gentleness) and compression which arise when there is subjectivity in the performance evaluation. An example of these cost can be the cost of a supervisor spending too much resources on the evaluation process. These are direct costs for the organization. There are also indirect costs of leniency and compression from the employee perspective. For example, if an employee experience that the incentives is not in line with its performance (to low), they might lower their performance because working harder does not necessarily mean a higher performance evaluation. In line with these reasoning, employees with higher performance ratings won’t feel the pressure for keeping up there ratings because
6 the already expected to get a favorable rating in the evaluation. This makes it more difficult to identify the top performance of an organization which can lead to unfavorable promotions, replacements and allocation decisions. These costs will also arise on a management level in organizations. Some managers tend to be more lenient and use more compression than other managers, which will lead to different results for different managers. This makes the performance evaluation of managers unable to compare with each other (Bol et al. 2016). These flawed performance evaluations and business decisions can lead to feelings of unfairness and frustration (Ittner et al. 2003).
As managers tend to be lenient in the performance evaluation, so the resistance of employees is at a minimum, it is an important process to motivate the employees. Colberg et al. (2007) has shown that employees attach great value to comparison of the performance evaluation to their peers, which affects their perception of fairness. If peers get a similar reward for less work, employees will experience stress and anger which can result in confrontation and reduced efforts (Colquitt et al., 2001) 1. Because managers do not want their best employees to
get demotivated, they will give their high performers a better score, so their employees get the feeling that they are rewarded for their hard work (Longenecker et al., 1987). I predict that supervisors will reward high performers with a better result than the low performance, therefore to following hypothesis shall be tested:
H1: The supervisors performance evaluation will be higher for the employees with a high result than for employees with a low result, for both new and experienced employee.
2.2 Job-hopping
Over the past decades there has been a switch from long-term employment for the same company to the phenomenon of job hoppers. This phenomenon was first described by Ghiselli (1974) who named it “hobo-syndrome”, which means that employees frequently change their jobs because of instinctive impulses. An important refinement was described by Dougherty et al. (1993) which argued that job-hopping is not about changing jobs, but frequently changing companies. This definition eliminates the job changes within companies and promotion based changes and gives a focus on inter-organizational transitions. Another important factor of these inter-organizational transitions is to know if these transitions are made voluntarily, because these transitions involve the costs of losing valued employees (Dalton et al., 1982). Maertz and Griffeth (2004) add to this definition that not all job-hopping decisions are made by instinctive impulses, but might be the result of employees making rational decisions. Given the above
1 An important assumption for this theory is that employees have some sort of general idea of how
their performance evaluations compares to their peers. This means there is some sort of insight in the peers evaluation as well as some self-knowledge about their own performance.
7 described characteristics we follow the definition of Steenacker and Guerry (2016), which states that the definition of job-hopping is making frequently voluntary inter-organizational transitions.
In the past, it was a bad sign when employees changed their jobs every couple of years. It lead to the general assumptions that job hoppers can’t hold down a job, can’t get along with colleagues, or that they simply can’t commit. In the perspective of the organization where the job-hoppers responsible for higher education costs. This is because the education of employees can be high and if they leave after a couple years the return of investment is relatively low (Campbell et al., 2012). This phenomenon is known as the human capital theory which states that investments in human capital increase future productivity and that the departure of an employee are associated with a loss of knowledge and implying intangible costs for the organization (Becker, 1964).
Tziner and Birati (1996) also argue that high employee turnover is associated with high costs. Examples of these costs are training costs, costs of additional recruiting, administrative costs and a loss of knowledge and experience. In addition to these more direct costs, if an employee voluntarily leaves a company, it might have a negative effect on the attitude of the employees who decided to stay. This can have a negative effect on the job satisfaction and productivity of these employees. Besides these negative effects that are associated with job-hopping, there might be some benefits of job-hopping for the economy. Nowadays it is even encouraged to switch jobs every couple of years for your own personal growth. High personal growth fosters a knowledge spillover effect, which is easily reallocated to the organizations that needs it the most (Steenacker and Guerry, 2016). According to Arthur et al. (2005) the number of job-hoppers will increase because of the boundaryless career theory, which will mostly effect the younger generations in the workforce. Arthur (1994) defined boundaryless career as a career which is no longer characterized by only one career opportunity and one organization, but a career that is driven by the individual. The most important aspects of a boundaryless career are an inter-organizational career, personal responsibility of his/her career and the development of a social network that supports this career path.
Since more employees stay for a shorter time on the job and leave after a couple of years to work for another company, there arise problems for the supervisors to evaluate their employees. For employees who work for a long lifetime by the same organization, supervisors get the opportunity to know the background of the employee, what is happening in their life, who do they react to problems and why and how do they make certain decisions. It indicates that as time passes supervisors are getting more understanding about their employees, e.g. the supervisor will get information about the attitude and job satisfaction of the employee. Shahnawaz and Jafri (2009) finds that job attitude is highly related to the intention to leave the company and job satisfaction is a good predictor for employee turnover. This means that the
8 behavior of employees to leave organization can be predicted by the attitude of the employee to the supervisor.
As managers can use all the information available to them and even more through cognitive information distortion (Bol & Smith, 2011), it is expected that supervisors will experience the attitude and job satisfaction of their employees. When managers use the indirect information about costs/benefits of job-hoppers and the effect on job attitude and satisfaction, this will most likely affect the employees evaluation. I predict that when the overall objective performance of employees is low (both the new and experienced employee), managers shall discriminate the experienced employee above the new employee. The experienced employee shall receive a higher rating because the supervisor as a better understanding of the employee and will reward loyalty. This is also in line with the theory of risk aversion, because the supervisor as a good understanding of the base line of the experienced employee.
H2: Supervisors performance evaluation will be higher for relatively experienced employees than for new employees when the overall performance is low.
Since less information about new employees is available, they have the opportunity to exceed expectations. Furthermore, there is no information available about future results of employees, so managers are not able to tell what the maximum capacity of a new employee is. I predicted that when the overall objective performance of employees is high, managers tend to give new employees a greater incentive for even more results next year. The supervisors already have a good idea about the maximum capacity of experienced employees, so the incentive role will be lower.
H3: Supervisors performance evaluation will be higher for relatively new employees than for experienced employees when the overall performance is high.
2.3 Growth opportunities
Promotions in organizations have been a fascinating subject in literature for the past decades. This fascinations comes from the different functions that a promotion can offer. The different roles of promotions can be (a) ensuring the best allocation of human recourse in an organization, (b) has an incentive aspect for employees and (c) forming coalitions within organization to give and maintain power to certain groups. For these reasons, promotions can have an effect on multiple stakeholders, which make promotions an important aspect of the human resource and evaluation process (Kaplan and Ferris, 2001). All these roles are part of the evaluation of employees which make this process subjective and even some sort of political within the company (Longenecker et al., 1987).
9 Even though the great effects of a promotion, the current development in the market shows that there are more and more supporters of the boundaryless career. Arthur (1994) argues that the rise of the boundaryless career is driven by a new form of economy in which changes occur in technology, flexibility and job security. This means that the old development of an employee with one employer is less and less an option for an employee. This old way of career making can be defined as a linear career of an employee in which an employer has some sort of responsibility about the career of the employee. This traditional career path is characterized by job security and an upward mobility in the organization (Eby et al., 2003; Hall and Moss, 1998).
This switch form the traditional career to the boundaryless career has an impact on the turnover of employees and human resource management. As previous discussed, a high employee turnover leads to different problems like a loss of knowledge, problems with organizational unity and maybe problems with continuity of a department. Because of these problems there are several chances for human resource management. One solutions to work with boundaryless career, is to reduce the internal boundaries of career path (e.g. a requirement of years of service in the organization). By giving employees more opportunities to grow within the company it is less likely that they will turn to other companies, which reduces the employee turnover (Parker and Inkson, 1999).
From the above we can state that a promotion of an employee has a great effect for the peers in the company and it is a way to satisfy good employees, so they keep working for the organization. Bol and smith (2011) stated that through cognitive information distortion managers will use all the information available to evaluate an employee. They will know that high performers will have to be satisfied so they keep working for the company and managers have the information which employees are the top performers. As high performers are the top priority for promotion decisions It is expected that managers are indifferent about how long an employee works for the company. So I predict that an employee will score high on growth opportunities as long as the performance of the employee is above the target.
H4: The supervisors performance evaluation about growth opportunities will be higher for the employees with a high result than for employees with a low result.
10
3 Method
To answer the research question a case-based experiment is used as the research method. In this case-based experiment different extreme scenarios are created and every participant is allowed to complete one scenario. This way the participants won’t see the manipulation that is created within this experiment. The next sections describe the sample selection, experimental design, variables, experimental procedure and some validity concerns.
3.1 Participants
The participants of the experiment work for an accounting firm in The Netherlands. The participants are from different disciplines of the company (audit, small and medium entities and taxes). For this experiment the employees of two different offices were used. The different disciplines on the different locations were informed about the experiment during a monthly meeting, after a brief explanation they received the questionnaire through a mail.
The questionnaire has been sent by the partner of the location and was followed up with a reminder after two weeks. In this process of sending reminders to the full population, the different disciplines and locations were visited to describe the reason of the questionnaire (master thesis), so that there were more people inclined to fill in the questionnaire. Of the 87 participants of the experiment, 57 respondent on the first moment. This means that 30 more participants completed the experiment after the reminder.
Libby et al. (2002) indicates that it is important to match the participants with the goal of the experiment. The goal of this experiment is to investigate if loyalty has an effect on the performance evaluation. By researching this subject it is a best match if participants are selected that experience this kind of scenario in real life. Therefore, these professionals were selected as participants in the experiment. The participants of this experiment have different knowledge about evaluating the performance of employees. All participants work in an accounting setting, but the level of evaluation and support responsibilities differ throughout the organization. The level of participants is from entry level junior employee to partner level. The participation was entirely voluntary and anonymized.
3.2 Experimental design
The experiment is a 2 x 2 between-subjects design. In this experiment there is a manipulation of the years of service of an employee so that the more experienced employee works several years for the same organization and the new employee works not even a year for the organization. It is important for this experiment to emphasize that the subject of this research is about the difference between an employee that just entered the organization and an employee who has worked for a long time in the organization. There will be no difference between they experience of the employees. Therefore, in the experiment is emphasized that
11 the employees is of the same age with the same years of experience and the same educational background, so that a bias about less experience is at its minimum.
Furthermore the results over the past year of the employees is manipulated, so an employee has performed above the target (high performance) or below the target (low performance). Important for this manipulation is that the participants get an idea if the employees is under or above their target so a quick judgement is possible. Therefore the manipulation of this variable is done with a percentage. This means that the sales results of the target, the high performing employee and low performing employee are presented as a percentage of the total sales.
3.3 Variables
The dependent variable is the subjective performance evaluation of new and experienced employees by their supervisors. For evaluating the employees, the participants received case information which included background information about the employee, objective ratings and years the employee worked for the organization. The participants have rated the employee on a scale from 1 to 10 to avoid the centrality bias described below. Another dependent variable is the growth opportunity of the employee. Based on the earlier information a supervisor received of the employees (results and years of service in the company), the participants rated the employees growth opportunity within the company on a scale from 1 to 10.
The independent variable consists of the objective measures used in the performance evaluation which indicate if an employee has a high or low rating and the years that an employee have worked for the organization. In this experiment the more new employee works for 11 months in the organization and the more experienced employee works 11 years for the organization. Both employees have 11 years of experience within the field of sales specialist.
The other independent variable is the result of the employee. The measures for high and low sales are defined in percentage of the total available sales. In this case the target is set as 95% of the total sales. For the low performing employee the result is a 80% sales turnover, a high performing employee accomplished a 100% of the possible sales. In the general case information there is information provided that gives the participants the idea that the target was possible for all employees in the sales team.
3.4 Experimental procedures
Participants of the experiment completed a performance evaluation of an employee in a hypothetical case setting. Their role as sales manager was described as overseeing and controlling a number of employees in their sales department. In their sales team was a mix of new and experienced employees. The experimental task was to perform an evaluation on one of the employees (this could be a new or experienced employee). The participants were given a set of objective results of the past year of the employee to make their assessment.
12 The case information informed the participants that it was the sales managers duty to conduct a performance evaluation every year for all the employees under their supervision. This evaluation was an end of year evaluation of the employee. The process for the evaluation was described as using all the available information they had about the employee. The instructions of the experiment noted that the participants had to rate the employee on two components and give an evaluation for another employee who had less or more years of service for the organization. The components which were evaluated, where the performance and the growth opportunities of the employee.
For the objective measure the participants received a percentage of the results related to the target that was set at the beginning of that year. The sales results of the employees were manipulated to be either high (100%) or low (75%). For the subjective part of the evaluation, the participants received some personal notes they had made (as sales manager) about the employee over the past year(s). These notes were manipulated only on the years the employee work for the company. Furthermore, the new and experienced employee received the same notes which gave the participants mostly work-related information about the years of experience, how the employees functioned in a team and some personal information. An important note for both employees is that they are at the same age in their early 30’s, so there is less confusion about the experience level of both employees.
After the participants gave a rating for both the performance and the growth opportunity, they were asked to evaluate another employee. When the participants where first in the scenario of a new employee with 11 months of experience in the same organization, they had to make a new performance evaluation of the experienced employee with 11 years of experience in the same organization. When the participant was first in the scenario with the experienced employee, they had to make a new performance evaluation for a new employee.
After this second performance evaluation, participants were asked some general questions. These questions were about gender, working experience, evaluation experience, educational level and how the participants experienced the ease and difficulty of the experiment. This questionnaire was done to evaluate at the end of the experiment if there was enough diversity within the group of participants. These questions were also asked to analyze if there is a difference for evaluating new and experienced employees and the feeling that the participants had about the information that was given to them.
Participants completed the experiment in an online setting by entering a URL into the web browser (Qualtrics), which they received through an email. The online survey program ensured that the participants were randomly assigned to one of the four scenarios. After the first participants had filled the questioner, a flaw in the randomizer was found. The first scenario (new employee with a high result) already contained 23 participants and the other scenarios were half way. To correct this, the first scenario was deleted out of the randomizer just before
13 the reminder was send. The distributed experimental case that was sent to the participants can be found in appendix I.
3.5 Testing the experiment
Before the experiment was send to the participants, the experiment was tested by four co-workers, which all filled the questionnaire twice. Testing the experiment is necessary because this experiment has never been distributed before, so there are no previous records about this experiment. Testing the experiment reduced the risk that participants would not understand the questions or that variables ware not defined enough. The test participants were asked to fill in the questionnaire twice and give some feedback about the experiment.
The tests showed that experiment was working as it should be. This means that the test participants received the experiment in the right order and were placed in random scenarios. Furthermore the tests gave the indication that all test participants understood the questions that were asked (to evaluate an employee) and that there is a difference between the result scenarios and the employee scenarios. After a test participant had conducted the experiment twice, they were asked face-to-face if they saw the changes in the experiment. The test participants described the changes (results, years of experience or both), which were in line with their scenarios.
3.6 Validity and considerations
It is important for an experiment that it is carried out as efficient and effective as possible. This means that an experiment should be conducted as economically as possible and still achieve a given level of effectives. An effective experiment provides evidence that the internal validity is as high as possible, while it still has enough external validity to generalize the results to a wider population (Libby et al., 2002). For the external validity, an important step is the selection of the participants which is discussed above in section 3.1. The internal validity is about a strong relationship between the dependent and the independent variable. This is why it is important to properly indicate the limits of these variables.
Using an experiment as a method gives some validity threats and limitations. Threats for this research design can be the central tendency, which is the extent of supervisors trying to avoid the extreme values. In this research design a 1 – 10 points scale is used, which was also used by Bol & Smith (2011). This scale proved to deliver good results in their paper. Another threat for this research in particular is the definition of new and experienced employees in the experimental design. In this thesis project a new employee means new to the organization and not new to the workforce. This master thesis is about the effect of loyalty and job-hopping of employees on the performance evaluation and not about the difference between young and older employees. It is an important note in the design of the experiment that the evaluated employee is in their early 30’s. Therefore, a table of the employee’s information is added to the
14 experiment. So beneath the written information a table is presented, showing all the important information. This means that at first the table presents the years of experience in total and after that the years of experience within the company. This reduces the risk of participants jumping to conclusions.
A limitation for this research method is the external validity. As earlier indicated, the participants of the experiment were professionals. This gives an indication that the results can be used for other organization. A limitation for the external validity is the use of the extreme scenarios in the experiment. A variable that creates a part of the extreme condition is the difference between the new and experienced employee, which is more than 10 years. By doing so, the opportunity of finding results on the hypothesis will increase, but these scenarios rarely occur in the real world.
15
4 Results
This chapter presents the results of the experiment. At first the sample of the experiment is described is this chapter. Second the overall results of the model are tested and discussed, which gives an indication if there are differences between the result groups (high and low) and the employee groups (new and experienced). The following section tests the results within the model and discusses the hypothesis about job-hopping. In the fourth section the reversed results are analyzed and discussed. This chapter ends with some supplemental analyses which contain and discuss the manipulation check.
4.1 Description of the sample
A total of 103 participants have participated in the experiment. Of the 103 filled out forms a total of 87 can be used for further analysis. The reason for the difference between the filled out and used forms, is that some forms were not fully completed. The 87 participants were randomized in the four scenarios which were made by the manipulation of the performance of an employee (high or low result) and the years that an employee worked for the company (new or experienced employee). The number of participants within the scenarios is shown in table 1.
Table 1
Number of participants in the experiment
New Employee Experienced employee Total
High Result 23 21 44
Low result 20 23 43
Total 43 44 87
The participants work for an accounting firm in The Netherlands, with different years of experience and with different evaluation experience. The average years of work experience is 9 years, with a minimum of 1 year and a maximum of 36 years. The average years of experience on evaluating other employees is 3 years and 9 months, with a minimum of zero years and a maximum of 22 years. The participants with zero years of experience in evaluating other employees is 33% of the sample population.
Of the participants a total of 34 of the 87 were female, which is 39,1% of the total. The average years of working experience for women is 7 years and 9 months and the average years of experience in evaluating employees is 3 years and 4 months. The average years of working experience for men is 9 years and 10 months and the average years of experience in the evaluating process for men is 8 years and 3 months.
Furthermore, there can be made a distinction in education level. A distinction can be made between the education levels WO (university degree), HBO (university of applied sciences
16 degree), MBO (associate degree) and High school degree. An overview of the participants educational level and gender is shown in table 2.
Table 2
Overview of educational level of the participants
Female Male total
WO 11 18 29
HBO 20 30 50
MBO 2 5 7
High school 1 0 1
Total 34 53 87
A total of 79 participants has an educational level of HBO or higher, which is 90,8% of the sample. The educational level is evenly distributed among men and female. In this experiment 91,2% of the women have an education level of HBO or higher, for men it is 90,6%. These numbers are not rare for an accounting firm in which different specialized professionals are working.
The data was also tested for normality which can be found in appendix II: statistical information about the sample. The graphics indicate that the overall sample has a normal distribution.
4.2 Testing the hypotheses of the overall models
In this section the average results for the performance evaluation and growth opportunities are tested and analyzed. To do this, the average results are calculated for every scenario and for every group total. At first the average results for performance evaluation are analyzed, which are presented in table 3.
Table 3
Average results of the performance evaluation of the employee
New employee* Experienced employee* Mean* (H1)
High Result (H3) 8,43 8,10 8,27
Low result (H2) 6,05 6,04 6,05
Mean 7,33 7,02 7,17
Significance of the difference between the bold means: * P < 0,01 (two-tailed)
The average results of the performance evaluation in table 3 show that there are some differences between the main groups (high versus low results and new versus experienced employees). By examining the main average results for the high and low results group, the high
17 This means there is a difference between the groups of 2,22 points. This indicates that there could be a difference between the groups which is expected with the first hypothesis.
The first hypothesis states that the supervisors performance evaluation will be higher for employees with a high result than for employees with a low result, for both new and experienced employees. To test the hypothesis, if there is a significant difference between the high and low results of the main average results, the overall average result of the high group (8,27) and the overall average result of the low group (6,05) are tested. For testing these two groups the one-way ANOVA analysis is used. The results are shown in the table below (table 4).
Table 4
Performance evaluation: results for testing hypothesis one
Descriptive N Mean Std. deviation Std. Error mean
High 44 8,27 1,020 0,154
Low 43 6,05 1,068 0,163
Total 87 7,17 1,527 0,164
ANOVA Sum of squares df Mean square F Sig. Between groups 107,780 1 107,780 98,897 0,000
Within groups 92,634 85 1,090
Total 200,414 86
The first part of the analysis shows the descriptive results of the two groups. The lower part of the table presents the results of the one-way ANOVA. The one-way ANOVA shows a F-value of 98,897 and an significance level of 0,000 which means the P < 0,01. Therefore, the first hypothesis which indicated a difference between the two groups, is accepted and H0 for this hypothesis is rejected. There is a significant difference between performance evaluation results of a group with a high result and the group with a low result. These results are consistent with the theoretical reasoning which is described in chapter two. A high performing employee shall receive a higher evaluation from the supervisor so there is less conflict and stress in the evaluation process, less stress with the peers of the employee which avoids mutual competition and the employee feels satisfied which leads to a higher chance that s/he keeps working for the organization (Colquitt et al., 2001; Longenecker et al., 1987; Parker and Inkson, 1999).
To support the results of the first hypothesis, two additional analysis are performed. These analysis test the difference between the high and low groups with only the new employee group or the experienced employee group 2. As a deduction from the first hypothesis, it is still expected
that there are differences between the high and low results groups.
18 At first the difference between the average results of the high and low group are tested with only the new employee group. As is presented in table 3, the average score on the performance evaluation for the new employee is a 8,43 for the high result group result and a 6,05 for the low result group. To test if there is a difference between the two groups an independent samples T-test was performed. The statistical output is presented in table 10, which can be found in appendix III. The independent samples T-test shows a T-value of 7,838 and a significance level of 0,000. This means that P < 0,01 which indicates that there is a significant difference between the two groups. Next, the difference between the high and low results group for an experienced employee is tested. For the experienced employee group, the average results of the high group was a 8,10 and a 6,04 for the low group (table 3). These groups are tested with an independent samples T-test which can be found in table 11 in appendix III. The independent samples T-test for this analysis shows a T-value of 6,183 which means the P < 0,01. This means that there is a significant difference between the high and low result group when the employee is experienced. Both test results indicate that, indifferent of the years of experience of an employee within an organization, the performance evaluation by supervisors is high as long as the results of the employee are high. These two tests support the results and the theory for H1.
The last test of the overall model for the performance results is the difference between the evaluation of the new and experienced employee. The average results for the new employee is a 7,33 and for the experienced employee a 7,02, which can be found in table 3. By using these average results, the manipulation of the results is no longer present. There is no significant difference expected between the two groups, because hypothesis two and three indicate opposite directions of the hypotheses. The second hypothesis expects that experienced employees get a higher score than new employees when the overall score is low, where the third hypothesis expects a higher result for new employees than for experienced employees when the overall result is high. This means that the two hypothesis contradict each other, so there is no difference between the two average results. The overall mean of the new employee (7,33) and the overall mean of the experienced employee (7,02) are tested with an one-way ANOVA. This analysis shows a F-value of 0,854 which means the P > 0,1. This indicates that there is no significant difference between the overall average results of the evaluation of a new employee and an experienced employee. This test is in line with the prediction that hypotheses two and three contradict each other.
Within the evaluation process the participants were also asked to give a rating for the growth opportunities of the employee at the same moment the participants evaluated the performance of the employee. The average results of the experiment for growth opportunities are presented in the table on the next page (table 5).
19 Table 5
Average results of the growth opportunities of the employee
New employee* Experienced employee* Mean* (H4)
High Result 7,87 7,71 7,80
Low result 6,15 5,75 5,93
Mean 7,07 6,68 6,87
Significance of the difference between the bold means: * P < 0,01
The average results of the high and low group (7,80 and 5,93) indicate that there is a difference between the overall results of these groups. This is in line with the fourth hypothesis (section 2.3) which states that there is a significant difference between the high and low result groups, but no difference between the years of service groups. This is because management and the human resources need to satisfy the high performers in an organization, so they will not leave the company (Parker and Inkson, 1999). To test the difference for the fourth hypothesis, the average of the high result group (7,80) and the average of the low result groups (5,93) are tested with an one-way ANOVA test. The results of this test are presented below in table 6.
Table 6
Growth opportunities: results for testing hypothesis four
Descriptive N Mean Std. deviation Std. Error mean
High 44 7,80 0,765 0,115
Low 43 5,93 1,404 0,214
Total 87 6,87 1,461 0,157
ANOVA Sum of squares df Mean square F Sig. Between groups 75,659 1 75,659 59,574 0,000
Within groups 107,950 85 1,270
Total 183,609 86
The result of this analysis shows a F-value of 59,574 which means the P < 0,01. This means that there is a significant difference between the average of high and low result groups for growth opportunities. Therefore, the fourth hypothesis is accepted and H0 for this hypothesis is rejected. This means that supervisors give the high performers in an organization a significantly higher score, so the impulse for employees to leave the company decreases. As there is no relevant personal information involved in the experiment, this can be seen as an action of the supervisor to bind and motivate the employee to the organization (Kaplan and Ferris, 2011).
20 For further analysis of the average high and low results within the new and experienced employee groups, the average results are tested the same way as the performance evaluation results (with only average results of the new or experienced employee). At first the difference between the high and low result of the new employee is tested 3. The average result of the
growth opportunities for a new employee is 7,87 for the employee with a high result and 6,15 for the employee with a low result. For testing the difference between the groups the independent samples T-test is used. The statistical output is presented in table 12, which can be found in appendix III. The T-value of this test is 5,613 and gives a significance level of 0,000. This means that P < 0,01 which means that there is a significant difference between the high result group and the low result group within the new employee group. This test is also used to test if there is a significant difference between the growth opportunities of the high and low result group of experienced employees. The average result of the experienced employee for the high group is 7,71 and for the low result group 5,74. To test the difference between these two groups an independent samples T-test is used (results are added to appendix III, table 13). This test also presents a significance level of 0,000. As P < 0,01, the test shows that there is also a significant difference between growth opportunity score of an employee with a high result and an employee with a low result within the experienced employee group. The results of these tests support the above statement for accepting the fourth hypothesis, stating that supervisors give a higher rating for high performing employees. This is in line with the theory that high performers have a better chance for promotion, because organizations try to avoid employee turnover losses (Parker and Inkson, 1999).
This section tested the total average results of the models and its hypotheses. The tests show that for both the performance evaluation as the growth opportunity, there is a strong indication that there is a difference between a high performing employee and a low performing employee. Therefore, the first and the fourth hypotheses are accepted in this section. There are no indication, that on the total average results there is a difference between the new and the experienced employee. In the next section the hypothesis within the model are tested.
4.3 Hypotheses tests within the model
In this section the groups within the model are tested and analyzed which includes the second and third hypothesis. These hypothesis predict the main effects of this experiment. Testing within the model means that the average results for the performance evaluation of new and experienced employees are tested with the manipulations of a high or low result. As shown in table 3, there is an indication that there are differences between the evaluated performance for new and experienced employees when the overall results are high, but there is no real
21 indication of a difference in the low result groups. An overview of the results of the performance evaluations is given below in figure 1.
Figure 1: Performance evaluation score of new and experienced employees with a high or low result
At first the second hypothesis is tested. This hypothesis states that supervisors performance evaluation will be higher for relatively experienced employees when the overall performance is low, than for the new employees. To test this hypothesis the following groups are used: new versus experienced and an overall manipulation for the results which is low. The performance evaluation process of the experiment gave the experienced employee with a low result a mean of score 6,04 and the new employee with a low result a mean score of 6,05. To test the hypothesis, the independent sample t-test is used. The tested variable is the supervisors score on performance and the grouping variable is the scenarios new/low and experienced/low. The results are presented in the following table (table 7).
Table 7
Performance evaluation: results for testing hypothesis two
Group statistics N Mean Std. deviation Std. Error mean New employee (low) 20 6,05 1,050 0,235
Experienced employee (low) 23 6,04 1,107 0,231
Levene’s test Independent samples T-test
F Sig. T df Sig.
Equal variances assumed 0,005 0,944 0,020 41,000 0,984
22 The first part of the analysis shows the descriptive results of the two groups. The lower part of the table presents the results of Levene’s test and the independent samples T-test. Levene’s test s used to test if it is possible to assume the variances are equal. The significance score of 0,944 gives a P > 0,1, which means there is no significant difference between the variances of the groups. Therefore equal variances are assumed, which means the independent samples T-test results are used from that row. The result of this independent sample T-test shows a T-value of 0,02 and a significance score of 0,984, which means the P > 0,01. Despite the fact that the hypothesis predicts only a one tailed result, if the significance score is divided by 2 it still is not enough to prove the hypothesis. This means the second hypothesis, which stated that experienced employees receive a higher performance evaluation than new employees when the results are low, is not accepted and H0 is not rejected. This indicates that both groups get an equal evaluation from supervisors, which means that employees receive a low performance evaluation result which is independent if they work 11 years or 11 months for the company.
The above test shows that if the overall results of the employees are low, a new or experienced employee receives the same performance evaluation. This is against the reasoning that supervisors will compensate the experienced employee more because of past performance results. This gives an indication that not all information available is used in the performance evaluation as was mentioned by Bond et al. (2007) and Bol & Smith (2011). The tests above indicate that when the overall performance is low, past performances have likely enough no effect on the performance evaluation of employees.
Next the third hypothesis is tested which predicts that the performance evaluation of the supervisors will be higher for more new employees than for experienced employees when the overall performance is high. This hypothesis is tested with again the two groups, new versus experienced employees, but this time with an overall manipulation where all results are high. The performance evaluation process gave the experienced employee with a high result an average score of 8,10 and the new employee with a high result a mean score of 8,43. To test this hypothesis, the independent sample t-test is used. The tested variable is the supervisors score on performance and the grouping variables are the scenarios new/high and experienced/high. The results are shown in the table on the next page (table 8).
23 Table 8
Performance evaluation: results for testing hypothesis three
Group statistics N Mean Std. deviation Std. Error mean New employee (high) 23 8,43 0,945 0,197
Experienced employee (high) 21 8,10 1,091 0,238
Levene’s test Independent samples T-test
F Sig. T df Sig.
Equal variances assumed 0,008 0,927 1,106 42,000 0,275
Equal variances not assumed 1,099 39,806 0,279
Levene’s test gives a significance score of 0,927, so equal variances are assumed. The result of this independent sample T-test shows a T-value of 1,106 and a significance score of 0,275. The hypothesis predicted a one sided effect which is that the performance evaluation result of a new employee is higher than the performance evaluation result of an experienced employee. As this is a two tailed test and the hypothesis only focusses on one side of the distribution, the significance level can be divided by 2. This gives a significance level of 0,1375, which still means that the P > 0,01 and there is no significant difference between the groups. This means the third hypothesis, which stated that new employees receive a higher performance evaluation than experienced employees when the results are high, is not accepted and H0 is not rejected.
As the desired significance level has not been achieved, the test shows that there is no significant difference between the two groups. This indicates that employees receive the same results from supervisors when the performance is high, despite the manipulation of the years of service for the organization. Unfortunately the test did not give evidence for the theory, but that does not mean there is no effect all together. The difference between the two groups of 0,33 is marginal, but there is a small difference. There can be several reasons why the test was not significant, maybe there arise greater differences in the performance evaluation when an employee exceeds the 100%, or when the experiment is executed in a different environment or with another scenario.
In this section hypothesis two and three were tested. The tests presented no indication that there is a difference between the performance evaluation of a new or experienced employee. Therefore, the second and third hypotheses are not accepted. The following section analyzes the reversed results of the experiment.
24 4.4 Testing reversed results
After the participants gave a score for the performance evaluation and growth opportunities, a new screen appeared with a reversed question for the performance evaluation. Participants that were in a scenario for a new employee received the question what score they would give the employee if the employee had been in the company for 11 years. The same was done for the participants in the scenario with the experienced employee, they were asked what score they would give an employee that had been in the company for 11 months. For this thesis these results are called the reversed results. The results of the reverse performance evaluations are shown in the next table (table 9).
Table 9
Average results of the performance evaluation of the employee
From months to years From years to months Mean*
High Result** 7,70 8,52 8,09
Low result* 5,25 6,65 6,00
Mean* 6,56 7,55 7,06
Significance of the difference between the bold means: * P < 0,01 / ** P < 0,05
In this section the reversed scores for the performance evaluation process will be analyzed. At first the average results of the overall model are tested, which means that the average results of the high and low result are tested and the average results for the groups that evaluated from years to months and from months to years. The statistical results have been added to appendix IV.
For testing the difference between the total average result of the high group (8,09) with the total average result of the low group (6,00) the one-way ANOVA test was used (table 14 in appendix IV). As expected there is a significant difference between the groups with a high and low result. These are expected because results emerged during the analysis of the normal performance evaluation process and the growth opportunities.
Next the overall average results of the groups that gave an evaluation from months to years (new to experienced) and from years to months (experienced to new) are analyzed. From this point a new kind of experiment started. On this moment in the experiment the participants had already evaluated an employee based on the knowledge that the employee was 11 months or 11 years working within the organization. Now they had the opportunity to compare two employees with the same result, but with different years of experience within the organization. This comparison leads to a significant difference between both groups of almost 1 full point as is shown in table 9. The difference between the groups are tested and the statistical results have also been added to appendix IV. Several tests are performed to analyze the difference between the from months to years and from years to months groups. Table 15 presents the
25 results of the overall average results between the groups, table 16 presents the difference between the groups when the overall result is high and table 17 presents the differences when the overall result is low. The tests shows significance levels between the 0,003 and 0,012 which mean all tests are significant with at least P < 0,05. This means there is a strong indication that there is a difference between the two groups (months to years and years to months).
The participants that gave a performance evaluation for the from months to years group first evaluated a new employee, so they now evaluated an experienced employee. As shown in table 9, participants that evaluated a new employee first (months to years) gave a lower evaluation to experienced employee and participants that evaluated the experienced employee first (years to months) gave a higher evaluation for the new employee. This indicates that the moment participants had to compare a new employee with an experienced employee, they gave a better performance evaluation to the new employee. The test of table 15, 16 and 17 in appendix IV support these results. This means that there is an indication that new employees receive a higher performance evaluation score when comparison is possible.
These results can be explained by the theory that new employees receive lower expectations than experienced employees. As Longenecker et al. (1987) earlier stated, all information available to a supervisor will be used in the performance evaluation. Since the base line of a new employee can’t be properly determined for the time being and the future results of an experienced employee can be better estimated, the supervisor gets positively biased to the new employee. This may have to do with motivating the new employee and create more job happiness to reduce employee turnover (Parker and Inkson, 1999; Shahnawaz and Jafri, 2009). In this case the supervisor chooses for a more uncertain but possible higher result, than for more secure results which will probably not surprise. As supervisors are risk-averse, the real risk shall not be about the uncertain results of the employee in the future, but about keeping the employee turnover as low as possible.
4.5 Supplemental analyses
To examine the possibility that the results are driven by the participants with no evaluation experience, a robustness check was performed. Therefore, the results of hypothesis three and four and the outcome of the reversed results are analyzed with participants that have at least one year of experience in the evaluation process 4.
The results for analyzing hypothesis three and four with experienced participants show no difference with the results described above (see also table 18 and 19). As hypothesis three stated a difference between the performance evaluation of a new and experienced employee when the overall results are high, there is still no significant difference between these two groups. This indicates that the earlier conclusion for rejecting H3 is still correct with only
26 experienced participants. The difference between the groups is even smaller. Furthermore, testing hypothesis four again with experienced participants shows there is still a difference between the growth opportunities between the groups that scored a high or low result. Since the new average results on this subject even have increased, accepting H4 and rejecting the H0 is still correct. These results indicate that scores on performance evaluation and growth opportunities are not driven by the years of service for the same organization.
Also, the difference between the months to years and years to months groups is tested with experienced participants. The tests in section 4.4 present a significant difference between the groups, but when only experienced participants are used the results are no longer significant (test is added in appendix IV, table 20). Even though there is still a difference between the two groups of 0,69 in favor of the new employee, the difference between the two groups has become smaller. This can indicate that experienced supervisors show a bit more compression when they are able to compare two employees. This indicates that to accept that there is a possible difference between the two groups when compared, further research is needed.
4.6 Evaluation of the experiment
At the end of the experiment the participants were asked how they experienced the ease and clarity of the case-based experiment. Both questions were measured on a scale from one to five. For ease, the participant was asked how they experienced the case from very easy to very difficult. The average score on this question was a 2,5, which means the most participants experienced the case easy to indifferent. Next the clarity of the experiment was measured, from very unclear to very clear. The average score was a 3,5 which means that most participants where indifferent about the clarity or thought the case was clear for them. There are only two real outliers in the dataset, at which participants probably didn’t read the question well enough. This is possible because the questions were written in the opposite direction of each other. If the outliers are not used in the analyses, the outcome is still the same. This indicates that the experiment was clear and easy enough for the participants.
27
5 Conclusion
The subject of this research was to verify whether the number of years of service at the same company has an influence on the performance evaluation of an employee. This question emerged when evaluating the differences in subjective performance evaluations research (Bol et al., 2016; Bol and Smith, 2011; Longenecker et al., 1987) and the effects of job-hopping (boundaryless career) (Arthur, 1994; Arthur et al., 2005).
The theory indicates a difference between the two groups which is associated with different expectations of employees by supervisors and compression in combination with laziness of supervisors in performance evaluations (Longenecker et al., 1987; Colquitt et al., 2001). In accordance with Bol and Smith (2011) a different approach is expected for employees with a high or low result. Furthermore, a difference is expected between a new and experienced employee. As the overall result of employees is low, it is expected that there is more compensation for the more experienced employee above the new employee in the performance evaluation. This is because of past the knowledge of the employee (a more reliable baseline can be expected) and the risk aversion of supervisors. When the overall result of employees is high, it is expected that a new employee receives a better performance evaluation score than the more experienced employee because of the higher exceeding expectations and binding the employee to the organization. Binding the employee to the organization is important to avoid the different costs which are associated with a high employee turnover (Becker, 1964; Tziner and Birati, 1996; Campbell et al., 2012).
To test the above reasoning an experiment was conducted with participants from an accounting firm in The Netherlands. Through the experiment there is no support found for this reasoning. Despite the close results for the manipulation when the overall performance is high, there is no indication in the experiment that employees with different years of experience are treated differently. From these results this thesis indicates that even though job-hopping has still no positive association with holding a job, job-hopping has no significant effect on the performance evaluation. This could be a trend in the economy, in which job-hoppers are more widely accepted.
An important aspect of the human recourse process are promotions (Kaplan and Ferris, 2001). Promotions can be a good way to keep the high performing employees within the organization and reduce the employee turnover (Parker and Inkson, 1999). Therefore, it is expected that a high performing employee has a higher rating for growth opportunities than a low performing employee. Another reason that a high performing employee receives a higher rating for growth opportunities than their peers, is to avoid stress and anger of the high performing employee (Colquitt et al., 2003). The results of this thesis indicates that high performing employee indeed receive a higher rating on growth opportunities. It did not matter
