Subjective Performance Evaluation and Trust: Perceptions of Distributive Justice

(1)

Subjective Performance Evaluation and Trust:

Perceptions of Distributive Justice

Master Thesis Organizational & Management Control

Student Daisy de Calonne

S3541347

University of Groningen Faculty of Economics and Business MSc BA: Organizational & Management Control

Supervisor Dr. van Veen – Dirks

(2)

2

Abstract

This study contributes to existing literature about the use of subjective performance measures in evaluation systems and the perceived justice of employees on the outcomes of their evaluations. It provides more insight into the differences between several subjective performance measures and their influence on perceived distributive justice. The role of social context in the performance evaluation system was tested by including trust as a moderating variable to the relationship between subjective performance measure and distributive justice. The study was conducted to find answers on two questions: ‘Does subjective performance evaluation affect the perceived distributive justice about the outcomes?’ and ‘In what extent does the level of trust in the supervisor influence the relationship between subjective performance evaluation and perceived distributive justice?’. Using data of 56 employees of multiple organizations, it was found that two indicators of subjective performance measure had a significant effect on the level of perceived justice on promotion distribution: personal-related performance measure negatively affected the perceived justice and work-related performance measure positively affected the perceived justice. However, the subjective measures did not significantly affect all perceived justice, there was no significant relationship found between overall perceived justice on rewards and subjective performance measures. The first hypothesis was partially supported. No evidence was found for the positive moderation of trust on the relationship between the use of subjective performance measure and distributive justice, it was only found that a lower level of objective measure did significantly affected the level of perceived justice on promotion distribution. The second hypothesis was not supported.

(3)

3

1. Introduction

Supervisors in firms often use performance evaluation systems to assess the performance of employees and to determine their compensation and decide which employees deserve promotions. Many of these evaluation systems include objective performance measures because these measures are easily observable, such as quantity of sales. Although these measures can be observed, other type of evaluations might also be included when objectivity does not allow supervisors to assess all relevant actions of the performance and abilities of employees (Bol & Smith, 2011). Including subjective performance evaluation in these systems is desirable, according to the informativeness principle that states that subjective evaluations should be implemented because they add more information about individual’s actions and behaviour. Subjective performance measures improve the contract between a supervisor and employee (Holmström, 1979). Therefore, many supervisors use subjective performance evaluation in the evaluation systems (Grund & Przemeck, 2012).

While both objective and subjective performance measures are widely used in practice by firms, the latter is more studied by researchers. Studies found evidence that subjective performance evaluation is useful to enhance the alignment of incentives (Baker, Gibbons, & Murphy, 1994) though others found disadvantages about subjective performance evaluation (Gibbs, 2008; Moers, 2005). The reason why researchers focus on subjective performance evaluation is the strong dependency on supervisor’s perceptions when these evaluations are used (Bol & Leiby, 2018). Subjective performance evaluation is associated with potential favouritism, leniency, strictness and unintentionally bias. Leniency bias indicates that supervisor evaluate the performance of the employee higher than objective measures would justify, while strictness bias leads to lower compensation or performance ratings for employees with higher performance than other (Prendergast & Topel, 1993). The positive results for employees when leniency or unintentionally bias occurs could result in positive perceptions from employees on the outcomes of their evaluations, while the negative results when strictness bias of favouritism occurs could result in negative perceptions. These perceptions are described as distributive justice. In this study, it is examined if the use of subjective performance measure have an influence on the perceived distributive justice of employees. This leads to the first research question:

(6)

6

Previous research showed that social context plays a crucial role on the reaction of employees to the outcomes of their evaluation process (Levy & Williams, 2004). Researchers propose that trust is crucial in the relationship between supervisor and employee, especially when the supervisor plays a substantial role in the performance evaluation process. Supervisors play a more important role in performance evaluation when they use subjective measures, because their perceptions are needed to evaluate their employees. The importance of trust in the perceived justice was examined by Mulvaney (2017). He found that the perceived justice of employees on the performance evaluation depends on the level of trust an employee has in their supervisor. Research showed that the higher the level of trust, the higher the perceived justice will be on the performance evaluation process and outcomes (Mulvaney, 2017; Cho & Lee, 2012). These studies show that trust is important when the supervisor uses subjective performance measure, due to the high involvement in the performance evaluation and that trust also ensures that employees perceive the outcomes of this performance evaluation higher. This leads to a potential moderating role of trust in the relationship between subjective performance evaluation and distributive justice. This results in the following research question:

Does trust in the supervisor positively moderates the relationship between subjective performance measures and perceived distributive justice?

(7)

7

measures that have the potential to lead to biases. The research focuses on 1) the effect of subjective performance measure and the effect on perceived distributive justice, and 2) the positive role of trust in this relationship.

(8)

8

2. Literature Review

It is important to give insight in the boundary condition of this research before presenting the variables used in the study. This study focuses on the relationship between supervisor and employee from the employee’s perspective. Distributive justice is explained based on how employees assess the level of fairness on the outcomes of performance evaluation. Trust in this study explains the level of trust an employee has in their supervisor.

2.1 Agency Theory

Rewards and promotions determined by supervisors often rely on observable objective performance indicators and subjective evaluations (Campbell, 2008). Objective performance indicators are “only informative about the measurable aspects of an employee’s job and provide no incentives for the more qualitative aspect, like cooperation and innovation” (Moers, p. 69, 2005). A greater use of subjective performance measures within the performance evaluation is necessary when supervisors are not able to obtain objective indicators on an individual level to assess the performance of their employee (Grund & Przemeck, 2012). Subjective performance measures are in this study defined as “superior’s subjective judgment about qualitative performance indicators” (Moers, p. 68, 2005). This shows that subjective performance measures are more informative about the qualitative aspects of an employee’s job. Hence, higher levels of subjective measures is usual when objective indicators are incomplete, when performance is difficult to measure alone with objectivity, or when objective indicators are susceptible to data manipulation (MacLeod & Tan, 2016; Gibbs, Merchant, Stede, & Vargus, 2004). Subjective measures are also used to filter out the uncontrollable factors, because these type of factors are perceived as unfair by employees when they can potentially decrease their performance ratings (Bol & Smith, 2011). Subjective measures are used to protect the risk for employees from receiving lower compensation, and to protect the firm from decreasing job satisfaction and motivation.

(9)

9

principal and agent, it allows more accurate judgments about the performance of an agent. Informativeness principle states that any observable performance measure that provides more information about the real actions and efforts of employees should be taken into account when compensation is determined, the use of observable information improves the contract between principal and agent and leads to greater effort intensity (Moers, 2005; Holmström, 1979). Thus, informativeness principle confirms the idea that supervisors should also implement subjective performance measures in their evaluation process.

2.2 Subjective Performance Measures

Subjective performance measures inform supervisors about the qualitative aspects of an individual’s work. Eccles and Crane (1988) studied how compensation of employees were determined at several banks, they found that compensation relied on a high level of subjective performance measures, such as the “quality of deals, the contributions to customer satisfaction, training of younger associates, and marketing” (p. 166). Moers (2005) stated that the quality of an employee’s product or service, good use of resources and adequate planning are examples of qualitative subjective performance measures.

Though subjective measures protect employees from uncontrollable objective indicators, it could also lead to potential bias. Qualitative performance measures increase the likelihood that supervisors let their preferences define the outcomes of the evaluation, which can influence the allocation of rewards (Moers, 2005). The perception of supervisors often determine these evaluations and outcomes, due to the fact that these perceptions are needed to score the qualitative aspects of performance, which is not the case when quantitative measures are used. The perceptions of supervisors lead to problems when the evaluation leads to distorted performance ratings.

(10)

10

affect the future career of the employee, which shows that bias plays an important role in the determination of rewards and promotions (Prendergast & Topel, 1993; Mathis & Jackson, 2010). The problem for firms when leniency bias or favouritism occur, is that it becomes difficult to differentiate between high and low performers. Leniency bias may lead to higher rewards for employees with low performance, while favouritism could also lead to lower rewards for high performers (Prendergast & Topel, 1993).

Several agency theorists showed that supervisors with residual claimants have incentives to under-report the performance of employees, because this reduces compensation costs (Baker, Gibbons, & Murphy, 1994). This is called strictness bias. This occurs when preferences of supervisors lead to lower compensation for employees with actual high performance. Strictness bias may lead to lower rewards for employees that they would have received when objective performance measures were only used in the appraisal system. It can cause inefficiencies for both parties: employees may feel discriminated and quit their job, and the firm could have higher turnover costs and a loss of human capital (Lunenberg, 2012).

Unintentional bias could also occur when supervisors are unintentionally influenced by the objective measures. This happens when the supervisor has knowledge about the quantitative ratings before evaluating on the subjective measures, individuals often tend to process information in a way that favours a previously held belief, desired outcome or preference (Bond, Carlson, Meloy, Russo, & Tanner, 2007). They may unintentionally adjust the subjective measures to the level of objective measures (Bol & Smith, 2011). These theories suggest that leniency bias, strictness bias, favouritism and unintentional bias could lead to lower efforts of employees when they are evaluated lower than expected, or higher efforts when they are evaluated higher than expected (Lunenberg, 2012).

2.3 Justice Perceptions of Employees

(11)

11

2.3.1 Perceived distributive justice. Bol and Smith (2011) indicated that employees will perceive the evaluation measures and outcomes as unfair when their ratings are not as high as expected, or when they do not record their actual efforts. This study focuses on the fairness perceptions of employees toward their evaluation outcomes, which is called distributive justice in research. The study focuses on this distributive justice because subjective performance measures could lead to biased performance ratings and hence, different outcomes for employees than they should have received. The perceptions of employees toward these outcomes are important, because it can affect motivation and job satisfaction (Taylor, Tracy, Renard, Harrison, & Carrol, 1995; Cropanzano, Bowen, & Gilliland, 2007). Distributive justice refers to “the distribution of the returns from decisions to a group of individuals” (Gillenkirch & Kreienbaum, p. 939, 2016). These returns refer to compensation or promotions, determined by the evaluation systems and measures supervisors use. Distributive justice describes the view of individuals on whether they received the outcome they deserved (Folger & Konovsky, 1989). Most theorists describe distributive justice according to equity theory, this theory requires the allocation of rewards to be related to the actions of individuals within the group. This theory states that individuals believe that the rewards they receive, should be distributed according to the level of contribution (Adam, 1965; Homans, 1974). Employees may have injustice perceptions toward the outcomes when they expected higher outcomes, and when they believe that they did not receive sufficient rewards according to their contributions (Crosby, 1976; van den Bos, Vermunt, & Wilke, 1997).

(12)

12

2.3.2 Subjective performance measures and distributive justice. Many scholars studied the consequences of injustice and violations of distributive justice. Bol (2011) stated that biases from performance evaluation influences the perceived justice about the appraisal system. It was illustrated that bias changed the outcome distribution of a compensation plan, which suggest that bias in performance evaluation affect the rewards or promotions employees receive. Subordinates consider more subjective feedback and measures from their supervisor as more informative and fair than objective quantitative measures (Bol & Smith, 2011). Bol and Smith (2011) indicated that supervisors in their study distorted information about employees’ performance when they wanted to be consistent with the objective measures. Supervisors in their study would asymmetrically adjust the subjective performance measures when an uncontrollable objective factor could result in an unfair outcome for employees. These evaluations resulted in evaluation outcomes affected by leniency and unintentional bias. Strictness bias did not occur, the supervisors would only adjust the evaluations upward, they did not adjust downward when objective uncontrollable measures resulted in favourable outcomes for employees. These results of Bol and Smith (2011) indicated that bias occurred when supervisors wanted to increase the perceived justice, strictness bias did not occur because this would result in decreasing perceived justice.

Despite these positive results of biases toward perceived justice, other theories did indicate that supervisors can downward the evaluations. Iqbal, Akbar, & Budhwar (2015) showed that biases can cause negative outcomes for employees, which results to employee dissatisfaction and low organizational commitment. This shows that strictness bias or favouritism could still occur, even when it will affect the perceived distributive justice. It is therefore predicted that biases resulting from subjective performance measures, influence the perceived distributive justice of employees. Both negative and positive effects of biases toward subjective performance measures were explained. These different effects resulted in the following null hypothesis:

(13)

13 2.4 The Moderating Role of Trust

Previous literature has suggested that social context plays a role in the performance evaluation process. Levy and Williams (2004) indicated that social context plays an important role in how employees will react to the evaluation process and outcomes. Research suggests that trust is critical in the relationship between supervisor and employee, it is suggested that employee’s trust in the supervisor is critical in situations where the supervisor plays a substantial role throughout the performance evaluation process, which is the case when supervisors use subjective performance measures (Cho & Lee, 2012; Dirks & Ferrin, 2002). Trust can be defined as “the willingness of a party to be vulnerable to the actions of another party, based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to control or monitor that party” (Mayer, Davis, & Schoorman, p. 712, 1995). In other words, if an employee expects that the supervisor has skills to properly evaluate them and believe that they would not use their control power to change the standard or values, they are more likely to trust the supervisor. The supervisor-employee relationship is different than other relationships in organizations, because supervisors have higher status, the power to exercise control and employees often depend on supervisors’ evaluations to receive promotions, pay raises and job security (Sitkin & Roth, 1993).

Mulvaney (2017) compared perceived justice among employees with low and high trust in their supervisors. Results suggest that employees with higher trust have higher procedural and distributive justice perceptions toward the appraisal system and outcomes, in comparison with employees with lower trust in their supervisor. This shows that higher levels of trust will increase the acceptance of the performance management system (Condrey, 1995). This supports the Social Cognitive Theory that indicates that employees with higher levels of trust in their supervisor will presumably react more favourable to the appraisal system and outcomes (Cho & Lee, 2012). The opposite is true when levels of trust are low, employees will react more negatively toward the system when they do not trust their supervisor (Mulvaney, 2012). Trust is therefore an important factor that influences the perceived justice of employees when they are evaluated by their supervisor. This resulted in the following hypothesis:

Hypothesis 2. Trust positively moderates the relationship between subjective performance measures and perceived distributive justice in such a way that the relationship between

(14)

14 Subjective Performance Measures

on an individual level

Perceived Distributive Justice from the employee perspective

Trust

from the employee perspective 2.5 Conceptual Model

(15)

15

3. Methodology

The research design of this study was deductive of nature to answer the research questions and to test the hypotheses. Quantitative data was collected by extracting a questionnaire among employees from different firms.

3.1 Data Sample

Data for this study was collected from multiple companies. They were targeted by telephone or e-mail and received a letter, which contained information about the objective of the research and the design of the questionnaire. Several companies were unable to cooperate, which leaded to changing the focus from firms to individuals. A total of 70 individuals participated in the questionnaire via an anonymous link on the system Qualtrics. Response analysis has been done in order to check the data on missing values, the final sample consisted of 56 respondents due to a large extent of missing values and non-usable responses. This indicates that 80% of the responses were usable for the study.

(16)

16 3.2 Data Collection

The questionnaire was designed with three other students with the same thesis topic. 56 items were measured, which conclude the items that were used by the other students to measure their independent variables. Hence, this study consisted of 29 items that were used to test the hypotheses. These items measured the dependent, independent and control variables.

In order to collect data, the questionnaire was sent to multiple companies and individuals to distribute the link. Anonymity was guaranteed by distributing an anonymous link, different links were designed for each company. Respondents were informed about the research in the introduction of the questionnaire. The introduction explained the objective of the study, confidentiality and their contribution to the research. All respondents were Dutch while the items were written in English, this meant that the items had to be translated in Dutch to ensure respondents could complete the questionnaire in their own language. Errors were removed by checking the items multiple times by all four students.

Difficulties were experienced during the data collection process. It appeared that most targeted companies perceived the questionnaire items as too sensitive, because the items were focused on employees’ perceptions, satisfaction and the trust in the supervisor. One company stated that the measures in the questionnaire were not used in their evaluation system. Seven firms did not or could not distribute the questionnaire, which are firms as: KwadrantGroep, Lode, ProCare, Axtion, Umaco, HTG and ABN Amro. Several contacts of companies that could not distribute the questionnaire, completed the questionnaire themselves. This leaded to a small response number from these companies, ranging from one to five responses per company. Examples of these companies are: Provincie Friesland, Meacasa and Voestalpine Polynorm. The responses from the contacts in these companies are grouped within the ‘others’ group. More response were expected when these companies were contacted, which resulted in disappointment when they were not able to distribute the questionnaire. Therefore, the sample size was limited due to the previous mentioned reasons.

3.3 Measurements of Variables

(17)

17

3.3.1 Distributive justice. The dependent variable distributive justice has been measured by the extent to which respondents were fairly rewarded. Ten items were used to measure the variable. The first item measured perceived justice on the performance ratings (see Table 1) on a five-point Likert-scale (1 = completely disagree, 5 = completely agree) based on the research of Harris (1988). Six items measured distributive justice on rewards based on the Distributive Justice Index of Price and Mueller (1986). The statements were measured to indicate the extent to which respondents were fairly rewarded, the items were measured on a five-point Likert-scale (1 = very unfair, 5 = very fair). Three items measured the perceived justice on promotion distribution (Witt and Nye, 1992). The statements were measured on a five-point Likert-scale (1= completely disagree, 5= completely agree) on which the employees indicated their agreement with the statements 8, 9 and 10 in Table 1.

An exploratory factor analysis using principal components extraction was conducted to determine the factor structure of the variable. The scree plot and eigenvalues indicated a two-factor solution that explain 61,71% of the variance. The two-factors were grouped in perceived justice on rewards (factor 1) and perceived justice on promotions (factor 2). Table 1 shows the factor loadings of the items from the highest to the lowest loadings. The perceived justice on rewards construct had a Cronbach’s alpha of 0.86, the perceived justice on promotions construct had a Cronbach’s alpha of 0.76. The constructs reached acceptable reliability despite the small sample size.

Table 1. Factor analysis with varimax orthogonal rotation

Factor Loading Factor 1: Distributive Justice Rewardings (ɑ = .864) 1 2 6. I am fairly rewarded for the the amount of effort I put in my work. .90 .00 7. I am fairly rewarded for the performance I deliver. .84 .15 5. I am fairly rewarded for the stressed and strains of my job. .80 .22 2. I am fairly rewarded for the responsibilities I take on. .70 .48 3. I am fairly rewarded considering the amount of experience. .52 .44 4. I am fairly rewarded considering the amount of education and training* .42 .57 1. My performance rating for this year represents a fair picture of my job

performance.

.40 .49

Factor 2: Distributive Justice Promotions (ɑ = .758)

(18)

18

10: Promotions in the organization are given to those who are best qualified. 8. The promotions that workers receive here are based on what the worker puts into his/her work.

-.06 .26

.75 .74

Note. KMO = 0.78. * Items 1 and 4 have higher loadings on factor 2, they are included in factor 1 due to more similarities in literature than the other variables.

3.3.2 Subjective performance evaluation. Eight items measured the level of

subjective performance measure used in the evaluation system. Five items on a five-point Likert-scale (1= completely disagree, 5= completely agree) were included to measure which subjective performance measures were used in the evaluation (see Table 2). Respondents indicated their agreement with the extent to which the subjective measures are used in their evaluation. The level of subjective performance measures was also measured by the level of objective performance measures1. Respondents indicated their agreement with the extent to which they were evaluated on the objective measures in Table 2. (Moers, 2005; Baker et al., 1994).

An exploratory factor analysis using principal components extraction was conducted to determine the factor structure. The scree plot and eigenvalues indicated a three-factor solution that explain 67,54% of the variance: personal-related subjective measures (factor 1), work-related subjective measures (factor 2) and objective measures (factor 3). Table 2 shows the factor loadings of each item on the three constructs. The first factor consisted of the items that measured involvement and reliability of the employee, these items are more related to personal behavior. Quality, level of innovative solutions and use of sufficient resources were measures that were grouped in the second factor. These items are more work-related than the other subjective measures. The last factor consisted of the objective measures. For each factor, Cronbach’s alpha was calculated based on the grouping of items in table 2. The first construct had a Cronbach’s alpha of 0.74, the second construct had a Cronbach’s alpha of 0.62, and the last construct had a Cronbach’s alpha of 0.57. The second and third constructs have lower alpha’s, though both do not increase when items in the construct are deleted. These lower values are according to Taber (2017) and van Griethuijsen et al. (2014) valued as ‘moderate’ (construct 2) and ‘acceptable’ (construct 3).

(19)

19 Table 2. Factor analysis with varimax orthogonal rotation

Factor Loading Factor 1: Personal-related subjectivity (ɑ = .737) 1 2 3

8. Subjective measure: reliability. .87 .01 -.18

7. Subjective measure: involvement. .86 .09 .18

Factor 2: Work-related subjectivity (ɑ =.619) 5. Subjective measure: sufficient use of resources. 4. Subjective measure: quality of service.

6. Subjective measure: contribution to innovative solutions*.

-.08 .27 .57 .84 .66 .56 -.03 -.04 .08 Factor 3: Subjectivity relation to objectivity (ɑ = .571)

3. Objective measure: quantity of sales/customers/products. 1. Objective measure: performance compared to budget. 2. Objective measure: percentage of absence.

.00 .20 -.18 -.67 .02 -.17 .44 .81 .76 Note. KMO = 0.67. * Table 2 shows that ‘innovation’ had the highest loading on factor 1, though the item was not included in the first factor, loadings between factor 1 and 2 differ only .011. Innovation has more similarities to the other items, which results in a higher Cronbach alpha when this item is included.

3.3.3. Trust in the supervisor. The items to measure trust in the supervisor were based on the Interpersonal Trust Scale of Read (1965). The level of trust has been measured using four items on a five-point Likert-scale (1= not at all, 5= to a very high extent). Table 3 shows the items used to measure the level of trust.

(20)

20 Table 3. Factor analysis with varimax orthogonal rotation

Factor Loading

Factor 1: Overall trust (ɑ = .806) 1

4. My supervisor’s decisions are justified by other considerations whenever they make decisions against my interests.

.93

1. My supervisor will take advantage of opportunities to further my interests. .80 3. My supervisor will keep me fully informed about things that might concern me. .79 2. I feel free to discuss problems with my supervisor. .68 Note. KMO = 0.70.

(21)

21

4. Results

The descriptive statistics give an insight in the data of the sample. Correlations between variables are showed to indicate how the variables used in the study effect each other. The hypotheses were tested using multiple linear regression analysis.

4.1 Descriptive Statistics

Table 4 shows the descriptive statistics of the variables (observations, means, standard deviations, observed range, possible range). As shown in the table, most measures range from the minimal to the maximal value which indicates that the answers of respondents varied in this sample. The descriptive statistics also show that the total observation number is 56, though there exists one missing value for supervisor tenure. Furthermore, the mean score for perceived justice on rewards is higher than the score for perceived justice on promotions. This indicates that employees perceived the allocation of rewards as more fair.

Table 4. Descriptive statistics of the variables

N Mean SD Observed range Possible range Predictor variables

Distributive justice: rewards 56 3.35 .73 1.86-4.86 1-5 Distributive justice: promotion 56 3.08 .80 1.00-4.67 1-5

Subjectivity: personal 56 3.99 .63 2.00-5.00 1-5 Subjectivity: work 56 3.54 .69 1.00-4.67 1-5 Subjectivity: objective 56 3.35 .84 2.00-5.00 1-5 Trust 56 3.50 .70 1.50-4.75 1-5 Control variables Gender 56 1.23 .43 1-2 1-2 Supervisor tenure 55 3.47 5.00 0-23 Na Education level 56 4.00 1.68 1-7 1-7 Age 56 2.98 .65 2-5 1-5 Company size 56 3.46 .87 1-4 1-4

(22)

22

Table 5 shows correlation between the dependent and predictive variables. It can be noticed that there are several significant relationships. Perceived justice on rewards is positively significant with perceived justice on promotions (r = 0.52, p < 0.01) which is expected because both measure the level of distributive justice. The table shows that the relationships between the subjective performance measure constructs differ, personal-related measure only has a significant positive relationship with related measure (r = 0.11, p < 0.05) while the work-related measure construct had a significant negative relationship with the objective construct (t = -0.30, p < 0.05). The negative correlation between these variables was unexpected, because both should measure the level of subjectivity. Trust had a significant relationships with both perceived justice constructs, respectively reward (r = 0.31, p < 0.05) and promotion (r = 0.50, p < 0.01).

Table 5. Pearson Correlation analysis (N=56)

1 2 3 4 5 6

1. Perceived justice: reward 1

2. Perceived justice: promotion .516** 1

3. Subjectivity: personal .107 .002 1

4. Subjectivity: work .050 .298* .334* 1

5. Subjectivity: objective .308* -.072 .006 -.302* 1

6. Trust .306* .502** .128 .394** -.201 1

Note. * p <.05, ** p <.0

4.2 Regression with Perceived Justice: Rewards

Multiple linear regression analysis was performed to examine the relationships between the dependent, independent and control variables and to test the hypotheses. Using the enter method, six control variables were entered in step 1, subjective measure constructs were entered in the second step, trust was entered in step 3, and the interaction of trust and subjective performance measure was entered in the last step. The regressions in this chapter include the perceived justice on rewards as dependent variable.

(23)

23

found that subjective performance measure constructs were not significant predictors of perceived justice on rewards. Further examination showed that none of the independent variables were significant predictors. This indicates that subjective performance evaluation did not significantly positively or negatively affect perceived justice on rewards, which supports the first hypothesis.

4.2.2 Trust. The variable that measured trust was included in the third step of linear multiple regression as an independent variable to test if the variable predicted distributive justice. The results indicated that there was a significant effect between the control variables, subjective evaluation, trust and distributive justice (R² = 0.338, F(10,45) = 2.296, p < 0.05). Trust explained 8,6% in the variance of distributive justice and was the only significant predictor of distributive justice (β = 0.348, t = 2.415, p < 0.05). These results show that trust in the supervisor had a positive effect on distributive justice. The interaction of subjectivity and trust was included in step 4 to test if trust positively moderated the relationship between subjective performance measures and distributive justice. The results indicated that the interaction did not significantly predict distributive justice. These results suggest that there does not exists a positive moderating role of trust between the independent and dependent variables, which does not support the second hypothesis.

4.2.3 Sub-regressions. Multiple linear regression was used to test the effect of each subjective measure construct on perceived justice on rewards, without the inclusion of the other constructs. First, table 7 show that personal-related subjective measures did not significantly predict distributive justice. This suggests that this type of subjective measures does not affect the perceived justice on rewards. Results show that the model did significantly effect distributive justice when trust was included (R² = 0.303, F(8,47) = 2.555, p < 0.05). Further analysis showed that trust (β = 0.330, t = 2.584, p < 0.01) did significantly predict distributive justice. However, the interaction of personal subjective measures and trust did not significantly affect the perceived justice on rewards.

(24)

24

rewards. It was also found that the interaction of work subjective measure and trust did not significantly effect perceived justice.

Table 6. Multiple linear regression on distributive justice: rewards

Note. * p <.05. Predictor variables are centered in the multiple linear regression analysis to avoid multicollinearity for interaction. Non-centered variables were also tested, which resulted to the same significant and non-significant results.

(25)

25

Results in table 7 suggest that objective performance measure did not significantly affect the perceived justice on rewards. The regression analysis showed that trust was an significant

predictor of the perceived justice on rewards, the model in step 3 becomes significant (R² = 0.337, F(8,47) = 2.989, p < 0.01) and trust is the only significant predictive variable to

effect the level of perceived justice (β = 0.361, t = 2.875, p < 0.01). It was found that the interaction of objective performance measure and trust did not significantly affect the perceived justice on rewards. The results from this regression analysis confirm that it can be concluded that subjective performance measure does not effect the perceived justice on rewards, and that trust does not positively effect the relationship between subjective performance measure and distributive justice.

Table 7. Multiple linear regression for distributive justice: rewards

Note. * p <.05, ** p <.01, *** p <.001

4.3 Regression with Perceived Justice: Promotions

Multiple linear regression was also used to examine if a relationship between subjective performance measure and perceived justice on promotion distribution exists, and to examine

Step 2 Step 3 Step 4

Variables B SE B β B SE B β B SE B β

Subjective pers. .02 .15 .02 -.02 .15 -.01 .06 .18 -.06

Trust .34 .13 .33** .35 .14 .34**

Trust x Sub. pers -.11 .21 -.08

ΔR² .00 .10* .00*

ΔF .02 6.68 .25

Subjective work .14 .15 .13 -.04 .16 -.03 -.08 .17 -.08

Trust .34 .13 .33* .34 .15 .33*

Trust x Sub. work -.10 .14 -.11

(26)

26

the role of trust in this relationship. The six control items were entered in step 1, subjective performance measure constructs were entered in step 2, trust was entered in the third step, and the interaction of trust and subjective performance measure was entered in step 4.

4.3.1 Subjective performance evaluation. Table 8 shows that firm size was the only

control item that significantly effect the level of perceived justice (β = -0.367, t = -2.202, p < 0.05). It was found that there exists a collective significant effect between the control items,

subjective performance measure and perceived justice (R² = 0.337, F(9,46) = 2.598, p < 0.05). The individual predictors were further examined, the results show that personal-related subjective measure (β = -0.277, t = -2.016, p = 0.05) and work-related subjective measure (β = 0.541, t = 3.626, p = 0.001) were significant predictors of perceived justice on promotions. By means of this analysis it can be concluded that subjective performance measure does significantly effect the level of perceived justice.

4.3.2 Trust. Trust was included in the third step of the regression analysis as an independent predictor of perceived justice on promotions. Results show that trust significantly affect the level of perceived justice (R² = 0.443, F(10,45) = 3.579, p = 0.001). Further analysis indicated that work subjective measure (β = 0.338, t = 2.185, p < 0.05) and trust (β = 0.387, t = 2.927, p < 0.01) were significant predictors of perceived justice on promotions. It was also found that the interaction of trust and subjective performance measure did not significantly effect the level of perceived justice on promotions.

(27)

27

Table 8. Multiple linear regression for distributive justice: promotions

Note. * p <.05, ** p <.01, *** p <.001.

Results of the multiple linear regression analysis suggest that there exists a collective significant effect between the control items, work-related subjective measure and perceived justice on promotion allocation (R² = 0.276, F(7,48) = 2.609, p <0.05). Further examination showed that work-related subjective measure was a significant predictor of perceived justice on promotions (β = 0.396, t = 3.006, p < 0.01). These findings suggest that subjective performance measure based on the work-related aspects positively effect distributive justice. Results of the regression also indicated that trust significantly predicted the dependent variable (R² = 0.40, F(8,47) = 3.909, p <0.001). Further analysis indicated that trust was the only construct to significantly predict perceived justice (β = 0.415, t = 3.114, p < 0.01). It was also found that

(28)

28

the interaction of work subjective evaluation and trust did not significantly predict the level of perceived justice on promotions.

Table 9 shows that objective performance measure did not significantly effect perceived justice on promotion distribution. Including the trust construct in the third step, did lead to a significant collective effect of the items in the model (R² = 0.373, F(8,47) = 3.496, p < 0.01). Results in table 9 show that trust was a significant predictor of the perceived justice on promotions (β = 0.507, t = 4.15, p < 0.001). These findings suggest that trust has an effect on the level of distributive justice. Moreover, results of the regression analysis indicated that the collective effect still holds when the interaction is included in the fourth step (R² = 0.423, F(9,46) = 3.74, p = 0.001). Further analysis indicated that the interaction of objective performance measure and trust did significantly affect the perceived justice on promotion allocation (β = 0.249, t = 1.985, p = 0.05). This suggests that trust has a significant effect on the relationship between a lower objective performance measure and perceived justice on promotions.

Table 9. Sub-regression for distributive justice: promotions

Step 2 Step 3 Step 4

Variables B SE B β B SE B β B SE B β

Personal subjective -.08 .18 -.06 -.14 .15 -.11 -.07 .18 -.06

Trust .58 .14 .52*** .56 .14 .50***

Trust x Pers. Sub .14 .22 .10

ΔR² .00 .24** .01**

ΔF .20 18.38 .43

Work subjective .46 .15 .40** .23 .16 .20 .14 .17 .12

Trust .47 .15 .42** .44 .15 .39**

Trust x Work. Sub -.18 .14 -.19

(29)

29

5. Discussion

The aim of this study was to examine the effect of the use of subjective performance measures on the level of perceived distributive justice. Furthermore, the goal of the study was to identify positive effects of trust on the association between subjective performance measure and perceived distributive justice. Previous research focused on the effect of trust on perceived justice or the effect on subjective performance measure, while this research was set up to find a moderation of trust on these two variables. Findings of the factor analysis, suggest that the perception of justice differs between compensation factors: rewards and promotions. The findings resulting from the regression analysis did suggest that differences between the two perceptions exist, which resulted in significant different relationships between variables.

5.1 Subjective Performance Measures and Perceived Distributive Justice

This study found significant negative and positive effects of subjective performance measure on perceived distributive justice. Results suggest that only two subjective measure constructs influence the level of distributive justice: personal- and work related subjective measures. It was found that a higher level of personal-related subjective measures, such as the involvement and reliability of an employee within the organization, negatively affect the level of perceived justice on promotion allocations. These finding suggest that employees will perceive the promotion allocation as more unfair, when supervisors use personal-related measures to evaluate them. This relationship between personalistic criteria and perceived fairness has been examined by Pearce, Branyiczki, and Bakacsi (1994). They found evidence that employees rewarded on personalistic criteria were more likely to report that the evaluation system was unfair, the employees in these organizations were also more likely to report that their supervisor engaged in favouritism. The latter suggest that supervisors have more opportunities to bias the performance of their employees when they evaluate them with personalistic criteria.

(30)

30

would not engage in biased activities. These speculations could both explain the differences. However, previous studies did not include distinctions between personal-related subjective measure and work-related subjective measure and the potential bias that could result from both measures.

This study did not suggest the same significant results for the effect of personal-related subjective measure on perceived justice on rewards. Kaplan and Ferris (2001) examined the differences of promotions and other rewards, such as pay raises, and their effect on perceived justice. They suggest that potential inequities would be greater when promotions are distributed in comparison with pay raises, which means that employees are more likely to perceive the distribution of promotions as more unfair than other rewards. The researchers also indicated that promotions serve a more critical role in human resource than other rewards, which could explain the differences that were found in this study. The main difference is the indivisibility of promotions compared to other rewards, not everyone can be promoted at the same time and promotions do not occur frequently, while pay raises are more distributed (Kaplan & Ferris, 2001; Barret-Howard & Tyler, 1986). The differences in the findings suggest that hypothesis 1 can be partially supported. Subjective performance measure only has a relationship with distributive justice when employees are asked to perceive the promotion distribution.

(31)

31

The findings of the study suggest that trust does not affect the relationship between subjective performance measure and distributive justice, but trust does significantly influences the perceived justice on rewards and promotion distribution. These findings correspond to findings of previous studies (Mulvaney, 2017; Condrey, 1995; McCauley & Kuhnert, 1992; Ilgen, Fisher & Taylor, 1979). Mulvaney (2017) stated that employees with higher trust in their supervisor than other employees, had higher distributive justice perceptions. McCauley and Kuhnert (1992) also indicated that the fairness perceptions of employees depended heavily on the level of trust in the supervisor. This study found significant evidence that the level of trust positively affects the level of distributive justice, which is a confirmation of previous studies.

5.3 Theoretical and Managerial Implications

Even though the hypotheses could not be fully supported, this study theoretically contributed to existing organizational literature. This study took an open perspective to the relationship between subjective performance measure and distributive justice because potential biases resulting from subjective evaluation might result in both positive or negative outcomes for employees. Significant negative and positive outcomes were found in the study between different types of subjective performance measures. Therefore, this study raises awareness on the potential outcomes of different measures used in evaluation systems. The findings of the study suggest that personal-related performance measures such as involvement and reliability, have a negative effect on the perceived justice of promotion distributions. Opposite results were found for work-related performance measures. These results raise the discussion about which subjective performance measures lead to biases in the evaluation of employees. Previous research did not made a distinction between different subjective performance measures and the effect on employees’ perceptions, while the results of this study suggest that only personal-related subjective measures negatively affect the extent to which employees perceive the distribution of promotions as fair.

(32)

32

perceived justice on different rewards, such as promotions, pay raises, bonuses, profit sharing and other financial rewards that are distributed across departments.

It was also found that trust was a significant positive indicator of distributive justice. This confirms previous research and shows that even with a smaller sample, this higher level of trust influenced the level of perceived distributive justice.

This study also has managerial implications. The significant results for subjective performance measure on distributive justice indicate that supervisors should place more weight or use more measures related to the work and effort of an employee, instead of more personal characteristics such as involvement and reliability. The study also shows that managers should invest in the relationship with their employees to increase trust. This study showed that a higher level of trust ensured that employees perceived the outcomes of performance evaluation as more fair. Investing in the relationship with an employee will not only increase the perceived justice, but potentially also satisfaction (Brockner & Wiesenfeld, 1996).

5.4 Limitations

A number of limitations must be acknowledged for this study. First, the findings suggest that employees perceive the distribution of rewards other than promotions differently than promotions. It must be considered that the differences between these results might arise from the design of the questionnaire. The items that measured the extent to which employees were fairly rewarded were formulated differently than the items that measured the extent to which promotions were fairly distributed in the organization. Items that measured distributive justice on rewards were constructed as personal statements that began with “I”, which leads employees to assess their own rewards. The items about promotions were based on how fair the promotions within the organization were determined, this could lead employees to asses these items differently. The problem with these differences is that the latter items could suffer from social comparison bias when the respondents were not promoted before and others did. Employees could perceive the distribution of promotions as more unfair when they compare their situation with the situation of colleagues. This potential bias could be solved when employees were first asked if they ever got promoted, which could be used as a control variable in the regression.

(33)

33

about trust, distributive justice or satisfaction were too sensitive. The latter was included in the questionnaire by the other students. It is possible that the items of the other students were perceived as too sensitive, but it cannot be confirmed that the sample would be larger when the questionnaire only included the items used in this study. Despite the small sample, the Cronbach alpha’s were sufficient for the items to show significant results.

(34)

34

6. Conclusion

The main goal of the research was to contribute to the literature about effect of subjective evaluation on perceived justice and the importance of trust in this relationship. An empirical research was conducted where the following research questions were attempted to be answered.

Does a relationship between subjective performance evaluation and perceived distributive justice exist?

Does trust in the supervisor positively moderates the relationship between subjective performance measures and perceived distributive justice?

Previous theories suggested that subjective evaluation measures could result in biased evaluation, because supervisors’ perceptions are of high importance when these measures are used. These biases could result in both positive and negative outcomes for employees, which is why it was suggested that subjective performance evaluation could have a positive or negative effect on the perceived fairness of employees about their rewards. Trust was involved to find out how trust could influence the perceived justice when subjective measures and potential biases were used. Regression analysis was used to test the data acquired from 56 individuals.

The results showed that only perceived justice on promotion decisions would significantly have a relationship with subjective evaluations (H1). Personal-related subjective

measures negatively affected the level of perceived justice on promotions, while work-related subjective measures showed a positive relationship with the justice perceptions. However, these significant relationships were not found when perceived justice on rewards were measured. This is why hypothesis 1 was partially supported.

Furthermore, a positive significant moderating effect of trust was found between the level of subjective performance measures used and distributive justice on promotional decisions (H2). This effect was only found when the level of objective measures were low, which might

(35)

35 6.1 Future research

This study offers a contribution to previous studies concerning subjective performance measure and distributive justice. Although the evidence from this research suggest that subjective performance evaluation has no influence on the perceived justice on rewards, it might be different for other subjective evaluations that can be used in performance evaluation. For instance other qualitative indicators that are more focused on competences or learning goals, which has been used more in recent evaluation systems. The study also suggest that differences between personal-related and work-related subjective performance measures exists. Further research could make a distinction between the two constructs to find out how the different type of measures influence perceived justice or other perceptions of employees. It would be advised to look at the different aspects of subjective measures in future research, especially because this has not been done before in previous studies.

In addition, it would be valuable to perform a longitudinal research in one organization. This would give more insight into the relationships in the organization, since this is a crucial factor in perceived distributive justice. It would also improve the insights in the performance measures used by supervisors, considering that these measures could be different for each supervisor and therefore might have different effects on perceived distributive justice.

(36)

36

References

Adams, J. S. (1965). Inequity in social exchange. Advances in Experimental Social Psychology, 267–299.

Baker, G., Gibbons, R., & Murphy, K. (1994). Subjective performance measures in optimal incentive contracts. The Quarterly Journal of Economics, 109(4), 1125–1156.

Barret-Howard, E., & Taylor, T. (1986). Procedural justice as criterion in allocation decisions. Journal of Personality and Social Psychology, 50(2), 296-304.

Bol, J. C. (2011). The determinants and performance effects of managers’ performance evaluation biases. The Accounting Review, 86(5), 1549–1575.

Bol, J. C., & Leiby, J. (2018). Subjectivity in professionals’ incentive systems: Differences between promotion- and performance-based assessments. Contemporary Accounting Research, 35(1), 31–57.

Bol, J. C., & Smith, S. D. (2011). Spillover effects in subjective performance evaluation: Bias and the asymmetric influence of controllability. The Accounting Review, 86(4), 1213–1230.

Bond, S. D., Carlson, K. A., Meloy, M. G., Russo, J. E., & Tanner, R. J. (2007). Information distortion in the evaluation of a single option. Organizational Behavior and Human Decision Processes, 102(2), 240–254.

Bretz, R. D., & Milkovich, G. T. (1989). Performance appraisal in large organizations: Practice and research implications. (Working Paper no. 89-17). Ithaca: Center for Advanced Human Resource Studies, School of Industrial and Labor Relations, Cornell University.

Brockner, J., & Wiesenfeld, B. M. (1996). An integrative framework for explaining reactions to decisions: interactive effects of outcomes and procedures. Psychological bulletin, 120(2), 189.

Campbell, D. (2008). Nonfinancial performance measures and promotion-based incentives. Journal of Accounting Research, 46(2), 297–332.

Cho, Y. J., & Lee, J. W. (2011). Performance management and trust in supervisors. Review of Public Personnel Administration, 32(3), 236–259.

Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in organizations: A meta- analysis. Organizational behavior human decisions processes, 86(2), 278-321.

Condrey, S. E. (1995). Reforming human resource management systems: Exploring the importance of organizational trust. The American Review of Public Administration, 25(4), 341–354.

Cropanzano, R., Bowen, D. E., & Gilliland, S. W. (2007). The management of organizational justice. Academy of Management Perspectives, 21(4), 34–48.

Crosby, F. (1976). A model of egoistical relative deprivation. Psychological Review, 83(2), 85- 113.

(37)

37

Folger, R. (1987). Distributive and procedural justice in the workplace. Social Justice Research, 1(2), 143-159.

Folger, R., & Konovsky, M. A. (1989). Effects of procedural and distributive justice on reactions to pay raise decisions. Academy of Management Journal, 32(1), 115–130. Gibbs, M. (2008). Discussion of nonfinancial performance measures and promotion-based incentives. Journal of Accounting Research, 46(2), 333–340.

Gibbs, M., Merchant, K. A., Stede, W. A. V. der, & Vargus, M. E. (2004). Determinants and effects of subjectivity in incentives. The Accounting Review, 79(2), 409–436.

Gillenkirch, R. M., & Kreienbaum, H. (2016). What guides subjective performance evaluation: Incentive alignment or norm enforcement? Review of Managerial Science, 11(4), 933–957. Grund, C., & Przemeck, J. (2012). Subjective performance appraisal and inequality aversion. Applied Economics, 44(17), 2149–2155.

Harris, C. (1988). A comparison of employee attitudes toward two performance appraisal

systems. Public Personnel Management, 17(4), 443–456.

Holmström, B. (1979). Moral hazard and observability. The Bell Journal of Economics, 10(1), 74-91.

Homans, G. C. (1974). Social behavior: It’s elementary forms. New York: Harcourt, Brace Jovanovich.

Ilgen, D. R., Fisher, C. D., & Taylor, M. S. (1979). Consequences of individual feedback on behavior in organizations. Journal of applied psychology, 64(4), 349.

Iqbal, M. Z., Akbar, S., & Budhwar, P. (2015). Effectiveness of performance appraisal: An integrated framework. International Journal of Management Reviews, 17(4), 510-533. Kaplan, D. M., Ferris, G. R. (2001). Fairness perceptions of employee promotion systems: A two-study investigation of antecedents and mediators. Journal of Applied Social Psychology, 31(6), 1204-1222.

Levy, P. E., & Williams, J. R. (2004). The social context of performance appraisal: A review and framework for the future. Journal of Management, 30(6), 881-905. Lunenberg, F. C. (2012). Performance appraisal: Methods and rating errors. International Journal of Scholarly Academic Intellectual Diversity, 14(1), 1-9.

MacLeod, W. B., & Tan, T. Y. (2016). Optimal contracting with subjective evaluation: The effects of timing, malfeasance and guile (NBER Working Paper no. 22156).

Major, B., & Deaux, K. (1982). Individual differences in justice behavior. Equity and justice in social behavior, 43-76.

Mathis, R. L., & Jackson, J. H. (2010). Human resource management [Pdf] (13th ed.). Mason: Cengage learning.

Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20(3), 709–734.

McCauley, D. P., & Kuhnert, K. W. (1992). A theoretical review and empirical investigation of employee trust in management. Public Administration Quarterly, 265-284

(38)

38

Mulvaney, M. A. (2017). Examining the role of employee participation, supervisor trust, and appraisal reactions for a pay-for-performance appraisal system. Public Organization Review, 19(2), 201–225.

Pearce, J., Branyiczki, I., & Bakacsi, G. (1994). Person-based reward systems: A theory of organizational reward practices in reform-communist organizations. Journal of Organizational Behavior, 15(3), 261-261.

Prendergast, C., & Topel, R. (1993). Discretion and bias in performance evaluation. European Economic Review, 37(2-3), 355-365.

Price, J. L., & Mueller, C. W. (1986) Handbook of organizational measurement. Marshfield, MA: Pitmann.

Sheppard, B. H., Lewicki, R. J., & Minton, J. W. (1992). Organizational justice: The search for fairness in the workplace. New York: Lexington Books.

Sitkin, S. B., & Roth, N. L. (1993). Explaining the limited effectiveness of legalistic “remedies” for trust/distrust. Organization Science, 4(3), 367–392.

Spector, P. E., & Brannick, M. T. (2011). Methodological urban legends: The misuse of statistics control variables. Organizational Research Methods, 14(2), 287-305.

Ramamoorthy, N., & Flood, P. (2004). Gender and employee attitudes: The role of organizational justice perceptions. British Journal of Management, 15(3), 247-258. Read, W. H. (1962). Upward communication in industrial hierarchies. Human Relations, 15(1), 3–15.

Taber, K. S. (2017). The use of Cronbach’s alpha when developing and reporting research instruments in science education. Research in Science Education, 48(6), 1273-1296. Taylor, M. S., Tracy, K. B., Renard, M. K., Harrison, J. K., & Carroll, S. J. (1995). Due process in performance appraisal: A quasi-experiment in procedural justice. Administrative Science Quarterly, 40(3), 495.

Van den Bos, K., Vermunt, R., & Wilke, H. A. M. (1997). Procedural and distributive justice: What is fair depends more on what comes first than on what comes next. Journal

Personality and Social Psychology, 72(1), 95–104.

Subjective Performance Evaluation and Trust: Perceptions of Distributive Justice