Investigating the Determinants of Perceived Procedural Fairness in ...

JAMAR Vol. 11 · No. 1 2013Investigating theDeterminants of PerceivedProcedural Fairness inPerformance EvaluationMahfud Sholihin*Richard Pike**AbstractPrevious management accounting andcontrol system studies have examined theimpact of perceived procedural fairness onmanagers’ attitudes and behaviour.However, relatively little research hasbeen conducted on the factors whichaffect perceived procedural fairness; andtheir results are inconclusive. Theobjective of this paper is to investigate themain determinants of perceivedprocedural fairness.Using a sample of managers from threemajor organisations with headquarters inEurope, this study finds that perceivedprocedural fairness is affected by goalrelatedvariables in the form ofparticipation in setting performancetargets, the goal-attainment-reward link,and the specificity of goals to be achievedby managers; but is not affected bywhether the performance measures usedto evaluate performance are primarilyfinancial or nonfinancial.KeywordsGoal-Attainment-Reward LinkGoal SpecificityPerformance Measurement and EvaluationProcedural FairnessTarget-Setting Participation* Gadjah Mada University, Indonesia**Bradford University, UKIntroductionThe performance measurement, evaluation,and reward system (PMERS) is an importantcomponent of Management Control Systems(MCS). The importance of fairness in thedesign of PMERS has been recognised in thenormative literature on MCS. Anthony andGovindarajan (1998, p. 556) argue that topmanagement should be aware that “objectives,goals and standards are likely to providestrong incentives only if managers perceivethem as fair” (emphasis added). Similarly,Kaplan and Atkinson (1998, p. 682) highlightthe importance of fairness in PMERS asfollows:“There are important behaviouralconsiderations that the performancemeasurement system must reflect. First andabove all, the individual must believe thatthe system is fair … Absent this belief, themotivational potential of incentivecompensation will be lost.” (emphasisadded).Previous empirical studies in managementaccounting and control systems (e.g.Lindquist, 1995; Libby, 1999; Libby, 2001;Wentzel, 2002; Lau and Tan, 2006; Sholihinand Pike, 2009) find that fairness is negativelyassociated with dysfunctional attitudes andbehaviour, such as job-related tension andbudgetary slack, and enhances functionalbehaviour, such as trust and organisationalcommitment, and positively associated withoutcomes such as task satisfaction, jobsatisfaction and performance.Considering the importance of perceivedfairness in the design of PMERS, it isimportant to identify and understand the mainfactors affecting such perceptions. However,the literature reveals that prior studies (e.g.Hopwood, 1972; Otley, 1978; Lau andSholihin, 2005; Lau and Moser, 2008;Hartmann, Naranjo-Gil, and Perego, 2009)have yielded inconclusive, even contradictory,evidence on the particular PMERS elementsaffecting the perception of fairness.Hopwood (1972) examined whethersupervisory evaluative style (how superiorsevaluate the performance of subordinates)affects the subordinates’ attitudes, including29

JAMAR Vol. 11 · No. 1 2013subordinates’ perception of fairness ofperformance evaluation processes. In doing so,he constructed three categories of performanceevaluation style: Budget Constrained (BC)style (budgetary information is used in a rigidmanner), Profit Conscious (PC) style(budgetary information is used in a moreflexible manner, and Non-accounting (NA)style (budgetary information is seen as beingof secondary importance). His resultsindicated that BC style is perceived as less fairthan the other styles. Otley (1978) replicatedHopwood’s (1972) study and found, contraryto Hopwood’s study, that supervisoryevaluative style is not significantly associatedwith perception of fairness of performanceevaluation processes.Lau and Moser (2008) examined the effect ofnonfinancial performance measures onsubordinates’ perception of proceduralfairness, arguing that their use may lead tohigher perceptions of fairness in performanceevaluation processes (procedural fairness).They also argue that nonfinancial measures aremore consistent with the procedural fairnessrules/criteria proposed by Leventhal (1980).Their findings show a positive associationbetween nonfinancial measure usage andprocedural fairness.Lau and Sholihin (2005) contended and foundthat both nonfinancial and financial measuresare positively associated with proceduralfairness, although the effect of financialmeasures on fairness was higher compared tothat of nonfinancial measures. This is incontrast with Hopwood’s (1972) finding thatthe effect of Non-Accounting style on fairnessis significantly higher than that of the Budget-Constrained style.A more recent study by Hartman et al. (2009)argues that the use of subjective performancemeasures is positively associated withprocedural fairness. However, their findingsdo not support this argument. Indeed,procedural fairness is associated with the useof objective performance measures.Hartmann and Slapnicar (2009) argue thatprocedural fairness is not affected byperformance measures or metrics used toevaluate subordinates’ performance but isaffected by the degree of formality in the useof the performance evaluation system. Theirempirical findings, however, show nosignificant association between formality andprocedural fairness.The foregoing summary of relevant priorstudies present a confusing, evencontradictory, picture of the antecedents ofprocedural fairness in PMERS. Given theimportance of this variable, as identifiedearlier by Anthony and Govindarajan (1998)and Kaplan and Atkinson (1998), the desire toclarify and better understand the determinantsof perceived procedural fairness providesstrong motivation for this study. In this paper,we investigate the determinants, orantecedents, of procedural fairness perceptionin performance measurement, evaluation andreward systems.This study is different from previous studies,particularly Sholihin and Pike (2009), in thefollowing ways. First, whilst Sholihin and Pike(2009) investigates the consequences ofprocedural fairness, this study investigates thedeterminants of procedural fairness. Second,the variables examined by Sholihin and Pike(2009) are procedural fairness, distributivefairness, trust, organisational commitment, jobsatisfaction, and managerial performance. Thevariables examined in this study areprocedural fairness, participation in targetsetting, transparency of the goal-attainmentrewardlink, goal specificity, financialmeasures, and nonfinancial measures. Third,whilst Sholihin and Pike (2009) employsorganisational justice theory, in this study weuse organisational justice and goal-settingtheories.Drawing on organisational justice literature(e.g. Lind and Tyler, 1988; Leventhal, 1980)and goal-setting theory (Locke and Latham,1990), this study argues that a manager’sperception of procedural fairness is influencedby goal-related variables in the form ofparticipation in setting performance targets,the goal-attainment-reward link, and thespecificity of goals to be achieved bymanagers; but is not affected by the form ofperformance measures used to evaluateperformance, whether financial (objective) ornonfinancial (subjective) measures, as arguedby Lau and Sholihin (2005), Lau and Moser(2008), and Hartmann et al., (2009).The proposition that procedural fairness isdetermined by the degree of participation intarget setting draws on the literature on30

JAMAR Vol. 11 · No. 1 2013procedural justice, which has consistentlyfound that disputants perceive the procedure asfair if they have process control (i.e. sufficientopportunity to present their case), oftenreferred to as ‘voice’ (Thibaut and Walker,1975; Folger, 1977; Lind and Tyler, 1988). Inaddition, participation is consistent withLeventhal’s (1980) procedural fairness rules.The goal-attainment-reward link is proposedas an antecedent of procedural fairnessbecause clarity and transparency are outlinedby Leventhal (1980) as important componentsof procedural fairness rules. In relation to goalspecificity, Lau and Sholihin (2005) argue thatthe adoption of specific performancemeasures, are more likely to give rise tospecific goals and targets for employees topursue, than where there are no specificperformance measures for performanceevaluation. The existence of specific goals andtargets, in turn, is likely to affect employeebehaviour because the specificity of goalsenhances the direction and clarity of tasks tobe performed and the roles of the subordinates.The rest of the paper is organised as follows.Section two will discuss literature review andhypotheses development. Section three,describes the research method employed inthis study. Section four will present findingsand discussions. This paper will be closed byconclusions, limitation and suggestion forfuture research in section five.Literature Review and HypothesesDevelopmentParticipation and Procedural FairnessIn this study, participation is conceptualised asparticipation in setting performance goals ortargets, which include financial andnonfinancial targets. This is a departure frommost previous accounting studies whichconceptualise participation more narrowly asbudgetary participation.Procedural fairness can be conceptualised asthe judgments on the fairness of social normswhich deal with how decisions are made andhow individuals are treated by authorities andother parties (Lind and Tyler, 1988). However,since this study specifically deals with theperformance measurement, evaluation andreward system, Folger and Konovsky’s (1989)definition of procedural fairness is adopted,namely, the perceived fairness of the meansand procedures used to determine the amountof reward or compensation employees receive,i.e. the fairness of all aspects of theorganisation’s procedures that are used by thesuperior to evaluate a subordinate’sperformance, to communicate performancefeedback and to determine the subordinate’srewards such as promotion and pay increases.Leventhal (1980) contended that theperception of procedural fairness will beaffected by six rules: consistency, biassuppression, accuracy, correctability,representativeness, and ethicality of theprocedures. The consistency rule states that tobe fair a procedure must be appliedconsistently across persons and across time.This implies that all parties have the samerights under the procedures and are treatedsimilarly, and that the procedure is enacted thesame way each time it is used. The biassuppressionrule stipulates that procedures arefair if the decision maker does not have avested interest in any specific decision, and ifthe decision maker is not influenced by priorbeliefs. The accuracy rule states that to be fairprocedures should be based on as much goodqualityinformation and informed opinion aspossible, while the correctability rule statesthat opportunities must exist to modify andreverse decisions made. Therepresentativeness rule stipulates that “allphases of the allocative process must reflectthe basic concerns, values, and outlooks ofimportant subgroups in the population ofindividuals affected by the allocative process”(Leventhal, 1980, pp.44-45). This rule isclosely related to power sharing andparticipatory decision making. Finally, theethicality rule states that to be fair, proceduresmust be compatible with fundamental moraland ethical values.Most of the above requirements for proceduralfairness can be promoted through participationin target setting. For example, participation isconsistent with the characteristics ofrepresentativeness, correctability, accuracy,bias-suppression and ethicality in that itpermits subordinates to reflect their concernsand values, gives opportunity to modifydecisions, can be used as a way of sharinginformation, provides opportunity forsubordinate managers to correct anyinappropriate prior beliefs held by theirsuperiors, and is consistent with the moralvalue that individuals should have the31

JAMAR Vol. 11 · No. 1 2013opportunity to engage in setting goals. It istherefore reasonable to propose thatparticipation is positively associated withprocedural fairness. Additionally, Lind andTyler (1988, p. 236) argue that “theopportunity to exercise voice (participation)constitutes a visible marker of groupmembership … mute procedures are seen as… unjust because they appear to deny fullmembership rights to those denied voice”(parentheses added) and “…one of the mostpotent determinants of the procedural fairnessof a social decision-making procedure is theextent to which those affected by the decisionare allowed to participate in the decisionmakingprocess through the exercise ofprocess control or voice” (p. 176). Similarly,Early and Kanfer (1985) argue that sinceparticipation gives opportunity for input, itprovides the individual with perceived masteryor control over a situation; thus participationmay enhance perceived fairness.In a budgeting context, Wentzel (2002) studied88 cost centre managers of a downsizedhospital in the US and found that budgetparticipation is positively associated withprocedural fairness. Similar results were foundby Lau and Tan (2006) among managers ofmanufacturing companies in Singapore. Thisstudy therefore argues that a positiveassociation holds between participation insetting targets and procedural fairness.Ha1: Participation in target setting ispositively associated with procedural fairness.Goal-Attainment-Reward Link andProcedural FairnessThe management control systems literaturesuggests that the performance evaluation andreward system is an important component ofthe management control system designed tomotivate organisational members to performbetter and in accordance with theorganisation’s objectives (Otley, 1999;Merchant and Van der Stede, 2003).According to Emmanuel et al. (1990) andMerchant and Van der Stede (2003), one of theprincipal means of motivating subordinatestowards effective performance is to linkorganisational rewards to the level ofperformance achieved. This implies that thelink between the achievement of goals orperformance targets set and rewards should betransparent. A recent empirical study byKominis and Emmanuel (2007), using asample of middle managers in a large UKbasedfinancial institution, found that theeffect of transparency in the performancerewardlink on motivation is indirect via thevalue of extrinsic rewards. In other words,Kominis and Emmanuel’s (2007) studysuggests that the effect of transparency of theperformance-reward link on motivation maybe mediated by other variables, and in theirstudy this was the value of extrinsic rewards.Drawing on Emmanuel et al.’s (1990) andMerchant and Van der Stede’s (2003)contentions and extending Kominis andEmmanuel’s (2007) findings, this study arguesthat the transparency of the goal-attainmentrewardlink is likely to enhance motivation viaperceived procedural fairness. This is becausea transparent rewards system should beconsistent, unbiased, and accurate, all of whichare characteristics of procedural fairness(Leventhal, 1980). Therefore, this studyhypothesises that transparency of the goalattainment-rewardlink will be positivelyassociated with procedural fairness.Ha2: Transparency of the goal-attainmentrewardlink will be positively associated withprocedural fairness.Goal Specificity and Procedural FairnessGoal specificity refers to “the extent to whichthe goals are clearly defined by a supervisor”(Fang et al., 2005). In the context ofperformance measurement, evaluation andreward systems the goals can be financial ornonfinancial. Organisations which develop andadopt specific performance measures, whetherfinancial or nonfinancial, for performanceevaluation are more likely to develop specificgoals and targets for employees to pursue,than are organisations which have no specificperformance measures for performanceevaluation. Further, the existence of prespecifiedgoals is likely to provide clearerunderstanding (goal clarity) for organisationalmembers and indicate how they will beevaluated. In addition, goal specificity andclarity informs employees of theirresponsibilities and performance targets. Inother words, the existence of specific goalswill guide employees in deciding where theyshould direct their attention and effort.32

JAMAR Vol. 11 · No. 1 2013In relation to procedural justice, the existenceof goal specificity/clarity is in line with theconsistency and bias-suppression rules, as aspecific goal is likely to facilitate theapplication of procedures consistently acrosstime and persons and be less biased and lesssubjective. In addition, the existence ofspecific goals indicates that there is a standardto follow which provides opportunity formanagers to challenge/rebut evaluation. Anempirical study by Greenberg (1986) withsamples of middle managers in the US foundthat the ability to rebut/challenge evaluationand consistent application of standards areimportant determinants of perceivedprocedural fairness. Hence this studyhypothesises that goal specificity is positivelyassociated with procedural fairness.Ha3: Goal specificity is positively associatedwith procedural fairness.Research MethodData and SampleData used in this study were gathered bymeans of questionnaire survey in three majororganisations, followed by interviews withselected respondents. The organisations wereapproached to participate in the study becausethey were leading companies in their sectorthat had been operating PMERS for some timewith varying degrees of satisfaction.CO1 is part of a highly centralised globalorganisation operating in a specialist, researchdrivensector whose headquarter is inSwitzerland. Much of its recent performancegrowth is through reducing complexity, usingfewer suppliers and closing smaller plants.CO2 and CO3 are financial servicesorganisations, providing specialist mortgageand savings products through a large numberof branches. Their corporate offices are both inthe UK.For the purpose of filling in the questionnaire,purposive sampling was employed. Theprinciple in this type of sampling is to get allpossible cases that fit particular criteria(Neuman, 2003). The criteria to be included asthe samples are: 1) managers have beenworking for the organisations for more thanone year; 2) they have been evaluated by theirsuperiors; and 3) they have received theperformance evaluation feedback. Thesecriteria are to ensure that the samples are valid.The survey was administered as follows.Working closely with managers from the threeorganisations, a preliminary notification wascirculated encouraging managers to participatein the survey. After obtaining seniormanagement permission to conduct theindependent research study the surveyinstrument was distributed to potentialrespondents, together with assurance ofconfidentiality. Reminders were sent one,three and seven weeks after the originalmailing. The survey package and reminderletters were sent via e-mail. Respondents couldreturn the completed questionnaireselectronically by e-mail or send a hard copyversion by post.The distribution of the survey instruments is asfollows: 102 to CO1, 99 to the CO2, and 95 toCO3. Of 296 questionnaires distributed, 174were returned (55 from CO1, 52 from CO2,and 67 from CO3) yielding a total responserate of 59 per cent. Careful examinationrevealed that 9 responses (1 from CO1, 2 fromCO2, and 6 from CO3) were not usable,yielding a total of 165 usable responses (56%).Respondents were invited to take part infollow-up interviews or group discussions.These interviews sought to better understandthe results of the quantitative results and gaininsight into the context in which the PMERSoperated.Variables and Their MeasurementThe dependent variable in this study isperceived procedural fairness and theindependent variables are participation intarget setting, the transparency of the goalattainment-rewardlink, and goal specificity. Inaddition to these variables, the study includedthe type of performance measures in the modelas the controlling variable. To increase thevalidity of the measures and to be morecomparable with previous studies, allmeasures were taken from previous literature. 11 All of the instruments were pilot-tested prior tothe distribution to the respondents, on executiveMBA students of an UK university.33

JAMAR Vol. 11 · No. 1 2013Procedural FairnessThis variable is measured using a four-iteminstrument developed by McFarlin andSweeney (1992). In management accountingstudies it has been used by, for example, Lauand Sholihin (2005) and Lau and Tan (2006).Respondents were requested to rate thefairness of the procedures used to evaluatetheir performance, to communicateperformance feedback, and to determine theirpay increases and promotion, ranging from 1(very unfair) to 7 (very fair).Participation in Target SettingTo measure this variable, respondents wereasked to indicate their level of agreementusing a 7-point Likert-type scale, ranging fromstrongly disagree to strongly agree, on a singleitem of “my superior allows me to participatein setting my performance goals/targets”. Thismeasure is adapted from the goal-settingquestionnaire developed by Locke and Latham(1984).Goal-Attainment-Reward LinkTo measure this variable, respondents wereasked to indicate their level of agreementusing a 7-point Likert-type scale, ranging fromstrongly disagree to strongly agree, on a singleitem of “my rewards are tied to theachievement of the performance goals”. Thisinstrument is adapted from Kominis andEmmanuel (2007).Goal SpecificityTo measure goal specificity, three items weretaken from Fang et al. (2005): (1) my superiorspecifically explained what my performancegoals are; (2) I have very specific performancegoals in my job; and (3) I understand the exactlevel of my assigned performance goals.Respondents were requested to indicate theirlevel of agreement to the above items, using aseven-point Likert-type scale ranging from 1(strongly disagree) to 7 (strongly agree).Performance MeasureThe performance measure is employed as thecontrolling variable. To measure this variable,respondents were requested to indicate howmuch importance they thought theirsupervisors attach to certain performancemeasures when evaluating their performance,using a seven-point Likert scale, anchored 1(no importance) and 7 (always important) with0 if not applicable. 2 The performancecategories used are based on those of Ittner etal. (2003), which they consider to be importantdrivers of long-term organisational success.The categories are: relations with customers,relations with employees, operationalperformance, product and service quality,alliances with other organisations, relationswith suppliers, environmental performance,product and service innovations, communityperformance, and financial performance.Findings and DiscussionsAs mentioned in the previous section, a totalof 296 questionnaires were distributed to threedifferent organisations and 174 were returned.Of these responses, 9 were unusable due to asubstantive part of the questionnaire beingincomplete, resulting in the 165 finalresponses used in this study. The demographicinformation of the respondents, presented intable 1, consists of the average number ofemployees that respondents directly managed,their length of service (tenure) in theorganisation, tenure in their current position,and the length of time they have beensupervised by their current supervisor.The table indicates that respondents from CO1rated all demographic variables more highlythan the other two organisations, with CO2having far lower ratings for length of serviceand supervision under current supervisor.However, of particular importance to thisstudy is that all respondents had experiencedthe performance review process on at least oneoccasion.To examine whether response bias exists, nonresponsebias test was performed. We splitresponses into “early” and “late” responses.Following Hall (2008), “Early” is defined asthe first 20 percent of responses and “late” isdefined as the last 20 percent of responses.The responses from those two groups werecompared by running t-tests. Table 2 presentsthe results of non-response bias tests for eachvariable. The results indicate that the majorityof the scores of the variables do not differbetween the early and late responses.2 In the analysis, following Ittner at al. (2003), 0was converted into 1.34

JAMAR Vol. 11 · No. 1 2013However, there are some differences,particularly for CO2. For example, earlyrespondents indicated that they are evaluatedmore through nonfinancial measures, andperceived that the goal-attainment-reward linkis more transparent and the reward distributionis fairer. It is difficult to offer a good reasonfor the particular differences but it is sensibleto interpret results of these variables withcaution. However, the table shows there is nosignificant difference for the dependentvariable studied, i.e. procedural fairness, inany organisation.Table 1: Demographic Information of the Respondents Means (with Range Shown in Brackets)CO1 CO2 CO3Employees reporting to respondent 8(1-60)7(1-60)5(1-21)Tenure in organisation 3 16.5(1-36)5.6(1-19)15.4(2-37)Tenure in current position 7.8(1-36)2.7(1-8)5.2(1-25)Supervision under current superior 3.7(1-18)2.6(1-8)2.5(1-6)Table 2: Results of Non-Response Bias TestsCO1 CO2 CO3VariableMean p- value Mean p- value Mean p- valuedifferencedifferencedifferenceNonfinancial measures 0.574 0.361 0.827 0.004 0.094 0.784Financial measures 0.200 0.869 1.015 0.156 0.628 0.443Participation 0.500 0.459 0.492 0.161 0.263 0.562Goal-attainment-reward link -0.100 0.906 1.121 0.002 0.462 0.317Goal specificity -0.367 0.597 0.005 0.986 0.336 0.297Procedural fairness 0.525 0.355 0.466 0.211 0.444 0.340Table 3: Descriptive Statistics of Variables StudiedOrganisation/Variable Minimum Maximum Mean StandardDeviationCO1Nonfinancial measures 1.78 7.00 4.01 1.33Financial measures 1.00 7.00 4.88 2.30Participation in target setting 1.00 7.00 4.60 1.64Transparency of goal-attainmentrewardlink1.00 7.00 3.82 1.75Goal specificity 1.33 7.00 4.51 1.57Procedural fairness 1.75 6.00 3.89 1.23CO2Nonfinancial measures 1.67 6.56 3.83 1.03Financial measures 1.00 7.00 4.74 1.99Participation in target setting 3.00 7.00 5.43 1.09Transparency of goal-attainmentrewardlink3.00 7.00 5.87 0.93Goal specificity 2.00 7.00 5.51 1.01Procedural fairness 3.00 6.50 5.16 0.96CO33 Tenure (both in the organisation and job) and the period with their current superior are given in years.35

JAMAR Vol. 11 · No. 1 2013Nonfinancial measures 1.00 7.00 4.95 1.90Financial measures 1.00 7.00 4.97 1.47Participation in target setting 1.00 7.00 5.02 1.31Transparency of goal-attainmentrewardlink1.00 7.00 4.95 1.90Goal specificity 3.33 7.00 5.57 0.83Procedural fairness 1.75 6.25 4.52 1.07The descriptive statistics of variables studiedare presented in table 3. This indicates that forthe performance measures variable, the highestscores both for nonfinancial and financialmeasures are found in CO3 (for nonfinancialmeasures mean=4.95; SD=1.90 and forfinancial measures mean=4.97; SD=1.47). Theresults suggest that in this organisationrespondents perceived that their superiors relyon both financial and nonfinancial measures.The lowest score for nonfinancial measuresand financial measures are found in the CO2(for nonfinancial measures mean=3.83,SD=1.03 and for financial measuresmean=4.74, SD=1.99). For the dependentvariable, i.e. procedural fairness, the lowestscore is found in CO1 (mean= 3.89; SD= 1.23)and the highest score is found in the CO2.Interviews with selected respondents fromCO1 established that they viewed theprocedure as somewhat unfair because theperformance evaluation system was vague,inadequate and offering little reward foremployees attaining performance measures. Inaddition there was little opportunity forpromotion, low or nonexistent pay rises, andlittle incentive through financial ornonfinancial rewards. This is also evidencedby the lower scores for the transparency of thegoal-attainment-reward link. The followingcomments gathered from CO1 managersreinforce this:“… Yearly appraisal is merely formality”(marketing managers).“Performance evaluation is inadequateand vague...at present the appraisal systemis informal and not enforced…innovationand creativity are not being rewarded andencouraged…no rewards for meetingperformance target …nothing happenwhether they (targets) are achieved or not”(R & D manager).“I’m asked to give more effort than otherswho appear to get more reward financiallyand credit for what they do… (but I’m) notbeing given the credit for (my)achievement” (production manager).The highest score of procedural fairness isfound in CO2. The following quote gainedfrom the group discussion reinforced thesurvey findings:“I thought the process was fair though I amnot happy with my pay rise but expect thatthis has been affected by the credit crunchwhich has impacted on everyone’s pay andbonus”.To test the hypotheses, data were analysedusing both multiple regression and PLS. 4 Priorto the main analysis, ANOVA tests betweengroups were performed to see whether thereare any differences among sub-samples. It wasfound that variability of variance between thegroups (i.e. organisations) for proceduralfairness, participation, transparency of goalattainment-rewardlink, and goal specificity issignificant (see table 4). Therefore in runningregression we controlled the organisationeffect.In addition to the ANOVA test, theinherentassumptions relating to the adequacyof regression models - normality,homoscedasticity, multicolinearity andlinearity - were tested. Normality refers to theshape of data distribution for a metric variableand its correspondence to the normaldistribution. Hair et al. (1998) argue that thisassumption is the most fundamentalassumption in regression analysis. To assessthe normality, Cohen and Cohen (1983)recommend an examination of residual scatter4The regression was executed using an OLSapproach with SPSS software and then verifiedusing PLS with PLS Graph 3 software.36

JAMAR Vol. 11 · No. 1 2013plots of the dependent and independentvariables. The results of the test indicate thatthe normal probability plots of the residuals ofthe regression models used in this study arescattered along a relatively straight line. Thismeans that the residuals are normallydistributed and thus, the normality assumptionis met.With homoscedasticity, it is assumed thatdependent variables exhibit constant varianceacross the range of independent variables(Hair et al., 1998). This assumption can betested by plotting the residuals against thepredicted values (Cohen and Cohen, 1983;Hair et al., 1998). The homoscedasticityassumption is not violated when the plots ofthe residuals against the corresponding fitted(predicted) values for the models show that allplots are scattered randomly betweenapproximately equal horizontal bands with nodiscernible patterns or systematic variations.The results indicate that the homoscedasticityassumption is not violated.Another assumption tested is multicollinearity.Multicollinearity occurs when the correlationsamong variables are so high that certainstatistical analysis cannot be performed.Multicollinearity signifies that at least twoseparate variables are measuring the samething. Some indications of multicollinearityare: the correlation between independentvariables being greater than 0.85 (Kline,1998); the tolerance levels for the variablesand any value being very small, less than 0.1;and the value of variance inflation factors(VIF) being greater than 10 (Kline, 1998;Pallant, 2005). The regression model showsthat the tolerance levels for the variables areall above 0.1, and that the value of varianceinflation factors (VIF) is less than 10. Hence,there is no multicollinearity problem. Inaddition, none of the correlations betweenindependent variables is greater than 0.85.Linearity refers to the extent to which anychanges in dependent variables are associatedwith the changes in the independent variables.Following Cohen and Cohen (1983), scatterplots of residuals were examined. The resultindicates that the appropriateness of the linearmodels is not violated.Table 5 presents the results of bivariatecorrelation analysis and table 6 presents theresults of regression analysis. Table 5 showsthat participation, reward link, and goalspecificity are positively associated withprocedural fairness. On the other hand, neitherfinancial nor nonfinancial measures areassociated with procedural fairness. Hence, theresults provide initial support for thehypotheses that procedural fairness is affectedby participation, reward link, and goalspecificity, but do not provide initial supportfor the hypotheses that procedural fairness isaffected by the type of performance measureused to evaluate the respondents’ performance.In addition, the table also indicates that allgoal-related attributes (goal settingparticipation, goal-attainment-reward link, andgoal specificity) are correlated significantlywith each other. The table also reveals thatfinancial performance measures are perceivedto be more specific compared to that ofnonfinancial measures.Table 6 presents the results of regressionanalysis with procedural fairness as thedependent variable and goal-related attributes(goal-setting participation, goal-attainmentrewardlink, and goal specificity) as theindependent variables, controlling forperformance measures (financial andnonfinancial) and the company. The tableindicates that participation, reward link andgoal specificity significantly affect proceduralfairness, although the effect of goal specificityis marginally significant (p

JAMAR Vol. 11 · No. 1 2013Table 4: The Results of ANOVA TestsVariable F SigProcedural fairness 17.580 0.000Participation 4.448 0.013Transparency 29.802 0.000Goal specificity 14.042 0.000Financial measures 0.152 0.859Nonfinancial measures 0.579 0.449Table 5: The Results Of Correlation AnalysisPF Part Link GS FM NMProcedural Fairness 1(PF)0.447** 0.570** 0.467** 0.059 0.005Participation1(Part)0.433** 0.484** 0.070 0.148Reward link1(Link)0.515** 0.225** 0.128Goal specificity1(GS)0.191* 0.083Financial measures1(FM)0.223*Nonfinancial1measures (NM)** Correlation is significant at the 0.01 level (2-tailed).* Correlation is significant at the 0.05 level (2-tailed).Table 6: The Results Of Regression Analysis Using OLS Approach.Variables Coefficient Coefficient value p-valueConstant b 0 1.759 .001Participation b 1 .199 .036Transparency link b 2 .415 .000Goal specificity b 3 .169 .097Financial measures b 4 -.064 .444Nonfinancial measures b 5 -.077 .349Company b 6 .015 .862Adjusted R 2 = 36.8%F = 10.993; p= 0.000Table 7: The Results of Regression Using PLS Approach (T-Statistics in Brackets)VariablePath toProcedural fairnessParticipation in target setting 0.191 (2.218)*Goal-attainment-reward link 0.374 (4.771)**Goal specificity 0.175 (2.077)*Financial measures -0.071 (1.214)Nonfinancial measures 0.112 (0.821)Company 0.026 (0.314)R 2 41.4%** Significant at the 0.01 level (2-tailed).* Significant at the 0.05 level (2-tailed).38

JAMAR Vol. 11 · No. 1 2013Overall, the results of statistical analysissupport the arguments that procedural fairnessis affected by participation in target setting,transparency of the goal-attainment-rewardlink, and goal specificity. As argued earlier,these three variables are consistent with thejustice rules proposed by Leventhal (1980) ascontended in this study. However, the resultsdo not support the argument that the use offinancial and nonfinancial measures isassociated with procedural fairness (c.f. Lauand Sholihin, 2005; Lau and Moser, 2008).The interview results support the abovefindings. For example, a production managerfrom CO1 emphasised the importance of goalspecificity as follows:“… well defined targets help me to getorganised which improves myperformance…if targets are not welldefined (I am) not able to achieve them”.Another manager, a marketing manager, fromthe same company said:“…(specific goals) allow me (to be) morefocused on achieving targets, (to make)more focused effort”.The above two managers argued that specificgoals/targets are important because they helpthem to achieve better performance. Hence, itis likely that goal specificity is positivelyassociated with procedural fairness because itis instrumental in achieving betterperformance (c.f. Locke and Latham, 1990;2002). To some extent, it supports theinstrumental approach to procedural fairness,i.e. an individual prefers fair proceduresbecause it will give them better outcomes(Lind and Tyler, 1998). A similar reason israised by a marketing manager from CO2:“Specific goals are important to mebecause specific goals motivate me tohigher performance and provide me withrequired focus”.With regard to participation, the abovemarketing manager offers reasons why it isimportant:“Participation allows me to give input and…I feel involved in business and …I canexpress my opinions and I can influenceoutcomes”.Likewise, the production managers argued:“Participation gives me a betterunderstanding of how to achieve targetsand understand why I have to do it … itimproves my motivation …”.However, a health and safety manager saidthat participation was “…just for personalsatisfaction”. Presumably, part of thesatisfaction is an assurance that the evaluationprocedures are conducted in a fair manner.The above quotes suggest various reasons whymanagers value participation. This isconsistent with previous studies inorganisational justice on the positive effects ofparticipative decision making on proceduralfairness (see Greenberg and Folger 1983; Biesand Shapiro 1988; Magner et al. 1992). Ehlenand Welker (1996) identified whyparticipation affects the perception of fairness.They are: (1) participation serves to enhance aperception of the sincerity of the positiontaken by a decision maker; (2) participationmay generate a perception of enhanceddecision quality; (3) participation sends amessage to participants about how they areperceived by others; and (4) people desirecontrol over decision outcomes which affectthem, and participation is seen as a way toachieve this control.Conclusion, Limitations andSuggestion for Future ResearchThe objective of this study is to investigate theantecedents of procedural fairness in thecontext of PMERS. This study is important asprevious studies of fairness within PMERScontext produce inconclusive findings onfactors affecting fairness perception. Drawingon organisational justice literature and goaltheory the study argues that perceivedprocedural fairness is affected by goal-relatedvariables, i.e. participation in settingperformance targets, the goal-attainmentrewardlink, and the specificity of goals to beachieved by managers. Using a questionnairesurvey with samples derived from three profitorientedorganisations, one manufacturingcompany and two financial servicesorganisations, this study finds that perceivedprocedural fairness is affected by participationin setting performance targets, the goalattainment-rewardlink, and the specificity of39

JAMAR Vol. 11 · No. 1 2013goals to be achieved by managers.Additionally, the study finds that the type ofperformance measures (whether financial ornonfinancial) used to evaluate managers is notassociated with perceived procedural fairness.It would seem that much of the prioraccounting literature has overemphasised theimportance that performance measure typesmay have on procedural fairness, whetherbudget constrained, profit conscious or nonaccounting,to the neglect of more criticalvariables.This study contributes to the literature ofprocedural fairness in the context of PMERSby providing empirical evidence on factorswhich influenced perceived proceduralfairness. Looking at the practical implicationsof the study, the results suggest that superiorsshould focus on goal-related variables, i.e.participation in setting performance targets,the goal-attainment-reward link, and thespecificity of goals to be achieved bysubordinates in designing PMERS rather thanon the types of performance measures.The study, however, should be interpretedcautiously for the following reasons. The firstone is the inherent limitations associated withsurvey method. Future study should examinethe issue using other approaches, such asexperimental design. Secondly, this studydraws on managers from just threeorganisations. Future study could examine themodel using samples derived from moreorganisations. Thirdly, some variables weremeasured using a single item. However, tocompensate this we have conductedinterviews. Future study could use the bettermeasure to study the same topic..Notwithstanding these limitations, we believethat this study provides clarity on factorsaffecting perceived procedural fairness.ReferencesAnthony, R. and Govindarajan, V. (1998),Management Control Systems, Irwin, Chicago.Bies, R. J. and Shapiro, D. L. (1998), “Voiceand justification: Their influence onprocedural fairness judgments”, Academy ofManagement Journal, September, 3: pp. 676-685.Cohen, J. and Cohen, P. (1983), AppliedMultiple Regression: Correlation Analysis forthe Behavioral Sciences, L. ErlbaumAssociates, New Jersey, USA.Earley, P. C. and Kanfer, R. (1985), “Theinfluence of component participation and rolemodels on goal acceptance, goal satisfaction,and performance”, Organizational Behaviorand Human Decision Processes, 36 (3): pp.378-390.Ehlen, C. R. and Welker, R. B. (1996),“Procedural fairness in the peer and qualityreview programs”, Auditing: A Journal ofPractice & Theory, 15 (1): pp. 38-52.Emmanuel, C. R., Otley, D. T., Merchant, K.A. (1990), Accounting for ManagementControl, Chapman and Hall, London, UK.Fang, E., Evans, K. R., Zou, S. (2005), “Themoderating effect of goal-settingcharacteristics on the sales control systems-jobperformance relationship”, Journal of BusinessResearch, 58 (9): pp. 1214-1222.Folger, R. (1977), “Distributive andprocedural justice: combined impact of"voice" and improvement on experiencedinequity”, Journal of Personality and SocialPsychology, 35 (2): pp. 108 - 119.Folger, R. and Konovsky, M. A. (1989),“Effects of procedural and distributive justiceon reactions to pay raise decisions”, Academyof Management Journal, 32 (1): pp. 115-130.Greenberg, J. (1986), “Determinants ofperceived fairness of performanceevaluations”, Journal of Applied Psychology,71 (2): pp. 340-342.Greenberg, J. and Folger, R. (1983),“Procedural justice, participation and the fairprocess effect in groups and organizations”,In: PAULUS, P. (Ed.) Basic Group Process,Springer-Verlag, New York, USA.Hair, J. R., Anderson, R. L., Tatham, R.E.W.C. B. (1998), Multivariate Data Analysis:With Readings, Pearson, New Jersey, USA.Hall, M. (2008), “The effect of comprehensiveperformance measurement systems on roleclarity, psychological empowerment andmanagerial performance”, Accounting,40

JAMAR Vol. 11 · No. 1 2013Organizations and Society, 33 (2-3): pp. 141-163.Hartmann, F. And Slapnicar, S. (2009), “Howformal performance evaluation affects trustbetween superior and subordinate managers”,Accounting, Organizations and Society, 34 (6-7): pp. 722-737.Hartmann, F., Naranjo-Gil, D., and Perego, P.(2009), “The effects of leadership styles anduse of performance measures on managerialwork-related attitudes”, European AccountingReview, 19 (2): pp. 275-310.Hopwood, A. G. (1972), “An empirical studyof role of accounting data in performanceevaluation”, Journal of Accounting Research,Supplement (10): pp. 156-182.Ittner, C. D., Larcker, D. F., Randall, T.(2003), “Performance implications of strategicperformance measurement in financialservices firms”, Accounting, Organizationsand Society, 28 (7-8): pp. 715-741.Kaplan, R. S. and Atkinson, A. A. (1998),Advanced Management Accounting, PrenticeHall, New York, USA.Kline, R. (1998), Principles and Practice ofStructural Equation Modelling, The GuilfordPress, New York.Kominis, G. and Emmanuel, C. R. (2007),“The expectancy-valence theory revisited:Developing an extended model of managerialmotivation”, Management AccountingResearch, 18 (1): pp. 49-75.Lau, C. M. and Moser, A. (2008),“Behavioural effects of nonfinancialperformance measures: The role of proceduralfairness”, Behavioral Research in Accounting,20 (2): pp. 55 - 71.Lau, C. M. and Sholihin, M. (2005),“Financial and nonfinancial performancemeasures: How do they affect jobsatisfaction?”, The British Accounting Review,37 (4): pp. 389-413.Lau, C. M. and Tan, S. L. C. (2006), “Theeffects of procedural fairness and interpersonaltrust on job tension in budgeting”,Management Accounting Research, 17 (2): pp.171-186.Leventhal, G. S. (1980), “What should be donewith equity theory? New approaches to thestudy of fairness in social relationships”, In:Gergen, K., et al. (Eds.) Social Exchange:Advances in Theory and Research, PlenumPress, New York.Libby, T. (1999), “The influence of voice andexplanation on performance in a participativebudgeting setting”, Accounting, Organizationsand Society, 24 (2): pp. 125-137.Libby, T. (2001), “Referent cognitions andbudgetary fairness: A research note”, Journalof Management Accounting Research, 13 (1):pp. 91-105.Lindquist, T. M. (1995), “Fairness asantecedent to participative budgeting:Examining the effects of distributive justice,procedural justice and referent cognitions onsatisfaction and performance”, Journal ofManagement Accounting Research, 7 (1): pp.122-147.Lind, E. A. and Tyler, T. R. (1988), The SocialPsychology of Procedural Justice, PlenumPress, New York, USA.Locke, E. A. and Latham, G. P. (1984), GoalSetting: A Motivational Techniques that WorksPrentice-Hall, New Jersey, USA.Locke, E. A. and Latham, G. P. (1990), ATheory of Goal Setting and Task Performance,Prentice-Hall, New Jersey, USA.Locke, E. A. and Latham, G. P. (2002),“Building a practically useful theory of goalsetting and task motivation: A 35-yearodyssey”, American Psychologist, 57 (9): pp.705-717.Magner, N. R., Welker, R. B., Johnson, G. G.(1992), “Testing a model of voice, control,procedural justice, and organizationalcommitment with latent variable structuralequation analysis”, The International Joumalof Conflict Management, 3 (3): pp. 223-236.McFarlin, D. B. and Sweeney, P. D. (1992),“Distributive and procedural justice aspredictors of satisfaction with personal andorganizational outcomes”, Academy ofManagement Journal, 35 (3): pp. 626-637.41

JAMAR Vol. 11 · No. 1 2013Merchant, K. A. and Van Der Stede, W. A.(2003), Management Control Systems:Performance Measurement, Evaluation andIncentives, Pearson Education, Essex, UK.Neuman, W. L. (2003), Social ResearchMethods: Qualitative and QuantitativeApproaches, Allyn and Bacon, Boston, USA.Otley, D. T. (1978), “Budget use andmanagerial performance”, Journal ofAccounting Research, 16 (1): pp. 122-149.Otley, D. (1999), “Performance management:a framework for management control systemsresearch”, Management Accounting Research,10 (4): pp. 363-382.Pallant, J. (2005), SPSS Survival Manual,Open University Press, Beckshire, UK.Sholihin, M and Pike, R. (2009), “Fairness inperformance evaluation and its behaviouralconsequences”, Accounting and BusinessResearch, 39 (4): pp. 397-413.Thibaut, J. and Walker, J. (1975), ProceduralJustice: A Psychological Analysis, John Wiley,New York, USA.Wentzel, K. (2002), “The influence of fairnessperceptions and goal commitment onmanagers’ performance in a budget setting”,Behavioral Research in Accounting, 14 (1):pp. 247-271.42