Investigating the Determinants of Perceived Procedural Fairness in ...

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