Is outcome fairness used to make procedural fairness judgments when procedural information is inaccessible?

Social Justice Research, Vol 9, No. 4, 1996

Is Outcome Fairness Used to Make Procedural Fairness Judgments When Procedural Information is Inaccessible?

Joseph P. Daly ~3 and Thomas M. Tripp 2

In a study of relocation decisions at seven different sites, procedural fairness was shown to be more sensitive to outcome fairness when respondents had less time to gather information about decision procedures. We interpret this finding to show that inaccessibility of information about decision procedures moderates the influence of outcome fairness judgments on procedural fairness judgments, such that outcome recipients rely more heavily on outcome fairness as a basis for forming procedural fairness judgments when information about decision procedures is not available. A second, laboratory study is reported that confirms the information inaccessibility explanation in the first study. When procedural information is available, procedural characteristics may be the primary bases for procedural fairness judgments, but when such information is unavailable, procedural fairness will likely be more sensitive to self-interest concerns. Future research should therefore take contextual factors such as accessibility to procedural information into accoun~ given that there are likely to be differences on that dimension between organizational settings on the one hand and lega~ politica~ and dispute resolution settings on the other. Information about decision procedures, general~ accessible in legal, politica~ and dispute resolution settings, is often much less accessible in organizations.

KEY WORDS: procedural justice, distributive justice.

Research on "organizational justice" (Greanberg, 1987, 1990) focuses on people's judgments of the fairness of organizational decisions. Two types

Department of Management, Appalachian State University, Boone, North Carolina 28608. Department of Management & Systems, Washington State University. All correspondence should be addressed to Joseph Daly, Department of Management, Walker College of Business, Appalachian State University, Boone, North Carolina 28608.

327

0885-7466/96/1200-0327S09.50/0 • 1996 Plenum Publishing Corporation

328 Daly and Tripp

of fairness perceptions have received particular research attention: judgments of outcome fairness and procedural fairness. Outcome fairness (also known as distributive justice; Homans, 1961) judgments are evaluations of the results produced by an allocation decision; procedural fairness judgments (Thibaut and Walker, 1975) are evaluations of the process used to make the decision.

Thibaut and Walker (1975) advanced a model of procedural fairness which maintained that judgments of procedural fairness are based on outcome recipients' considerations of long-term self-interest. Lind and Tyler (1988) labeled that perspective on procedural fairness the "self-interest model." One of the model's predictions, according to Lind and Tyler, was that outcome fairness will influence procedural fairness judgments (p. 226). An example of the influence that outcome information can have on procedural fairness judgments can be seen in public opinion polls surrounding the O. J. Simpson trial and its attending outcome. In a CBS News poll taken before the verdict, only about 40% of African Americans indicated they believed the trial was fair while nearly 80% believed he received a fair trial when polled after the verdict of "not guilty" was known (Gottlieb, 1995).

Numerous early studies of legal and dispute resolution decisions (e.g., LaTour, 1978; Lind et aL, 1980; Walker et al., 1974) found that procedural fairness influences outcome fairness (i.e., outcomes are viewed as fairer when the procedures that generate them are considered fair). Lind and Tyler (1988) maintain that causality between the two variables is reciprocal and there is experimental evidence to show that outcomes (if not outcome fairness per se) can influence procedural fairness judgments (Conlon et at, 1989; Lind and Lissak, 1985). Additionally, outcome-based concerns have been shown to affect procedural fairness as well (Shapiro and Brett, 1993). There is some evidence to show that the influence of outcomes on procedural fairness judgments is enhanced by a number of moderator variables, specifically the degree of consideration apparently given to the decision by the decision maker (Conlon et at, 1989), the lack of recipient input into the decision (Paese, 1985), and the degree of impropriety apparent in the enactment of the procedure (Lind and Lissak, 1985).

Brockner and Wiesenfeld (1996) theorize that when people think about allocation decisions, they actively seek information about the decision in order to make sense of it. Furthermore, people more likely engage in sense-making activities when either the outcome is unfair or unfavorable, or the procedures appear unfair. Thus, for example, when outcomes are unfavorable, people look for information about the decision-making procedures, if they do not have such information already.

Information Inaccessiibility 329

For some decisions, readily available procedural information feeds this sense-making process. For example, in legal, political, or formal dispute resolution settings in the United States, due primarily to constitutional guarantees the decision-making procedures are typically visible to the par- ties involved and much information about how these processes work is made public (e.g.., the courts follow a considerable number of guidelines embodied in procedural law). However, for other decisions, little procedural informatior| exists to feed the sense-making process. In particular, managers in organizations have substantially fewer legal obligations to make available to organizational constituents information regarding their decision processes. Those U.S. laws that govern relationships between public officials and citizens in legal and political settings have their basis in the U.S. Constitution and deal with public rights and respons~ilities. The laws in the U.S. that apply to corporate governance, by contrast, are rooted in the law of private property (Berle and Means, 1932). American corporate charters, given by the states, do not acknowledge a role for employees in corporate governance, rather, the corporation's management is recognized as having the responsibility for important decisions (Stohr, 1975). Thus, as Ewing (1977, p. 3) put it,

For nearly two centuries Americans have enjoyed freedom of press, speech, and assembly, due process of law, privacy, freedom of conscience, and other important fights--in their homes, churches, political forums, and social and cultural life. But Americans have not enjoyed these l~erties in most companies, government agen- cies, and other organizations where they work. Once a U.S. citizen steps through the plant or office door at 9 a.m., he or she is nearly righfless until 5 p.m., Monday through Friday. #

As a consequence,' of this state of affairs, a great many decisions in U.S. firms--particularly strategic decisions--are typically made by upper-level managers and announced to employees, with no accompanying information as to how the decision was made. (However, in less public decisions, managers often give accounts of the procedural aspects of a decision; Bies and Shapiro, 1987, 1988; Bies, et al., 1988).

This difference in the accessibility of procedural information in the two types of settings suggests the following questions: What do people do in attempting to make sense of an allocation decision when they have no information about procedures? What do they do with the external cues they do have (e.g.,, outcome information)7 More generally, how related are procedural and distn'butive fairness? Does the strength of their relationship vary by the amount of procedural information available in the setting such that distributive fairness can be interpreted to influence more strongly pro-

#In contrast to the American system, German law explicitly requires that employees be given a say in corporate governance.

330 Daly and Tripp

cedural fairness judgments when procedural information is absent versus when it is present?

Research on the relationship between distributive fairness and procedural fairness fails to provide a consistent picture of the strength of that relationship. In legal/political field settings, Casper et al. (1988) assessed r = .29; Tyler et aL (1985) found r = .28. Among organizational field studies, intercorrelations between procedural and outcome fairness by contrast are generally high (e.g., FryxeU and Gordon, 1989, r = .57; Konovsky and Cropanzano, 1991, r = .53; Sweeney and McFarlin, 1993, r = .85). When intercorrelations are high in legal settings, these often occur where the stakes are not as great for the participants and thus procedural law imposes fewer requirements regarding the presentation and consideration of evidence (e.g., Tyler's, 1984, study of traffic and misdemeanor cases; r = .77).

These studies suggest that the relationship between procedural and distributive fairness might vary according to the context of the study. Spe- cifically, distributive fairness judgments may have a greater impact on procedural fairness judgments in those settings where less information about decision procedures is available. It is likely that workers will use outcome information to infer judgments about procedures when procedural information is lacking. Brockner and Wiesenfeld (1995) stated, "When people are in sense-making mode, external cues that address their information needs should be particularly influential" (p. 21). We suspect that when procedural information is lacking, outcome information often provides the most salient and relevant external cues. Outcome information is likely to be available to the outcome recipient under those circumstances, it is likely to be one of the few relevant external cues available in the absence of procedural information, and it is logical to expect that recipients would infer aspects of the decision process from its outcome. As Tyler and Lind (1992) have observed, "favorable outcomes Can have informational, as well as hedonic, value" (pp. 143-144), suggesting that group members may make interpretations about Tyler and Lind's proposed procedural fairness ante- eedents of standing, neutrality, and trust on the basis of outcomes as well as procedures. Tyler and Lind maintain, however, that the effect of procedural characteristics is likely to be more direct than that of outcomes.

In the two studies reported here, we test the prediction that where information about organizational decision procedures is inaccessible, procedural fairness judgments will be more strongly influenced by outcome fairness. That is, the effect of outcome fairness on procedural fairness is predicted here to be moderated by the accessibility of procedural information.

Hypothesis 1: Outcome fairness judgments will influence procedural fairness judgments.

Information Inaccessibility 331

Hypothesis 2: Outcome fairness will more strongly influence procedural fairness when information about decision procedures is relatively inaccessible than when such information is relatively accessible.

Our hypoflaeses are tested here in two studies. The first is a field study, using a sample of relocated employees. The second is a laboratory study.

STUDY 1 METHOD

Sample

Research Sexing

Employees were surveyed from seven private sector organizations that had recently moved to locations in the Chicago area. All of the respondents were involved in those relocations. The relocating organizations from which respondents were sampled were identified through contacts in the business community or via the local business media. Each company's management was contacted and asked to allow their employees to participate in a study of employee responses to relocation. The researcher promised cooperating managers a feedback report.

The fairness measures used in the study were gathered on the same instrument. Measures of contextual factors were gathered at each site from an informant, in each case the individual who oversaw the implementation of the relocation for the site. All measures were obtained within 2 to 4 weeks after their finn's relocation was complete. In most instances, employees filled out their questionnaires on their own time and mailed them to the author. At two sites (n = 41), employees filled out the questionnaires during working hours and the completed questionnaires were then collected by the author.

The facility :relocation decision was used here as the focus of respondents' fairness judgments for a variety of reasons. In relocation decisions, an organization's management determines whether a facility should be moved and, if so, where it should be located. Relocation decisions are typically made using a dosed decision process. It was important that a decision process that is frequently dosed be examined in the study, since the pur- pose of the study is to examine differences between open decision processes (the focus of most research to date) and closed ones. Facility relocations are discrete decisions that generally are not made very often, even by the managers of the largest organizations. Such decisions are therefore likely

332 Daly and Tripp

to be unprogrammed and regarded by both managers and employees as unique.

While employees in relocations are generally likely to have few formal opportunities to become apprised of the relocation decision process, there may be informal opportunities for them to acquire procedural information (e.g., through leaks or rumors). Of course, such information may be accu- rate or inaccurate. Stohl and Redding (1987) maintain that rumors are more likely to emerge in organizations over issues that (like relocation) are ambiguous yet consequential to employees.

A multiorganizational sample was sought to enhance variation in relocation procedures and outcomes. It must be emphasized, however, that variations in outcomes are likely to occur within as well as between organizations in a relocation setting. Differences in commuting distances can result, for example, as when a short-distance move brings the facility location closer to one employee's residence and further away from another's. Vari- ation in procedural fairness judgments can occur even when the same procedure is enacted, given that fairness perceptions are thought to the individually based.

Sample Characteristics

A total of 192 questionnaires were returned for a response rate of 68%. In the sample, males (57.5%) outnumbered females (42.5%). The majority of the employees in the sample were not required to move far-- the median distance of the new facility location from the employee's pre- move residence was 16 miles. Correspond~gly, few employees (24.6%) reported that they had moved their residence or planned such a move. The median age of respondents was in the range of 25 to 35 and their median education fell into the category "some college." Of the sample, 73.2% reported themselves to be salaried, 0.6% on a commission basis, 26.3% hourly. Few respondents (n = 18, 9.4%) were supervisors or managers. The bulk of the remainder (n = 129, 67.2%) were support staff, clerical, or warehouse personnel. A sizable number (n = 45, 23.4%) were factory workers. The modal family income was in the range of $20,000 to $30,000.

Measures

Fairness Measures

Measures of both outcome fairness and procedural fairness focused on employees' judgments regarding their firm's decision to relocate. The

Information Inaccessibifity 333

fairness measures used Likert formats with scale anchors ranging in five steps from 1 (stngngly disagree) to 3 (neither agree nor disagree) to 5 (strongly agree).

The outcomes of a relocation decision are often numerous and variable. It is extremely difficult to predict the pattern of outcomes an employee will experience in the context of a single facility relocation, let alone that of a multiorganizational sample, as represented here. Therefore, the outcome fairness items used here were global in nature, centering on the fairness of the decision rather than on specific outcomes, a measurement strategy that is frequently used in legal settings (e.g., Tyler, 1984). The two outcome fairness items read as follow: (i) "Management was perfectly jus- tiffed in making the relocation decision that they made."; and (ii) "Man- agement had every fight to make the relocation decision that they made." The alpha for both items was .70.

Four items were used to measure procedural fairness in the present study. They read as follow: (i) "The organization went about deciding to move in a way that was not fair to me" (reverse-scored). (ii) "The organization was fair to me in the way that it made the decision to relocate." (iii) "The way that management made the relocation decision was not fair to me" (reverse scored). (iv) "The steps that the company took to make the relocation decision were fair to me." The alpha for the procedural fairness scale was .87.

Response rates and means for the various fairness measures were compared for each of the seven sites in the study. No evidence of any relationship between fairness means and response rates was detected. Thus, restriction of range due to self-selection does not appear to have biased the analyses.

Operationaliaation of Information Inaccessibility

The length of relocation process was chosen as an indicator of information inaccessibility on the assumption that, the more abbreviated the relocation process, the less time there is for employees to gather information about the procedures used to make the relocation decision. Length of process information was obtained from the individual at each site who man- aged the relocation. Those informants were asked when management first announced its im:ention to relocate to the affected employees. Those data were compared with the date of the relocation itself to generate the length of process for the, site, which was recorded in numbers of months. Because length of relocation may indicate other constructs besides information ac-

334 Daly and Tripp

cessibility, we conducted a second, laboratory study where we manipulated information accessibility directly (see below).

The operationalization of information inaccessibility used here is a global, contextual variable and, as such, is less prone to error than the aggregated variables often used in cross-level research (Rousseau, 1985). No attempt was made to measure perceptions of information inaccessibility through employee self-report because it was deemed likely that respondents would overestimate their degree of access to procedural information in an effort to appear to be "in the know." The contextual indicator used here is more distant from the respondent's direct experience than would be the case with individual-level measures. However, it has the advantage of being free from common methods bias in estimates of relationships with fairness measures, which were obtained by self-report from respondents rather than informants.

STUDY 1 RESULTS

Means, standard deviations, and intercorrelations (along with signifi- cance levels) for the variables in the study are presented in Table I. Con- sistent with Hypothesis 1, outcome fairness is correlated with procedural fairness (r = .44, p < 0.001).

We used moderated regression to test Hypothesis 2. An Outcome Fairness • Length of Process interaction term was computed by calculating the cross-product of the outcome fairness and length of process variables. When outcome fairness and length of process were combined in a regression equation with their interaction term to predict procedural fairness, the resulting formula yielded a multiple correlation of .50. Thus, the three variables taken together accounted for 24.8% of the variance in procedural fairness (adjusted R 2 = .24, n = 180). In the first step, outcome fairness accounted for 19.1% of the variance in procedural fairness, F(1, 178) = 41.99, p < 0.001. Length of process accounted for an additional 3.5% of variance over that attributed to outcome fairness, F(2, 177) - 7.98, p <

Table L Correlation Matrix for Variables in Study 1 a

Variable ~ SD 1 2

1. Procedural fairness 3.49 1.03 -- 2. Outcome fairness (OF) 4.19 0.93 .44 -- 3. Length of process (mOnths) (LP) 9.71 6.86 -.36 -.44 -- 4. OF • LP interaction 37.94 23.36 -.22 .01 .85

aAll coefficients with an absolute value greater than �9 = .2 are significant to p < 0.01.

Information lnaccessibifity 335

0.01. The Outcome Fairness x Length of Process interaction accounted for an additional 2.2% of the variance in procedural fairness, F(3, 176) = 5.23, p < 0.05. The beta for the interaction term was negative (13 = -.187). Thus, the effect was in the expected direction: The relationship between outcome fairness and procedural fairness was shown to be weaker for employees who experienced a longer relocation process than for those experiencing a shorter process. 'l-his finding confu'ms Hypothesis 2. Outcome fairness appears to have a stronger effect on procedural fairness when procedural information is relatively inaccessible.

The Outcome Fairness x Length of Process interaction is depicted graphically in Fig. 1. The three regression lines in the graph were derived from a single regression formula (see Cohen and Cohen, 1983). 5 An in- spection of the slopes for the lines indicates that outcome fairness is more strongly related to procedural fairness when the relocation process is short than when it is long.

STUDY 1 DISCUSSION

Hypothesis 1 in this study predicted that outcome fairness would influence procedural fairness judgments. Hypothesis 2 predicted that the influence of outcome fairness on procedural fairness would intensify when procedural information was inaccessible. Both hypotheses were supported. The finding of a moderating effect for information accessibility on the relationship between outcome and procedural fairness suggests a boundary condition which describes when outcomes will play a greater role than procedural characteristics in predicting procedural fairness. As Tyler and Lind (1992) suggest, outcome information is likely to be used instead of information about procedures to evaluate procedural fairness when procedural information is inaccessible.

As mentioned above, the measure of outcome fairness used here focused on the fairness of the decision, a practice sometimes used in the measurement of outcome fairness in legal settings. A limitation of our outcome fairness measures lies in their wording: Neither of the outcome fairness items makes use of the word "fairness" per se, but rather each relies on phrases chosen as synonyms for fairness (i.e., "Management had every right . . ."; "Management was perfectly jus t i f ied . . . " ) . That practice was pursued in an effort to reduce common method variance within items re-

5The groupings for short, medium, and long process were as follows. In this sample, a site was characterized (i) as having a short process if the time interval from announcement to move was 3 months or less, (ii) as having a medium process if the interval was 9-12 months, and (iii) as having a long process if the interval was greater than 24 months.

336 Daly and Tripp

4 . 5 - -

4 n

Procedural Fairness

3 . 5 - -

3 - - - -

2 . 5 -

3

Outcome Fairness

Short process

Medium process

Long process

Short process: Y= .58(Outcome Faimess) +1.29 Medium process: Y= .45(Outcome Fairness) + 1.64

Long process: Y= .3 l(Outcome Fairness) + 1.98

I I I 4 5

Fig. 1. Interaction effect of outcome fairness and length of relocation process on procedural fairness (Study 1).

la t ing to the same const ruct , however it may have had the u n i n t e n d e d effect o f obscur ing the m e a n i n g o f those i t ems somewhat . 6

S tudy 1 has m o r e substant ia l l imi ta t ions still. I ts connec t ion o f l ength o f dec i s ion -mak ing process to in fo rmat ion inaccessibi l i ty is t enuous . T h a t is, the measu re , whi le serving to avoid c o m m o n m e t h o d var iance , is s o m e - w h a t d is ta l f rom the cons t ruc t and m a y car ry in fo rmat ion a b o u t va r i ab les

6One source of confusion could be the use of the word, "justification." Folger's (1986) Referent Cognitions Theory (RCT) proposes that justification is an element of fair procedures rather than fair outcomes. However, if the items we label as outcome fairness tapped into the procedural fairness construct instead/more, we would expect a higher correlation with the four items we do label as procedural fairness than the r = .44 we found. It may be then that respondents interpret justification to mean a justified outcome, in that the outcome is fight or fair, rather than a justified procedure in that good procedures were followed. At a minimum, if these justification phrases in the two items do not capture outcome fairness per se, they should capture antecedents to outcome fairness.


unrelated to information inaccessibility. For example, length of process is confounded with the company for which the respondents worked. Thus, length of process may tap into variances in management practices unrelated to the amount of procedural information relocated respondents received.

In addition, the length of the relocation process might itself be viewed as a standard of procedural fairness, that is, degree of advance warning (used by Brockner et at, 1994, to operationalize procedural fairness). The Brockner et aL ,(1994) study described a process by outcome interaction effect that has been seen in a number of studies (see Brockner & Wiesen- feld, 1996, for a review), such that procedural fairness restrains the effect of outcome negativity on recipient responses. However, among the respondents in our Study 1, length of process was negatively correlated with procedural fairness (r = -.36, p < 0.01). That is, respondents tended to view a longer process as less fair, perhaps reflecting an aversion to "dragging it out." It is theretbre unlikely that our Study 1 respondents viewed length of the relocation process in the same terms as those of Brockner et a t (1994), and the effect uncovered in that study appears to be different from the effect reported here.

A final limitation of Study 1 lies in its correlational design. Given that all of the key measures were gathered at the same time, the results of Study 1 do not allow us to make causal inferences. An experimental design is needed to overcome that limitation.

To deal with the above described limitations of Study 1, we undertook a laboratory study to test the same hypotheses as those tested in Study 1. In a laboratory study, more proximal representations of information inaccessibility can be used and causal direction can be assessed.

STUDY 2 METHOD

Sample

Seventy-nine college students taking business courses at both the undergraduate and MBA level participated in Study 2. Sixty-one percent of the students were male, with an average age of 30.8 years.

Design

Participants each read one of six fictional vignettes as part of a management class lesson. The vignettes were presented as newspaper articles. Each article reported an allocation decision made by a university admini-

338 Daly and Tripp

stration regarding a business college curriculum change that could seriously affect the number of extra courses students would have to take to graduate.

To make the articles more believable, and to not unduly frighten some students, we chose another, unnamed university about whose business college students knew nothing. Thus, the participants did not believe that the outcomes affected them and only made judgments as third-party observers. However, we chose the college environment so that students could under- stand and identify with the allocation recipients.

Students were given 5 rain to read the articles, and then asked to complete a three item questionnaire. Immediately following, students were debriefed and the vignettes were discussed as part of a class-lesson.

Independent Variables

The six vignettes varied among two dimensions/variables. First, the outcome was either fair or unfair. In the fair outcome condition, students were excluded from complying with the curriculum changes by a grandfather clause. Also, the grandfather clause applied to all business students. In the unfair condition, marketing business students--and only marketing business students; no other business students--had to comply with the curriculum changes, which meant taking (and perhaps paying for) four more courses, regardless of how close to graduation they were. Essentially, in the unfair outcome condition there was an unexpected burden, and this burden was not equally shared. This outcome should be judged as distributively unfair.

Second, the articles contained either (i) no information about the pro- eedures, (ii) contained information that the procedures were fair, or (iii) contained information that the procedures were unfair. In the fair procedure condition, the articles stated that (i) the decision was made based on the advice of an outside business school accreditation body, and (ii) the dean's goal was for the new courses to benefit the students. In the unfair procedure condition, the articles stated that (i) the decision was made by the business school administrators themselves, and (ii) the real goals of the clean were to increase enrollment, which would increase monies available to the dean to remodel his office, and to get the new curriculum changes published in Business Week to make the dean look good. Essentially, in the unfair procedure condition, the dean was portrayed as extremely biased in a serf-serving fashion, which Leventhal (1980) and others (see Lind and Tyler, 1988) have shown to be judged as procedurally unfair. See the Ap- pendix for the actual vignettes.


Dependent Variables

Participants rated the vignettes on three items. Each item used a 7- point Likert scale with anchors 1 (strong~ disagree), 4 (neither agree nor disagree), and 7 (strongly agree). The first item, "The new policy is a fair policy to implement," measured their judgment of overall fairness. The second item, "The decision-making procedure used to make the new policy is fair," measured their perception of procedural fairness. Finally, the third item, "The outcome is fair," measured their perception of outcome fairness.

STUDY 2 RESULTS

A 2 x 3 analysis of variance (ANOVA) (2 Outcome Fairness x 3 Pro- cedural Information) was performed on ratings of procedural fairness. The analysis revealed: a significant main effect of procedural information, F(2, 73) = 4.58, p < 0.02; a nonsignificant main effect of outcome, F(1, 73) = 0.05, ns; and a marginally significant effect of the interaction, F(2, 73) = 2.63, p < 0.08. The distribution of means (Fig. 2) was in the predicted pattern, i.e., a steep slope for the no procedural information condition and relatively shallow slopes for the two procedural information conditions.

To verify that the outcome fairness-procedural fairness relationship for the "no procedural information" condition was indeed different from those for the remaining procedural fairness conditions, we estimated zero-order rs for the no information condition and the aggregate of the two conditions in which procedural information was given and tested those for differences. In the no procedural information condition, r = .43 (n = 26, SD of procedural fairness = 1.27). In the procedural information present condition, r = -.14 (n = 53, SD of procedural fairness = 1.66). The two regression coefficients were significantly different from one another (Z = 2.39, p < 0.02). With the two procedurai information conditions collapsed, we performed a second ANOVA. It revealed nonsignificant main effects for procedural information, F(2, 75) = 0.72, ns, and outcome, F(1, 75) = 0.04, ns; and a significant interaction effect, F(1, 75) = 4.46, p < 0.03. Because of the predicted and actual high correlation between the procedural fairness and distributive fairness dependent variables (r = .58, p < 0.001), an ANCOVA was run using distributive fairness ratings as the covariate. The interaction of procedural information and outcome fairness manipulations on the procedural fairness ratings dependent variable remained significant, F(1, 78) = 4.84, t ' < 0.03.

Figure 2 shows that when procedural information was absent there was a large difference between unfair outcomes (~ -- 2.69, SD = 1.18) and

340 Daly and Tripp

Procedural Justice Ratings

[ ] Procedural Information- Fair Procedure

[] Procedural Information- Unfair Procedure

O No Procedural Information

I I Unfair Fair

Distributive Justice

Fig. 2. Interaction effect of distributive justice and lack of procedural information on procedural justice ratings (Study 2).

fair outcomes (~" = 3.77, SD = 1.16) on ratings of procedural fairness. A t test of the difference between the means was significant, t(1, 24) = 2.34, p < 0.03).

Manipulation Checks

The significant main effect of procedural information on ratings of procedural fairness confirms that participants did perceive the fair proce-


dure condition (~ = 3.77, SD = 0.23) as fairer than the no procedural information condition (~ = 3.23, SD = 0.28), which they perceived as fairer than the unfair procedure condition (~ = 2.57, SD = 0.28). However, a t test of outcome fairness on ratings of outcome fairness was nonsignificant, t(76) = 0.18, ns.

STUDY 2 DISCUSSION

Hypothesis 1 in this study predicted that outcome fairness would influence procedur-,d fairness judgments. Hypothesis 2 predicted that the influence of outcome fairness on procedural fairness judgments would intensify when procedural information was inaccessible. Only Hypothesis 2 received support.

Hypothesis 1 did not receive support most likely because participants could not tell (or deem important) the difference between the actual outcomes provided in the vignettes in terms of outcome fairness, perhaps because they acted only as observers, not as recipients. Nevertheless, as noted above, Fig. 2 shows that when procedural information was absent there was a large difference between unfair outcomes and fair outcomes on ratings of procedural fairness and the correlation coefficient between outcome fairness and procedural fairness for the "no procedural information" cases was substantial.

Study 2 demonstrates that people can judge the procedural fairness of an allocation decision by its outcome fairness. More precisely, this more likely will occur when they are deprived of information about the procedures. Much as Brockner and Wiesenfeld (1996)suggest, when people try to make sense of an allocation decision and consider its procedural merits, lack of information about procedures will not stop them from passing judgment on procedures (perhaps, sometimes, because the experimenter has asked them to pass judgment). When passing judgment, they may use what- ever limited information they have, even if it is only knowledge of the outcome.

However we do not suggest that the finding here replicates the process by outcome interaction that Brockner and Wiesenfeld (1996) document. They found that procedural fairness judgments can mitigate the effect of outcome negativity on recipient responses. However, we found that the interaction is more complex. That is, the initial judgment of the fairness of the procedures will depend upon the outcome's negativity when procedural information is lacking.

342 Daly and Tripp

GENERAL DISCUSSION

Study 1 found that in relocation decisions the length of time for the relocation decision was related to workers' ratings of the fairness of the relocation procedures. It also found that the length of time for the relocation process moderated the relationship between ratings of outcome fairness and ratings of procedural fairness. We theorized that what explained these findings was information inaccessibility: A shorter relocation decision process provides workers with less information about procedures, which in turn makes them more likely to more strongly draw inferences about procedures based purely on what they do know--the outcomes.

Study 2 directly tested that explanation in a laboratory. It conftrmed that when people have no information about procedures, they do use outcome information to form opinions about procedures. While limitations in both studies are noted above, we believe that, to a large extent, weaknesses in one study are compensated by strengths in another and the limitations of each study are nonoverlapping. For example, the use of an artificial setting in Study 2 is compensated for by the naturalistic setting of Study 1. Taken as a whole, the multiple studies reported encourage confidence in their findings.

When those judging the process behind a particular decision lack knowledge regarding how the decision was made, that does not necessarily imply that they do not have information about any procedural characteristics that have been held to be procedural fairness antecedents in the justice literature. Voice, for example, may be conspicuous by its absence when it is expected. However, a lack of voice was not likely to be seen by our Study 1 respondents as unfair. An earlier analysis of data from those respondents found that voice had no effect on procedural fairness for that sample (Daly and Geyer, 1994). Supplementary analyses by Daly and Geyer suggested that the absence of an effect was due to expectations--that is, that employee voice in relocation decisions is generally not expected and therefore an absence of voice was not seen to be unfair in that context. The finding is consistent with Greenberg et aL (1991), who found in a laboratory study that voice effects were lacking when expectations of voice were not supported by norms. Had voice been expected by respondents, it is conceivable that the absence of voice would have served as a means of evaluating procedural fairness apart from decision outcomes. The moderating effect of information inaccessibility on the relationship between outcome fairness and procedural fairness perceptions reported here may therefore be weaker in settings where proposed procedural fairness antecedents (e.g., voice) are expected by employees and are made salient by their absence.


The present study raises cautions about generalizing findings from studies of fairnes.~; in legal, political, and dispute resolution settings to organizational decisJions. That is not to imply that most fairness findings from other settings do not generalize to organizations, only that setting-specific factors ought to be taken into account in the specification and testing of theory in organizational justice. More specifically, in studies of organizational justice we cannot assume that outcome recipients have the same degree of access to procedural information that is typically enjoyed by recipients in legal, political, and dispute resolution settings. For example, some organizations make s o m e decisions through commit tees , teams, and/or task forces. These groups often represen tand communicate with various organizational constituencies. At other times or in other organizations, however, top management makes decisions behind closed doors, only to announce unalterable outcomes with little explanation of the procedures and criteria involved in reaching those decisions. We argue that when as- sessing perceptions of procedural fairness decisions in the field, it is important to know which setting is which.

Identifying information inaccessibility as a moderator with regard to the effect of outcome fairness on procedural fairness promises to be an aid in interpreting results from organizational field studies. As mentioned above, outcome fairness and procedural fairness often have been found to be highly related in organizational settings, while in legal and political settings, by contrast, the relationships between the two types of fairness judgments are often moderate to weak. However, as argued above, the decision process in most legal settings (and many political settings in the U.S.) is required to be largely open. and to follow the many procedural rules embodied in procedural law, which has it roots in the due process clause of the U.S. Constitution. The U.S. laws pertaining to corporate governance do not impose suc'.h requirements on managerial decision processes.

Implications for Practice

This study's findings regarding the effects of outcome fairness on procedural fairness have important implications for managerial practice. When managers keep their subordinates in the dark as to how a decision was made, their subordinates are likely to focus more heavily on outcome fairness in their responses to the decision. Keeping subordinates uninformed may invite subordinates to make inaccurate inferences about the procedures involved. That is, lacking evidence to the contrary, subordinates who receive unfair outcomes--outcomes that they do not see as just, ff only because the outcomes are unfavorable or unexpected--will judge the pro-

344 Daly and Tripp

cedures as unfair. They could become even more upset. Bies and Shapiro (i987, 1988) discussed the importance of account giving. Managers give accounts to prevent such negative judgments and attributions. Bies et al. (1997) study of revenge in the workplace also shows that upon receiving bad news or feeling harmed, workers search for an explanation, and invent attributions or pass judgment on the decision-makers, even in the total absence of information about the decision-making process. Such attributions (even uninformed ones) are part of the revenge-seeking process.

Where resources are scarce, it may be possible for managers to enhance employee knowledge of fair procedures, especially when they cannot deliver favorable outcomes to employees. Also, managers can inform employees of the fair decision-making procedures before decisions are made. If the managers get employee buy-in to such procedures, then employees may be predisposed to judge all outcomes that result from it as fair. That is, in the absence of outcome information, judgment of the procedures may influence the judgment of the outcome, especially when employees have no information about the outcomes.

In addition to (partially) compensating for the dissatisfaction of receiving poor outcomes, providing information about fair procedures should enhance a variety of outcomes desired by managers, including organizational commitment (Forger and Konovsky, 1989), satisfaction with super- vision (Alexander and Ruderman, 1987), and organizational citizenship behavior (Moorman, 1991). If managers want to reap the benefits of fairness but cannot afford the expense of providing highly favorable outcomes, they may be advised to focus on providing procedural information.

Implications for Theory and Research

As the studies in this special issue signify, there is growing interest in the relationship between distributive and procedural fairness among procedural fairness researchers. The authors are grateful to the editors of this special issue for having suggested the questions enumerated in this section. The present study raises many questions about the relationship between the two types of fairness judgments which could prove fruitful as points of departure for future research. One such question is, if people tend to use outcome fairness to infer procedural fairness when procedural information is absent, how then do they make outcome fairness judgments in advance of outcome information? Much as we speculate in the above section, do they use procedural fairness under those circumstances as a proxy for outcome fairness?

Information Inaccessibifity 345

A related question is, ff people are constantly scanning their environ- merits, looking for cues by which to make fairness judgments (as Brockner and Wieserffeld, 1996, argue), why is it so important for them to make such judgments even, when the relevant information is not available? One an- swer may lie in Lind et al.'s (1993) concept of a fairness heuristic, that is, the idea that people rely on fairness judgments to determine if an authority's decision is legitimate. Lind (1994) stated, "Fairness judgments function as a social diagnostic, that they serve as a cognitive device by which people can assess how things stand in their group and as a heuristic that can guide a person's resolution of the fundamental social dilemma" (p. t0). Lind de- fined the fundamental social dilemma as the choice an individual makes as to how much to subordinate one's individual interests to the larger group's or community's interests.

The notion that fairness is important to people is certainly consistent with our findings: People make strong judgments about fairness even with- out information. Researchers should explore further attribution processes involved in estimating one kind of fairness from another, including all the attribution errors involved and their consequences.

APPENDIX

Vignettes Used in Study 2

The vignettes were presented as articles clipped from a newspaper, accompanied by' graphics and a byline for a fictitious author to enhance believability. For each experimental condition, the opening paragraph of the vignette was as follows:

There is a movement among business schools to radically alter their course offer- ings. Many business schools plan to require some of their students to take foreign language courses.

The remaining sentences of the vignette varied by experimental condition. The specific vignettes for each condition are given below. (Note: All ma- terials were administered to study participants during the calendar year 1994.)

I. Unfair Outcome, No Procedural Information Condition

At [named crossed out] University, the administration decided last week to implement the following policy. All undergraduate marketing students who are graduating in 1994 or later will have to complete four courses in Japanese language, which is

346 Daly and Tripp

a very difficult language to learn. This policy applies to all currently enrolled marketing students, but not to other business students.

Z Fair Outcome, No Procedural Information Condition

At [named crossed out] University, the administration decided last week to implement the following policy. All undergraduate business students who are graduating in 1994 or later will have to complete four courses in Japanese language, which is a very difficult language to learn. Thus, all currently enrolled students do not have to take the four Japanese courses, although they may take the courses voluntarily.

3. Fair Outcome, Procedural Information Indicating Fair Procedure


The decision was based on the advice of the American Association of Collegiate Schools of Business--the main business school accreditation agency that sets course requirements. Although requiring the new language courses will provide additional monies for the business school, Dr. [name crossed out], the business school dean, has publicly stated that by far the biggest reason for adopting the policy is to encourage students to develop foreign language skills to deal with the increasingly global work environment.

4. Fair Outcome, Procedural Information Indicating Unfair Procedure


The decision was based on the advice of business school administrators, who indicated at a public meeting on the university campus last Thursday that their reason for adopting the policy is that students need to develop foreign language skills to deal with the increasingly global work environment. However, this newspaper has learned that the main reasons for the policy really are: 1) the projected enrollment from these courses will provide additional monies for Dr [name crossed out], the business school dean, to remodel his office, and 2) rumor has it that the dean believes the new policy will be written up in Business Week, which could give a boost to his career.

Information Inaccessibility

5. Unfair Outcome, Procedural Information Indicating Fair Procedure

At [named crossed out] University, the administration decided last week to implement the following policy. All undergraduate marketing students who are graduating in 1994 or later will have to complete four course.s in Japanese language, which is a very difficult language to learn. This policy applies to all currently enrolled marketing students, but not to other business students.

The decision was based on the advice of the American Association of Collegiate Schools of Business--the main business school accreditation agency that sets course requirements. Although requiring the new language courses will provide additional monies for the business school, Dr. [name crossed out], the business school dean, has publicly stated that by far the biggest reason for adopting the policy is to encourage students to develop foreign language skills to deal with the increasingly global work environment.

347

6. Unfair Outcome, Procedural Information Indicating Unfair Procedure

At [named crossed out] University, the administration decided last week to implement the following policy. All undergraduate marketing students who are graduating in 1994 or later will have to complete four courses in Japanese language, which is a very difficult language to learn. This policy applies to all curremly enrolled marketing students, but not to other business students.

The decision was based on the advice of business school administrators, who indicated at a public meeting on the university campus last Thursday that their reason for adopting the policy is that students need to develop foreign language skills to deal with the increasingly global work environment. However, this newspaper has learned that the main. reasons for the policy really are: 1) the projected enrollment from these courses will provide additional monies for Dr. [name crossed out], the business school dean, to remodel his office, and 2) rumor has it that the dean believes the new policy will be wdttan up in Business Week, which could give a boost to his career.

A C K N O W L E D G M E N T S

T h e r e sea rch r e p o r t e d in Study i was suppo r t ed by a g ran t f rom the D i s p u t e Reso lu t ion Resea rch Cen te r at Nor thwes te rn Univers i ty .

T h e au thors owe thanks to Bob Bies, Joe l Brockner , Bor i s Kabanof f , D e n i s e Rousseau , and this issue 's coedi tors for the i r c o m m e n t s and suggest ions .

R E F E R E N C E S

Alexander, S., and Rudennan, M. (1987). The role of procedural and distributive justice in organizational behavior. Soc. Justice Res. 1: 177-198.

348 Daly and Tripp

Berie, A., and Means, G. (1932). The Modern Corporation and Private Property, Macmillan, New York.

Bies, R. J., and Shapiro, D. L. (1987). Interactional fairness judgments: The influence of causal accounts. Soc. Justice Res. 1: 199-218.

Bies, R. J., and Shapiro, D. L. (1988). Voice and justification: Their influence on procedural fairness judgments. Acats Manage. J. 31: 676-685.

Bies, R. J., Shapiro, D. L., and Cummings, L. L. (1988). Causal accounts and managing conflict: Is it enough to say its not my fault? Communicat. Res. 15: 381-399.

Bies, R. J., Tripp, T. M., and Kxamer, R. M. (1997). At the breaking point: A social cognitive perspective on vengeance and violence in organizations. In Greenberg, J., and Giacalone, R. (eds.), Antisocial Behavioral in Organizations, Sage, Newbury Park, CA, pp. 18-36.

Brockner, J., Konovsky, M., Cooper-Schneider, R., Folger, R., Martin, C., and Bies, R. (1994). Interactive effects of procedural justice and outcome negativity on victims and survivors of job loss. Acad. Manage. J. 37: 397-409.

Brockner, J., and Wiesenfeld, B. (1996). Integrative framework for explaining reaction to decisions: Interactive effects of outcome and procedures. PsychoL Bull. 120: 189-208.

Casper, J., Tyler, T., and Fisher, B. (1988). Procedural justice in felony cases. Law Soc. Rev. 22: 483-507.

Cohen, J., and Cohen, P. (1983). Applied Multiple Regression Correlation Analysis for the Behavioral Sciences, 2nd ed., Erlbaum, Hillsdale, NJ.

Conlon, D., Lind, E. A., and Lissak, R. (1989). Nonlinear and nonmonotonic effects of outcome on procedural and distributive fairness judgments. J. Appl. Soc. PsychoL 19: 1085-1099.

Daly, J., and Geyer, P. (1994). The role of fairness in implementing large-scale change. Employee evaluations of process and outcome in seven facility relocations. J. Organ. Behav. 15: 623-638.

Ewing, D. (1977). Freedom Inside the Organization: Bringing Civil Liberties to the Workplace, McGraw-Hill, New York.

Folger, R. (1986). Rethinking equity theory: A referent cognitions model. In Bierhoff, H. Cohen, R., and Greenberg, J. (eds.), Justice in Social Relations, Plenum Press, New York, pp. 145-164.

Folger, R., and Konovsky, M. A. (1988). Procedural justice, distributive justice, and reacaons to pay raise decisions. Unpublished manuscript, Tulane University New Orleans, LA.

Fryxell, G., and Gordon, M. (1989). Workplace justice and job satisfaction as predictors of satisfaction with union and management. Acad. Manage. J. 32: 851-866.

Gottlieb, M. (1955, October 4). Deep split in reactions to the verdict. The New York Times, pp. A1, A12.

Greenberg, J. (1987). A taxonomy of organizational justice theories. Acats Manage. Rev. 12: 9-22.

Greenberg, J. (1990). Organizational justice: Yesterday, today, and tomorrow. J. Manage. 16: 2, 399-432.

Greenberg, J., Eskew, D., and Miles, J. (1991), August). Adherence to participatory norms as a moderator of the fair process effect: When voice does not enhance procedural justice. Paper presented at the Academy of Management, Miami, FL.

Homans, G. C. (1961). Social Behavior:. Its Elementary Forms, Harcourt Brace Jovanovich, New York.

Konovsky, M., and Cropanzano, R. (1991). Perceived fairness of employee drug testing as a predictor of employee attitudes and job performance. J. AppL PsychoL 76: 698-707.

LaTour, S. (1978). Determinants of participant and observer satisfaction with adversary and inquisitorial modes of adjudication. J. Pets. Soc: P~ycho/. 36: 1531-1545.

Leventhal, G. (1980). What should be done with equity theory? New approaches to the study of fairness in social relationships. In Gergen, IC, Grcenberg, M., and Willis, R. (eds.), Social Exchan~: Advances in Theory and Research, Plenum Press, New York, pp. 2%55.

Lind, E. A. (1994). Procedural justice and cultu~" Evidence for ubioutous process concerns. Working paper, American Bar Foundation.


Lind, E. A., Kulk, C., Ambrose, M., and Park, M. (1993). Individual and corporate dispute resolution: Using procedural fairness as a decision heuristic. Admin. ScL Quart. 38: 224-251.

Lind, E. A., Kurr.z, S,, Musante, L, Walker, L, and Thibant, J. (1980). Procedure and outcome effects on reactions to adjudicated resolution of conflicts of interests. J. Pets. Soc. Psychol. 39: 643-653.

Lind, E. A., and Lissak, R. (1985). Apparent impropriety and procedural fairness judgments. J. Exp. SOc. PsychoL 21: 19-29.

Lind, E. A., and Tyler, T. R. (1988). The Social Psychology of Procedural Justice, Plenum Press, New York.

MacCoun, R. J., Lind, E. A., Hensler, D. R., Bryant, D. L., and Ebener, P. A. (1988). Alternative Adjudication" An Evaluation of the New Jersey Automobile Arbitration Program, Institute for Civil Justice, Rand Corporation, Santa Monica, CA.

Moorman, R. H. (1991). Relationship between organizational justice and organizational citizenship behaviors: Do fairness perceptions influence employee citizenship? J. Appt PsychoL 76: 6, 845-855.

Paese, P. (1985). Procedural fairness and work group responses to performance evaluation procedures. Unpublished masters thesis, University of Illinois, Champaign.

Rousseau, D. (1985). Issues of level in organizational research: Multi-level and cross-level perspectives. In Cummings, L. L., and Staw, B. (eds.) Research in Organizational Behavior, Vol. 7, JAI Press, Greenwich, CT, pp. 1-37.

Shapiro, D. L. and Brett, J. M. (1993). Comparing three processes underlying judgments of procedural justice: A field study of mediation and arbitration. J. Pers. Soc. PsychoL 56: 1167-1177.

Stohl, C., and Redding, W. (1987). Messages and message exchange processes. In Jablin, F., Putnam, L., Roberts , K. and Porter, L. (Eds.), Handbook of Organizational CommunicatiorL" An Interdisciplinary Perspective, Sage, Beverly Hills, CA, pp. 45-502.

Stohr, R. (1975). Legal responsibilities, liabilities, and protection of directors In Bacon, J., and Brown, J. (eds.), Corporate Directorship Practices: Role, Selection and Legal Status of the Board, The Conference Board, New York.

Sweeney, P., and McFaflin, D. (1993). Workers' evaluations of the "ends" and the "means": An examination of four models of distributive and procedural justice. Organ. Behav. HurrL Decision Process. 55: 23-40.

Thibaut, J., and Walker, L. (1975). Procedural Justic~ Edbaum, Hillsdale, NJ. Tyler, T. (1984). The role of perceived injustice in defendant's evaluations of their courtroom

experience. Law Soc. Rev. 18: 51-74, Tyler, T., and Lind, E. A. (1992). A relational model of authority in gro~ups. In Zanna, M.

(ed.), Advances in Experimental Social Psychology, Vol. 25, Lexington Books, New York, pp. 115-191.

Tyler, T., Rasinski, K., and McGraw, IC (1985). The influence of perceived injustice on support for political authorities. J. Appt Soc. PsychoL 15: 700-725.

Walker, L, LaTour, S., Lind, E. A., and Thibaut, J. (1974). Reactions of participants and observers to modes of adjudication..L AppL SOc. PsychoL 4: 295-310.

Documents

Is outcome fairness used to make procedural fairness judgments when procedural information is inaccessible?