What are four important tactical tasks for a negotiator in a distributive situation to consider?

® Academy of Managemenl /ournal 1990, Vol. 33, No. 4. 833-846.

INFLUENCE OF SPEED OF THIRD-PARTY INTERVENTION AND OUTCOME ON NEGOTIATOR AND CONSTITUENT

FAIRNESS JUDGMENTS

DONALD E. CONLON University of Delaware

PETER M. FASOLO Human Resource Associates

Using a simulated organizational dispute, we tested the impact of a third-party's intervention on disputants' perceptions of procedural and distrihutive justice and satisfaction with the third party. We manipu- lated disputants' roles in the negotiation, the speed of third-party in- tervention, and the outcome imposed by the third party in a two- by-two-by-five factorial design. The results suggest that the speed of third-party intervention influenced perceptions of procedures more than perceptions of outcomes. Speed of intervention also influenced disputants differently depending on their role. Outcome influenced all measures of procedural and distributive fairness. We discuss implica- tions for managerial behavior in dispute resolution and consequences for the measurement of procedural justice.

Research on dispute resolution and resource allocation has identified two dimensions of fairness that are important to participants in disputes. Early research focused on distributive justice, or equity, the perceived fair- ness of the outcomes of a dispute. More recently, researchers have studied procedural justice, which is the perceived fairness of the procedures used to generate outcomes. The present study examined the influence of the speed of third-party intervention on disputants' perceptions of distributive and procedural justice.

Researchers have advanced theories of procedural justice based on work in legal (Thibaut & Walker, 1975, 1978] and organizational settings (Lev- enthal, Karuza, & Fry, 1980]. In their theory of procedure, Thibaut and Walker distinguished between process control, or the degree to which dis- putants exerted influence over a resolution procedure, and decision control, or the degree to which disputants exerted influence over the final outcome of a dispute. Leventhal and colleagues' allocation preference theory identi- fied a number of dimensions that influenced procedural justice, many of which overlapped with the dimensions Thibaut and Walker iden-

The authors would like to thank Allan Lind, Howard Garland, and this journal's reviewers for their comments on earlier drafts of this article. Thanks are also due to John Scott, Steven Roth, Cary Polkowitz, and Karin Muenning for their assistance in the data collection.

833

834 Academy of Management Journal December

tified (Lind & Tyler, 1988]. Researchers investigating procedural justice in organizations have examined performance evaluation and job transfers (Lis- sak, 1983], pay raise decisions (Folger & Konovsky, 1989], and dispute res- olution (Karambayya & Brett, 1989; Sheppard, 1984; Sheppard & Lewicki, 1987]. Many of the same effects found in legal contexts have also occurred in organizational contexts: for example, in both laboratory and field settings, people have seen performance evaluation procedures that provide a high level of voice, or personal expression (a component of process control], as more fair than procedures that limit input (cf. Lind & Tyler, 1988].

Some recent studies have suggested that the manner in which a proce- dure is enacted or interpreted plays an important role in subsequent evalu- ations of fairness. For example, using concepts found in research on lead- ership, Ross, Conlon, and Lind (1990] found that disputants involved with task-oriented mediators were more satisfied with mediation when the dead- line for a settlement was vague rather than explicit. Disputants involved with person-oriented mediators were equally satisfied whether the deadline was known or not. A similar pattern was found for disputants' distributive fairness ratings.

Ross and colleagues (1990] were the first to examine variation within a mediation procedure by varying the behavioral style of third parties. As Sheppard (1984] noted; there has been little research on disputants' reac- tions to variation withii*^ procedure. Assuming that other variations in enactment may influence participants' perceptions of distributive and pro- cedural justice, we examined one such variable, the timing of third-party intervention.

As McGrath and Rotchford (1983] noted, time is a pervasive aspect of behavior in organizations. Time can also play a prominent role in negotia- tions: disputants often set deadlines for reaching agreements or stall before making a concession in the hope that opponents will concede first. Shep- pard and Lewicki's (1987] study of 44 executives found that the timeliness of third-party intervention was an important aspect of fairness for managers. Views of negotiation put forth in previous research have suggested that, all thing being equal, people prefer procedures that can settle disputes quickly to those that cannot because the costs to all parties increase as time elapses (Ury, Brett, & Goldberg, 1988]. But can a dispute be settled too quickly?

Previous views of distributive justice, social exchange, and public choice (e.g., Adams, 1965; Walster, Walster, & Berscheid, 1978] have sug- gested that satisfaction with procedures, outcomes, and third parties is pri- marily, if not exclusively, a function of outcomes; thus, whether a settlement is reached quickly or slowly is unimportant. On the other hand, procedural justice theory and research (e.g., Lind & Tyler, 1988] have suggested that the speed of third-party intervention could have a significant impact on dispu- tants' perceptions by altering their opportunity to exercise voice. If a third party intervenes quickly in a procedure, it will allow little opportunity for disputants to express their opinions and may be seen as less fair than a longer procedure. In addition, disputants typically do not desire third-party

1990 Conlon and Fasolo 835

intervention unless they realize that they cannot resolve a dispute them- selves (Pruitt, 1981], and this realization may not yet have occurred if inter- vention is fast. Lastly, Greenberg's (1990] work suggested that rapid inter- vention may not appear to be as fair because it violates people's expectations of what a fair procedure should be.

Hypothesis 1: The quick imposition of an outcome by a third party wiJJ lead disputants to report low levels of procedural justice, distributive justice, and satis/action with the third party.

The present study also varied the outcome third parties imposed on disputants. Although most of the research on legal procedures has focused on all-or-nothing outcomes, agreements in which both parties receive some amount of the resource in question are typical in organizations. Little pre- vious research has examined variation in outcomes beyond total win and total loss.^ The present study employed five different outcomes that varied parametrically in their favorability to subjects.

When maintaining social harmony is important, dividing rewards or outcomes equally among disputants is likely to be the preferred distribution norm, and individuals are likely to be uncomfortable with exceedingly large rewards gained at the expense of the other disputants (Greenberg, 1987; Walster et al., 1978]. Thus,

Hypothesis 2: Although outcome satisfaction wiJJ in- crease as the value of an outcome increases, disputants who are co-workers wiJJ not see excessively favorable out- comes as fair.

This study also examined perceptions of satisfaction and fairness for two types of disputants. We called disputants who participated in a dispute resolution procedure "negotiators" and called those who did not participate but were affected by an outcome "constituents." The relationship between constituents and negotiators is similar to the principal-agent relationship discussed in agency theory (e.g., Eisenhardt, 1988], as constituents delegate the work of negotiating to negotiators. Gomparing these two types of dispu- tants allowed us to compare two roles whose occupants have similar stakes in an outcome—though negotiators may also have other interests at stake, such as prestige. In addition, most organizational disputes, both labor nego- tiations and less formal procedures, involve only a few employees in nego- tiations; the others must wait to hear the outcome and do not experience the procedure firsthand. Thus, studying these two roles seemed of practical importance.

Procedural justice studies have rarely manipulated roles, typically com- paring the reactions of participants in and observers of procedures (e.g., Latour, 1978]. The literature on procedural justice and participation suggests that a high level of participation in properly enacted procedures enhances

' Conlon, Lind, and Lissak (1989) is an exception.

836 Academy of Management Journal December

perceptions of fairness, even when participants have no control over out- comes (Musante, Gilbert, & Thibaut, 1983].

Hypothesis 3: Negotiators will perceive themselves as having higher levels of control over decisions and voice than constituents.

Although we expected negotiators to report greater voice and decision control than constituents, we expected constituents to see distributive jus- tice, procedural accuracy, and third parties more favorably than negotiators because, lacking specific information about how a procedure was enacted, constituents will have an idealistic vision and assume that it was executed properly and competently. Negotiators, who personally experienced the be- haviors of opponents and third parties, will have more negative perceptions.

Hypothesis 4: Constituents will report higher levels of dis- tributive justice, procedural accuracy, and satisfaction with third parties than negotiators.

Finally, role was expected to interact with the speed of third-party in- tervention. We expected negotiators' perceptions to vary with the speed with which an outcome was imposed because speed affects their opportunity for voice. Constituents, with no opportunity for voice in either case, should be unaffected.

Hypothesis 5: Constituents' perceptions of decision con- trol, voice, and third-party propriety wiJJ show JittJe vari- ation related to the speed of third-party intervention. Ne- gotiators' perceptions of these values wiJJ be lower under fast intervention than under slow intervention.

METHODS

Subjects and Research Design

Subjects were 197 undergraduate business administration students who participated in return for class credit and a chance at several monetary prizes. We conducted simulations of an organizational negotiation for groups of 12 to 30 subjects. The study had a two-by-two-by-five factorial design. Subjects took the roles of disputants, either negotiators or constitu- ents; third-party intervention in the dispute was either slow or fast; and the outcome received ranged from extremely unfavorable to extremely favor- able. The number of subjects in the cells formed by these variations ranged from 8 to 10.

Procedures

Subjects were seated at personal computers that presented the instruc- tions, experimental manipulations, and a questionnaire. Experimenters, who were blind to the assignment of subjects to experimental conditions, told subjects that they would be role-playing an organizational negotiation

1990 Conlon and Fasolo 837

and that they would interact with other subjects via the computer network. All subjects were told they had been randomly assigned to work in the organization's Boston office. Those assigned to be negotiators were given the role "Boston bargaining rep" and those assigned to be constituents were given the role "Boston sales manager." The instructions then said that the organization's Boston office was in conflict with its New York office and that the bargaining representative's task was to resolve the conflict with another manager, the New York bargaining representative. Subjects also learned that a third manager, the product manager, could assist them in settling the

dispute. The scenario was based on an actual negotiation that occurred in a large

securities firm. The New York and Boston bargaining representatives nego- tiated three issues: the size of transferable accounts, eligibility for promo- tion, and criteria for promotion. Subjects received information about the issues on the computer screen and in a manila folder. Following past re- search (Garnevale & Gonlon, 1988], we presented information about the is- sues on the screen in tabular form. The actual table used is in Carnevale and Gonlon (1988].

For each issue, there were nine proposal levels labeled A to I. Subjects were told that the bargaining representatives needed to agree on one pro- posal level for each issue. Points shown next to each level indicated its value to the disputants. Level I gave the most points to subjects on every issue. Subjects saw both their own and their opponents' point values and were told that the product manager, the third party, would see both disputants' point values.

On each round of negotiation, the bargaining representatives first saw the New York representative's issue chart and a proposal composed of three letters. Subjects then saw their own issue chart and the personal outcomes associated with particular agreements. They were told that the points for outcomes would be converted to lottery tickets at the completion of the study and that the seven winning tickets would split $300. The instructions emphasized that chances increased as lottery tickets increased, so subjects should try to get as many points as possible.

Subjects were reminded that the product manager, who was described as superior in the organization's hierarchy to the managers in the New York and Boston offices, could intervene in the negotiation. We told them that although the product manager might not get involved in the negotiation at all, he could impose an outcome at any time. They were also told that the product manager had a history of acting fairly and responsibly in helping managers resolve conflicts.

For negotiators in all conditions, the proposals generated by the New York manager (the opponent] produced few points. When the third party was to intervene slowly, the opponent's proposals on the eight rounds of negotiation were labeled AAA, AAA, AAB, AAB, ABB, ABB, BAB, and BAB. When intervention was to be fast, the opponent's were AAA and BAB, the

838 Academy of Management Journal December

same as the first and last offers made by the opponent when intervention was slow. Thus, negotiators' opponents conceded an equivalent amount in both conditions.

Negotiators were free to make counteroffers and send messages to their opponent on each round. When intervention was slow, the computer inter- rupted negotiators after their opponent's eighth proposal and told them that the third party had decided to impose a settlement. When intervention was fast, the interruption occurred after the opponent's second proposal. After receiving an outcome, subjects completed a questionnaire that elicited their perceptions of procedural and distributive justice, the third party, and the opponent. After they completed the questionnaire, we told subjects that the study was finished and gave them a debriefing.

Independent Variables

Disputant role. Negotiators, assigned the role of the Boston bargaining representative, were reminded that they were negotiating not only for them- selves, but also for their co-worker, the Boston sales manager. In fact, we encouraged subjects to believe that many co-workers in the Boston office were depending on their negotiation skills. Gonstituents, assigned the role of the Boston sales manager, were told that another person in the room was their bargaining representative. During the negotiation, constituents were encouraged to review the material provided in their folders about the orga- nization, the dispute, and the dispute resolution procedure. While constit- uents waited for the dispute to be settled, a message flashed on their com- puter screen: "PLEASE STAND BY: YOUR BARGAINING REP IS STILL NEGOTIATING."

Speed. The manipulation of third-party intervention varied how many rounds the bargaining representatives negotiated with their opponent before the third party imposed a settlement. Bargaining representatives in the fast intervention condition were interrupted before they made a proposal on the second round of negotiation and told that the third party had decided to impose a settlement. In pretests, subjects reached the second round in about two minutes. Gonstituents in the fast intervention condition were thus kept waiting for about two minutes before their computer screen announced that the third party had decided to settle the dispute during the second round of the negotiation. Bargaining representatives in the slow intervention condi- tion were interrupted before they made a proposal on the eighth round of negotiation. Gonstituents in this condition waited about ten minutes for the announcement that the product manager had decided to settle the dispute, the amount of time negotiators in the slow intervention condition took to reach this point in pretests.

Outcome. In the zero outcome condition, subjects were told that the third party had imposed settlement AAA, which gave them no points but gave their opponent 240 points. In the one-quarter outcome condition, GGG was imposed, providing 60 points to a given subject and 180 to the oppo- nent. In the one-half outcome condition, settlement EEE gave each disputant

1990 Conion and Fasolo 839

120 points. In the three-quarters outcome condition, GGG gave 180 points to a subject and 60 to the opponent. In the everything outcome condition, III gave 240 points to a subject and no points to the opponent. Subjects were told to write their settlement and its point value on a card in their folder.

Dependent Variables

Subjects indicated their judgments about the fairness of the settlement and other perceptions on six- and seven-point rating scales for a number of measures that we constructed. A manipulation check for the speed of third- party intervention consisted of two questions asking how long it took the dispute to be resolved and how fast it was resolved (a = .76). We measured outcome satisfaction with two questions that asked subjects how satisfied they were with the outcome of the dispute [a = .94) and measured outcome fairness with two questions on how fair the outcome was (a = .89).

Three indexes measured procedural justice. Decision control was mea- sured by asking subjects how much control and how much influence they had over the decision that was made (a = .64). We measured process control using two scales reflecting different components of the construct. Voice was measured by four questions: we asked for subjects' perceptions of their op- portunity to present evidence and express their views and asked about their belief that adequate information exchange preceded the settlement and that the third party gave proper consideration to their views (a = .68). Accuracy, a component of procedural justice [Leventhal et aL, 1980), was measured by three questions addressing beliefs that the negotiation process was an accu- rate way to reach a solution and that the third party had adequate knowledge of each disputants' positions (a = .71).

Finally, three indexes measured subjects' perceptions of the third party. Measures of satisfaction with the third party were five questions assessing subjects' satisfaction, trust, and resentment of the third party, their satisfac- tion with the third party's performance, and their willingness to talk to a manager superior to the third party to change their settlement (a = .87). Third-party fairness was assessed with three questions measuring percep- tions of the third party's fairness, lack of bias, and impartiality (a = .76). Finally, we measured perceptions of the speed of intervention with an index called intervention propriety: three questions asked if the disputants had had a reasonable amount of time to settle the dispute themselves, whether they could have reached an agreement by themselves, and whether the third party's assistance was necessary to reach agreement (a = .66). Table 1 shows means, standard deviations, and the intercorrelations among the measures.

RESULTS

Manipulation Checks

Subjects who received fast settlements reported that the dispute was resolved more quickly than did subjects who received settlements less

840 Academy of Management Journal December

TABLE 1 Means, Standard Deviations, and Correlations

Variables

Distributive justice 1. Outcome satisfaction 2. Outcome fairness

Procedural justice 3. Decision control 4. Voice 5. Accuracy of procedure

Third-party 6. Satisfaction 7. Fairness 8. Propriety of intervention

Means

3.19 2.85

1.53 1.95 3.20

3.14 2.91 2.87

S.d.

2.07 1.77

0.85 0.93 1.23

1.42 1.40 0.65

1

.76

.23

.36

.42

.78

.59

.35

2

.30

.35

.44

.71

.73

.40

Correlations

3

.54

.14

.27

.25

.07"

4

.29

.45

.40

.36

5

.61

.55

.34

6

.74

.41

7

35

' All correlations except this one are significant at p < .05.

quickly (x = 10.37 and 6.78, F^.^^ = 130.68, p < .001). Subjects who received none, one-fourth, one-half, three-fourths, or all of the points in dispute correctly identified the number of points that they received (F4 ̂ 77 = 211.17, p < .001) and their opponents received (F4 ̂ 77 = 423.84 p < .001). Using Newman-Keuls comparisons (Winer, 1971) set at the .05 level, we found all five levels of outcome for both questions to be significantly differ- ent from each other. Finally, negotiators correctly identified themselves as the Boston bargaining representative, and constituents identified themselves as the Boston sales manager (x = 3.96 and 2.96, respectively, F^ ̂ 77 = 648.31, p < .001).

It was possible that manipulating the speed of intervention would con- found subjects' perceptions of how competitive or cooperative their oppo- nent was, with opponents who moved from the first proposal to the last in two rounds rather than eight perceived as more cooperative. To examine this possibility, we used two questions measuring subjects' perceptions of their own and their opponent's competitiveness. There were no differences be- tween subjects in the fast and slow intervention conditions in perceptions of their opponent's (x = 2.26 and 2.07, F^ 177 = 1.00, n.s.) or their own com- petitiveness (x = 2.86 and 3.19, Fj,177 = 1.99, n.s.), strong evidence that all subjects perceived equivalent amounts of competitiveness in the disputes.

Major Analyses

Multivariate analysis of variance (MANOVA) was used to examine the overall influences of the three factors. Significant main effects emerged for speed of intervention (F^ ̂ 77 = 8.36, p < .001), role (F^ ̂ 77 = 23.43, p < .001), and outcome (F4 ̂ 77 = 18.57, p < .001) and for the role-by-speed (Fi,i77 = 4.37, p < .001) and outcome-by-role (F4177 = 2.15, p < .001) interactions. We report the results of univariate follow-up tests next.

Speed of intervention. Hypothesis 1, which predicted that fast third-

1990 Conion and Fasolo 841

party intervention would result in lower ratings of distributive justice, pro- cedural justice, and satisfaction with the third party than slow intervention, received strong support, especially for the procedural justice and third-party measures (see Table 2). Subjects for whom the imposition of an outcome was slow believed that their settlement was slightly more fair than subjects re- ceiving fast intervention, but this effect only approached significance (p < .11). Regarding procedural justice, subjects who received fast impositions reported having less decision control and less voice than those who received slow impositions. Lastly, subjects who received fast impositions were less satisfied with the third party and believed the intervention to be less proper than subjects who received a slow imposition.

Outcome effects. The manipulation of outcome strongly influenced all results (see Table 2). Hypothesis 2, regarding distributive justice ratings, received strong support: ratings of outcome satisfaction increase as the level of outcome increases, but ratings of outcome fairness peak when subjects receive three-fourths of the points and then decline.

The outcome imposed also strongly influenced measures of procedural justice: subjects who received three-fourths or all of the points in dispute reported having greater decision control and voice than all other subjects. In addition, subjects who received at least half the points in dispute perceived the procedure as more accurate than those receiving one-fourth, and the latter saw the procedure as more accurate than did subjects who received no points.

Finally, outcome influenced perceptions of the third party: as outcomes improved, satisfaction with the third party increased. Again, as with the outcome fairness and procedural accuracy measures, all subjects who re- ceived at least half the points in dispute believed the third party was equally fair, and these three groups reported greater third-party fairness than sub- jects in the remaining two conditions. Subjects who received at least half the points in dispute believed that the third party's intervention was more ap- propriate than subjects who received one-fourth, and the latter believed the intervention was more appropriate than did subjects who received no points.

Negotiators versus constituents. Role in the negotiation influenced a wide variety of measures (see Table 2). Hypothesis 3 received strong sup- port, with negotiators seeing themselves as having greater decision control and more voice than constituents. Hypothesis 4 also received strong support, with outcome satisfaction higher for constituents than negotiators. Constit- uents believed that the procedure was more accurate than did negotiators, and the former were also more satisfied with the third party and believed the third party was more fair than did the negotiators.

The role manipulation was also involved in several significant interac- tions. As predicted in Hypothesis 5, the interaction of the role and speed manipulations influenced three variables. Constituents' perceptions of de- cision control (F4 177 = 6.63, p < .011) and voice (F4 177 = 22.32, p < .001) were not strongly influenced by the speed of intervention: means were 1.15

842 Academy of Management Journal December

3 cd

SM

•V Cl

m p

os

1s 3

u

S H

e u

g u

ir te

rs =

• ^

ar te

r

3

2

41 **

* 2

2 3

OO

i n

u

CO

XI CO

oCO

CS rH

a C S

40 **

*

i n

CO CO

CO

u

CO

XI CO

in CO

in in rH

CO CS rH

96 **

*

tn

CD r-<

"6 6

' - '

CJ?

co" CM

rH

ta C S C S

00 **

*

OT

t v -

CM

CS (N

a CM CO

CO CO

" ^

ta CM

in

CO CM

12

CO 00 CO

CO CO

L

cn CM

CO

o CO

in" (N

rs CD

o in

r^ T-< CO

r H r-t T H CM

CD 00 i n CD i n CD CO CO OO CS

CO CS tN CO CD o rs cq CS OT CO CM r H CM CM

O o in rs in rH hs OT CS O

tv

CM

CO CO

o

CO CM

CO

CM

CM

CM O CO

CO

C31

0 0 CD

r H

t n CO

50 .

O l CM

CM

O l i n

o

CO CM

CO

CO r H

10 .

oo CO

CO

CJl oo

d

o o CO

CM 0 0

51 .

0 0

o CO

CO CM

ro le

-a

g a o

in ti

:>

1 "r

ff er

ei

X

-C

3 • a

a ° CO

I 5 § "o e e g s s s :2 s 5 3 3 05 3

a g O cd

•̂ 2

S.^

o o a > < CO •H JH

a o

•r- m •H Si

a o B »

§ g Q) . 2

t < " tv W '-' 3

- CU

ii jd t,

.!•§ C 3

S "3 01 IU

oi 3 IU° •:Sa) • "

:S «

r s <D rs iJ3 1 •3

0 i ^ cd

O DO

a «

cd O. B h

"o 2

g- S S q § .S I V V V ti '^ D. D. Q.

Conlon and Fasolo 843

and 1.20, respectively, for the decision control index and 1.51 and 1.72 for the voice index. However, negotiators' perceptions of decision control and voice were much lower under fast intervention than under slow intervention (x = 1.56 and 2.17, decision control, and 1.72 and 2.83, voice). The inter- action also influenced perceptions of the propriety of the third party's in- tervention {F4177 = 10.96, p < .001). Of the four types of subjects yielded by the interaction, negotiators under slow intervention perceived the third par- ty's behavior to be most appropriate (x = 3.24), and negotiators under fast intervention perceived it as least appropriate [x = 1.99). Constituents' per- ceptions of impropriety in the slow and fast interventions were between these two extremes (x = 2.92 and 2.47).

The outcome-by-role interaction indicated that negotiators generally see themselves as having more decision control than constituents, except when they receive only one-fourth of the points in dispute (F4 J77 = 4.79, p < .001). In addition, although the perceived level of voice increases for both negoti- ators and constituents as the rewards yielded by an outcome increase, the level for constituents declines at the upper end of the outcome range but the level for negotiators continues to increase (F4 177 = 3.94, p < .004). We did not predict these outcome-by-role interactions.

DISCUSSION

The hypotheses guiding this research received strong support. When a third party intervened quickly in a dispute, disputants reported low per- ceived decision control and voice. Compared to disputants whose third party intervened later, they were also much less satisfied with the third party and perceived the intervention as less appropriate. The speed of third-party intervention did not strongly influence perceptions of distributive justice.

A common assumption in research on dispute resolution is that it is beneficial to end conflicts as quickly as possible because protracted disputes raise the costs incurred by both sides. Prolonged disputes can also lead to the escalation of a conflict. Pruitt and Rubin described the course that conflicts can take over time: "First, relatively light, friendly, and inoffensive conten- tious tactics tend to give way to heavier moves . . . the number of issues in conflict tends to increase . . . a focus on specifics tends to give way to more global, all-encompassing concerns . . . motivation in escalating conflict shifts from an initial interest in doing well for oneself to beating the other side" (1986: 7).

Although we are not suggesting that managers ignore disputes until this "conflict spiral" is out of control, the present results suggest that it may be valuable to let subordinates in conflict wrangle with each other for a while before intervening. Having had adequate uninterrupted time to express their opinions, disputants will be more amenable to third-party suggestions and, if they are necessary, third-party settlements. Thus, despite pressures that may force managers to act quickly and decisively in organizational disputes, they must be careful that their haste to resolve conflicts does not result in

844 Academy of Management Journal December

disputants feeling that their rights to due process have been violated. Pre- vious research in political contexts has documented that when disputants see procedural fairness as low, they typically report low satisfaction and commitment to their institutions and representatives (Lind & Tyler, 1988]. Those findings suggest that procedures enacted too quickly may lower or- ganizational commitment among employees.

In addition, the present data clearly indicate that swift third-party in- tervention affects negotiators' procedural fairness judgments more nega- tively than it affects those of constituents. Negotiators, with their greater potential for using voice, were more affected by curtailment of their oppor- tunity to express opinions than were constituents, who did not have this opportunity for voice no matter how long or short the negotiation was.

These findings suggest that negotiators are more likely to suffer low satisfaction and commitment than constituents. However, agents often serve as interpreters for principals: upon dispute settlement, negotiators commu- nicate their feelings and perspectives to their constituents, who are likely to be starved for information regarding how the procedure was enacted. Clearly, the explanations a third party gives to justify intervention can have a powerful role in determining disputants' perceptions of satisfaction and fairness, and such explanations were not made in our study. In fact, the type of explanation a third party provides can attenuate or exacerbate feelings of injustice.

Research by Bies and Shapiro (1987] and Baron (1985] has demonstrated the powerful role that explanations can play in dispute resolution. It is likely to be particularly important for third parties to impart their causal accounts to negotiators before the latter have had a chance to devise their own inter- pretation of events, which they will subsequently pass on to their constitu- ents. This study suggests that, in the absence of any negative information regarding the procedure, third party, or execution of the procedure in a dispute, constituents primarily focus on the favorability of the outcome in determining their satisfaction and fairness, and negotiators focus on both how the procedure was enacted and the outcome received.

The explanations negotiators give constituents is also a topic worthy of investigation. As discussed in the agency theory literature (e.g., Eisenhardt, 1988], negotiators may prefer constituents to evaluate them on the basis of their behavior in negotiations, but constituents may prefer to evaluate ne- gotiators on the basis of the outcomes they attained. Such differences in performance evaluation criteria and expectations may not only affect con- stituents' perceptions of procedural and distributive justice but also their perceptions of their negotiators, leading to a possible escalation of within- group conflict.

The outcome effects on distributive fairness judgments in this study suggest that in certain contexts, the objective favorability of an outcome and its perceived fairness may have a curvilinear relationship. When disputants believe that resource distribution should be based on an equality norm, it seems likely that they will see both outcomes that are extremely favorable

1990 Conlon and Fasolo 845

and those that are extremely unfavorable to one disputant as unfair. We know of no research that has been able to demonstrate such a relationship, but it seems plausible. The present results certainly show a ceiling effect for measures of outcome fairness. Perhaps a negotiation task in which dispu- tants had a close relationship—as workers in the same office or as good friends, for instance— would produce results supporting this idea.

The laboratory context of this study and the use of student subjects suggest that caution be used when making generalizations to managers in organizational settings. Nevertheless, we based the three-issue dispute on an actual dispute that occurred in a securities company, and several growing trends, including networked computers, electronic mail, computer confer- encing, and "worksteading" (employees working at home], offer natural set- tings for organizational conflict resolution that are not too different from the context created in this study. Certainly, more research on managerial third- party roles and the reactions of constituents and negotiators is warranted. Future research should continue to examine variation within, and not just between, dispute resolution procedures. The present study suggests that further research on how managers acting as third parties enact conflict res- olution procedures is necessary, and such research will continue to enrich understanding of procedural and distributive justice.

REFERENCES

Adams, J. S. 1965. Inequity in social exchange. In L. Berkowitz (Ed.), Advances in experimen- tal social psychology, vol. 2: 267-299. New York: Academic Press.

Baron, R. A. 1985. Reducing organizational conflict: The role of attributions. Journal of Applied Psychology, 69: 272-279.

Bies, R. J., & Shapiro, D. L. 1987. Interactional fairness judgments: The influence of causal accounts. Social Justice Research, 1: 199-218.

Carnevale, P. J., & Conlon, D. E. 1988. Time pressure and strategic choice in mediation. Orga- nizational Behavior and Human Decision Processes, 42: 111-133.

Conlon, D. E., Lind, E. A., & Lissak, R. I. 1989. Nonlinear and nonmonotonic effects of outcome on procedural and distributive fairness judgments. Journal of Applied Social Psychology, 19: 1085-1099.

Eisenhardt, K. M. 1988. Agency- and institutional-theory explanations: The case of retail sales compensation. Academy o/Management/ourna/, 31: 488-511,

Folger, R., & Konovsky, M. A. 1989. Effects of procedural and distributive justice on reactions to pay raise decisions. Academy of Management Journal, 32: 115-130.

Creenberg, J. 1987. A taxonomy of organizational justice theories. Academy of Management Review, 12: 9-22.

Greenberg, J. 1990. Looking fair vs. being fair: Managing impressions of organizational justice. In B. M. Stavir & L. L. Cummings (Eds.), Research in organizational behavior, vol. 12: 111-157. Creenvirich, Conn.: JAI Press.

Karambayya, R., & Brett, J. M. 1989. Managers handling disputes: Third-party roles and per- ceptions of fairness. Academy of Management Journal, 32: 687-704.

Latour, S. 1978. Determinants of participant and observer satisfaction with adversary and in- quisitorial modes of adjudication. Journal of Personality and Social Psychology, 36: 1531-1545.

846 Academy of Management Journal December

Leventhal, C. S., Karuza, J., & Fry, W. R. 1980. Beyond fairness: A theory of allocation prefer- ences. In G. Mikula (Ed.), Justice and social interaction: 167-218. New York: Springer- Verlag.

Lind, E. A., & Tyler, T. R. 1988. The social psychology of procedural fustice. New York: Plenum.

Lissak, R. I. 1983. Procedural fairness: How employees evaluate procedures. Unpublished doctoral dissertation. University of Illinois.

McGrath, J. E., & Rotchford, N. 1983. Time and behavior in organizations. In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior, vol. 5: 57-101. Greenwich, Conn.: JAI Press.

Musante, L., Gilbert, M. A., & Thibaut, J. 1983. The effects of control on perceived fairness of procedures and outcomes. Journal of Experimental Social Psychology, 19: 223-238.

Pruitt, D. G. 1981. Negotiation behavior. New York: Academic Press.

Pruitt, D. G., & Rubin, J. Z. 1986. Social conflict: Escalation, statement, and resolution. New York: Random House.

Ross, W. H., Gonlon, D. E., & Lind, E. A. 1990. The mediator as leader: Effects of behavioral style and deadline certainty on negotiator behavior. Group and Organization Studies, 15:105- 124.

Sheppard, B. H. 1984. Third party conflict intervention: A procedural framework. In B. M. Staw & L. L. Gummings (Eds.), Research in organizational behavior, vol. 6: 141-190. Green- wich, Gonn.: JAI Press.

Sheppard, B. H., & Lewicki, R. J. 1987. Toward general principles of managerial fairness. Social Justice Research, 1: 161-176.

Thibaut, J. W., & Walker, L. 1975. Procedural justice: A psychological analysis. New York: Erlbaum/Halstead.

Thibaut, J. W:, & Walker, L. 1978. A theory of procedure. California Law Review, 66: 541-566.

Ury, W. L., Brett, J. M., & Goldberg, S. B. 1988. Getting disputes resolved. San Francisco: Jossey-Bass.

Walster, E., Walster, G. W., & Berscheid, E. 1978. Equity: Theory and research. Boston: Allyn & Bacon.

Winer, B. J. 1971. Statistical principles in experimental design (2d. ed.). New York: McGraw Hill.

Donald E. Conlon received his Ph.D. degree in business administration from the Uni- versity of Illinois. He is currently an assistant professor in the Department of Business Administration, University of Delaware. His research interests include disputant per- ceptions of procedural and distributive justice, third-party power, and decision making.

Peter M. Fasolo received his Ph.D. degree in psychology from the University of Dela- ware. He is an organizational consultant at Human Resource Associates, Englewood Cliffs, New Jersey. His research interests include the effects of procedural and distrib- utive justice on performance, organizational commitment, and interview judgments.