Repeat Player (N=35)          Non-Repeat Player (N=235)
Commercial or Employment Rules

Purchase Article Back | Chicago-Kent Home | Journal Home

Employee Rights and Employment Policy Journal

Volume 1 1999 Number 1




Employment arbitration is emerging as a controversial method for resolving disputes between employers and employees not represented by a union.(1) While an earlier study found only 4 of 111 employers used outside arbitration in 1991,(2) by 1995, the GAO found that 10 percent of all employers with 100 or more employees use binding arbitration for employment disputes, and as many as half of these may impose mandatory arbitration as a condition of employment.(3)
The use of employment arbitration continues to grow. 

Proponents argue that arbitration will provide an expeditious, low cost means for employees to get a hearing on statutory, contractual, or other claims arising out of their dismissal; they argue this is superior to no hearing at all, or in some cases, to protracted litigation.(4) Employment arbitration holds great promise as a dispute resolution process for employees who otherwise would never have a hearing on their claims against an employer. The procedural justice literature strongly suggests that the mere fact of a hearing before an apparently impartial decision-maker (rather than a settlement between lawyers with no arbitration hearing) will enhance employees' satisfaction with the outcome, whether they win or lose.(5) In other words, some hearing is better than no hearing at all. That employees will experience greater procedural justice through an arbitration hearing can only be to the good. Moreover, employers likely will benefit from the elimination of the outlier jury award, concerns over which have motivated substantial changes in personnel practices, at some significant cost.(6)

However, critics argue that mandatory arbitration is a new corporate tool used to advantage by large companies against consumers, employees, and other "little guys."(7) Commentators in the press have reported claims that employment arbitration will disadvantage women, because the panels are predominantly comprised of male arbitrators.(8) In particular, critics caution that in the absence of a union, the process is skewed against the employee, particularly the unrepresented employee.(9) Some have expressed concern that, as institutional repeat players who will use the arbitration process for multiple cases, employers have an advantage over employees, who are unlikely to use it more than once or twice in their lifetimes.(10) The industrial and labor relations community has expressed similar concerns about employers as repeat institutional players.(11)

The purpose of this article is to determine whether there is a measurable repeat player effect, to examine possible explanations for its existence, and to identify policy issues for employment arbitration that are raised by the effect. First, this article examines the role of the employer and the theoretical basis for a repeat player effect in employment arbitration. Second, it examines some of the empirical literature on employment arbitration. Third, it presents results of an empirical study of employment arbitration awards issued under the American Arbitration Association's Commercial and Employment Arbitration rules during 1993 and 1994,before recent rules changes; this study documents that the repeat player effect exists. Finally, it discusses some of the implications of the repeat player effect for employment arbitration policy. The focus here is solely on relevant empirical evidence and arbitration policy; judicial review of arbitration awards is outside the scope of this article.(12)

I. The Issue of the Repeat Player 

In Cole v. Burns International Security Services,(13) a federal court of appeals discussed for the first time at some length the issue posed by repeat player employers in non-union employment arbitration. Clinton Cole worked as a security guard for Burns Security. When Burns hired him, it required him to sign a "Pre-Dispute Resolution Agreement" as a condition of employment. This agreement waived Cole's right to a jury trial, and required him, at the employer's option, to submit any dispute relating to his recruitment, employment, or termination to binding arbitration under American Arbitration Association rules. The agreement covered statutory claims and common law claims, but excluded worker's compensation claims. The waiver of a jury trial was absolute, and took effect even if the employer chose to litigate instead of arbitrate. About two years later, Burns fired Cole, who in turn filed charges with the Equal Employment Opportunity Commission alleging race discrimination, racial harassment, retaliation for his writing a letter of support for another employee who had alleged sexual harassment, and intentional infliction of emotional distress. Cole filed a complaint in federal district court which was dismissed when the court granted Burns' subsequent motion to compel arbitration. Cole appealed. 

Judge Harry Edwards, noted labor law and dispute resolution scholar,(14) authored the majority opinion. The court held that the arbitration agreement was valid and enforceable under the Federal Arbitration Act ("FAA").(15) There has been a lively scholarly debate regarding the reach of the FAA,(16) but that question is outside the scope of this article. The court then turned to the enforceability of conditions of employment requiring employees to arbitrate statutory claims. The court distinguished between traditional labor arbitration in the context of collective bargaining, and mandatory arbitration of statutory claims in the context of non-union employment, noting that commercial arbitration substitutes for litigation, while labor arbitration substitutes for industrial strife. Moreover, labor arbitration molds a system of private law to govern a private contractual relationship which is part of the continuous process of collective bargaining. As a result, labor arbitration awards are subject to limited judicial review, and generally not subject to reversal for an error of law.(17)
The court then turned to the repeat player issue, observing that there are unique protections for both parties in collective bargaining: 

For example, because both unions and employers are repeat customers of arbitration and have a hand in selecting the arbitrator to hear their disputes, arbitrators who regularly favor one side or the other will not be hired again. As a result, arbitrators have a strong personal interest in crafting awards that will be respected as fair by both parties regardless of who wins or loses the particular dispute.(18)
The court noted that these structural protections inherent in collective bargaining are not present in commercial arbitration of statutory claims. 
Unlike the labor case, in which both union and employer are regular participants in the arbitration process, only the employer is a repeat player in cases involving individual statutory claims. As a result, the employer gains some advantage in having superior knowledge with respect to selection of an arbitrator.(19)
The court observed that a lack of public disclosure of arbitration awards also may systematically favor companies over individuals. The court noted problems with building a case for a pattern or practice of discrimination if all the complaints are arbitrated because the awards are confidential and not published, the problem that employers may structure arbitration to their advantage unilaterally, and that employers may shift the costs for arbitration to employees to create a financial barrier or disincentive to litigate or arbitrate. 

The court responded largely by citing Gilmer as having resolved the issue of whether agreements to arbitrate were enforceable notwithstanding these concerns.(20) The court reviewed the grass roots movement among major providers and professional associations of advocates and neutrals to bring minimal protection into employment arbitration through a voluntarily adopted Due Process Protocol for the Mediation and Arbitration of Statutory Employment Claims, and protection present in the American Arbitration Association National Rules for the Resolution of Employment Disputes.(21)

It then held that the employer could not effectively deprive the employee of a forum for resolving statutory employment claims by imposing the substantial costs of arbitration on the employee. Employers must bear the full freight of the arbitrator's fee. Notwithstanding its earlier discussion regarding the repeat player problem, the court observed:

Some commentators have suggested that it would be a perversion of the arbitration process to have the arbitrator paid by only one party to the dispute. We fail to appreciate the basis for this concern. If an arbitrator is likely to "lean" in favor of the employer - something we have no reason to suspect - it would be because the employer is a source of future arbitration business, and not because the employer alone pays the arbitrator. It is doubtful that arbitrators care about who pays them, so long as they are paid for their services.(22)
It noted the research on the repeat player effect reported in this article, but observed that it was difficult to know what to make of the results without more information about the merits of the cases. In addition, it asserted there were protections against arbitrators systematically favoring employers as the source of future business:
[T]here are several protections against the possibility of arbitrators systematically favoring employers because employers are the source of future business. For one thing, it is unlikely that such corruption would escape the scrutiny of plaintiffs' lawyers or appointing agencies like the AAA. Corrupt arbitrators will not survive long in the business. In addition, wise employers and their representatives should see no benefit in currying the favor of corrupt arbitrators, because this will simply invite increased judicial review of arbitral judgments. Finally, if the arbitrators who are assigned to hear and decide statutory claims adhere to the professional and ethical standards set by arbitrators in the context of collective bargaining, there is little reason for concern.(23)
The court assumed that an arbitral record systematically favoring employers would represent corruption, and would be promptly identified by the plaintiffs' bar and appointing agencies. Moreover, employers would not seek these arbitrators because it would invite increased judicial review of arbitrator awards. Lastly, the arbitrator code of ethics developed in the context of labor arbitration should prevent this result.

There are certain unanswered questions in the court's analysis of the repeat player issue. No one is keeping records in employment arbitration about the awards of particular arbitrators. This is not unusual; in labor arbitration the parties maintain their own institutional memories, and occasionally commercial services will sell reports regarding a given labor arbitrator's published awards. However, even in labor arbitration, less than ten percent of these awards are published. Thus, at present, it is impossible to determine whether a given arbitrator is systematically favoring the employer in question, or employers in general. Second, is it fair to assume that if an arbitrator's record shows that he or she ruled for the employer a majority of the time, or even seventy, eighty, or ninety percent of the time, that this necessarily represents corruption? Perhaps the merits of individual cases demanded this result. On the one hand, the court rejects evidence of the repeat player effect without more information about the merits of the individual cases surveyed;(24) on the other hand, the court reasons that labor arbitration is fair because the parties will strike an arbitrator who regularly favors one party.(25) The issue warrants further analysis.

Professor Marc Galanter gives us a framework for that analysis in his classic article on the limits of legal change.(26) In an extensive survey of literature about litigation and dispute settlement from a broad range of disciplines, he suggests that built in advantages in the legal system can best be identified and understood through an analysis of elements including the rules, the institutional facilities, lawyers, and parties. His typology of parties divides actors into the categories of repeat players (RPs) and one-shotters (OS).(27) Repeat players have frequent recourse to the courts or institutional facilities, while one-shotters will have few such occasions. He identifies a collection of advantages repeat players will generally enjoy over one-shotters, including: (1) experience leading to changes in how the repeat player structures the next similar transaction; (2) expertise, economies of scale, and access to specialist advocates; (3) informal continuing relationships with institutional incumbents; (4) bargaining reputation and credibility; (5) long-term strategies facilitating risk-taking in appropriate cases; (6) influencing rules through lobbying and other use of resources; (7) playing for precedent and favorable future rules; (8) distinguishing between symbolic and actual defeats; (9) investing resources in getting rules favorable to them implemented.(28) One-shotters, on the other hand: (1) have more at stake in a given case; (2) are more risk averse; (3) are more interested in immediate over long-term gain; (4) are less interested in precedent and favorable rules; (5) are not able to form continuing relationships with courts or institutional representatives; (6) cannot use the experience to structure future similar transactions; and (7) have limited access to specialist advocates. Galanter observes that most litigation is brought by repeat player plaintiffs against one-shotter defendants (landlord-tenant, creditor-debtor, prosecutor-accused criminal, and IRS-taxpayer being salient examples).(29)
The one-shotter plaintiff against the repeat player defendant is comparatively infrequent (and includes personal injury, for example).(30)

At the time of his article, Galanter observed that, outside the personal injury area, litigation in this category was not routine, but 

It usually represents the attempt of some OS to invoke outside help to create leverage on an organization with which he has been having dealings but is not at the point of divorce (for example, the discharged employee or the canceled franchisee). The OS claimant generally has little interest in the state of the law; the RP defendant, however, is greatly interested.(31)
During the over twenty years since publication of his article, litigation brought by employee plaintiffs against employer defendants has become, if not routine, certainly more publicized.(32) Certain scholars assert that the development of the wrongful discharge tort(33) has had significant economic consequences for employers.(34) It is not surprising then, given Galanter's analysis, that employers as repeat players might use this experience to structure their relationships and transactions with employees to better advantage.(35) Employment arbitration has been advocated as a superior means for resolving these disputes by reference to its success in the collective bargaining context. How does Galanter's typology of parties work in employment arbitration?

A. Absent Unions

Employment arbitration follows the model set by labor arbitration, where unions represent employees and are bound by a legal duty of fair representation.(36) However, the model of labor arbitration cannot be lifted wholesale into this new context, because the premise of an elected bargaining representative negotiating the underlying contract or charter for the arbitrator is missing. In labor arbitration, the parties generally investigate the background of an arbitrator, and use that information in arbitrator selection.(37) The labor union is an institutional repeat player. It selects arbitrators on a repeated basis, and maintains an institutional memory of past arbitral rulings. It generally pays half of the arbitrator's fee. There is empirical evidence that parties will strike a predictably biased arbitrator.(38) Arbitrators know this. Thus, there is empirical evidence that labor arbitrators respond to the financial incentive of repeat business by trying to be fair in one of two ways. One theory is that they either strive for objectively predictable awards, that is, the award most other arbitrators would make on the same facts;(39) the other is that they 'split the baby' by balancing the equities, this latter approach more commonly found in interest arbitration.(40)

Unions could, but generally do not, provide representation services in employment arbitration. Thus, the traditional repeat institutional player is missing. With it, some argue the traditional financial incentive may be absent for arbitrators to follow either of these models of fairness. Arbitrators take pride in the quality of their decision-making, and may not respond to the presence or absence of a financial incentive. Nevertheless, the structure of employment arbitration raises the question, and it warrants empirical investigation.

B. Substitute Repeat Players: The Plaintiff's Bar 

Galanter observes that lawyers may themselves serve as repeat players in the legal system.(41)  Lawyers representing clients in arbitration develop expertise regarding arbitrators. It is certainly true that management labor lawyers develop institutional memory similar to that of unions and their lawyers regarding arbitral performance. So too, the plaintiffs' bar, those lawyers who generally represent employee plaintiffs against employer defendants in claims arising out of Title VII of the Civil Rights Act, wrongful discharge torts, or contracts, could in theory develop such an institutional memory, as the court suggests in Cole.(42)  To the extent that they have a large employee caseload, they could serve as repeat players counterbalancing the repeat player employer. While there may be isolated exceptions, this does not appear to be happening, at least, not yet. AAA case files reveal that a substantial proportion of employees are unrepresented in employment arbitration. In addition, those employees who are represented use a wide variety of lawyers and law firms. The repeat player effect, if any, from lawyer representation is not yet measurable. Although researchers have found that parties fare better in labor arbitration when they have a lawyer and the other side does not,(43) there is no similar measurable effect yet in employment arbitration. The analyses using representation are not statistically significant.(44)

Moreover, there is reason to believe that most individual members of the plaintiff's bar may never successfully emerge as repeat players in employment arbitration. Galanter observes that the lawyers for one-shotters share certain characteristics. They tend to make up the lower status ranks of the legal profession, practice alone, have attended local or part-time law schools, have problems mobilizing a clientele, have fewer resources to devote to individual cases, and engage in more routine case handling.(45) To understand why this militates against successful repeat player plaintiff's counsel in employment arbitration, one must consider law firm economics. Lawyers generally receive compensation based either on an hourly rate, or on a contingent fee basis. Employee plaintiffs may be able to afford hourly rates if they are highly compensated white collar employees, or if they have independent resources. The fact that employment arbitration is faster and cheaper than litigation may mean that marginally more of these highly compensated white collar employees may be able to afford lawyers on an hourly basis. However, the vast majority of employees to whom employment arbitration will eventually apply will be modestly compensated blue and pink collar workers. These generally cannot afford to pay lawyer hourly rates. 

Thus, in order for them to obtain legal representation, they must have a case that a lawyer wants to take on a contingent fee basis. In a contingent fee arrangement, the lawyer is compensated by receiving a quarter to a third of the plaintiff's recovery. The lawyer must front a substantial investment in time and expenses in the case. Lawyers will only take a contingent fee case if they think they are going to win.(46)

In order for a lawyer to conclude he or she will win, the case must fall into certain predictable categories based upon established case precedent. Moreover, based on precedent and anecdotal evidence regarding local jury awards, the lawyer must place a certain value on the case. That value must be enough to at least pay for the time the lawyer puts into the case at his or her regular hourly rates. The hourly rates are set based upon firm overhead and targets for annual partner compensation. Often, a contingent fee case may provide a windfall or premium over hourly rates, which provides an incentive to overcome the uncertainty of recovery. Lawyers engage in this cost-benefit analysis every time they take a contingent fee case.

Employment arbitration has altered the equation for employees seeking legal counsel. Good lawyers may be less likely to take an employment arbitration case than a litigation case. In employment arbitration, the plaintiffs' bar cannot get to a jury. Empirical evidence shows that cash outcomes are systematically more modest in employment arbitration than in jury awards for wrongful dismissal cases,(47) although there is a risk that on the merits, employment arbitration and litigation cases may be apples and oranges.(48) All of this means that the cash value on a contingency fee is probably less for an employment arbitration case than it is for a traditional piece of employment litigation. Of course, the expenses of taking a case to arbitration are also lower than the expenses of traditional litigation; this is one of the great strengths of arbitration. However, it is possible that good lawyers, who correctly estimate the case's true worth, often may not take employment arbitration cases on a contingent fee basis. They may be less likely to take an employment arbitration case than a litigation case if the two are equivalent on the merits. A survey of employers who have adopted arbitration systems revealed that many believed that juries posed too great a risk of run-away monetary awards.(49) An early empirical study confirms this perception.(50)

Lastly, even less able lawyers may be unlikely to take employment arbitration cases. Anecdotal evidence suggests that certain plaintiffs' lawyers will take cases on a contingent fee basis, cases which they know are losers on the merits, file lawsuits in state or federal court, and then settle the cases for the nuisance value of the lawsuit. Sometimes, even the nuisance value of a case in federal court can reach $10,000. However, anecdotal evidence suggests that employers generally do not settle employment arbitration cases as readily. The process is quicker and cheaper, so it is easier for an employer to take the position they will arbitrate every case, to avoid a reputation for settling which might encourage more claims. A recent survey confirms that employers settle fewer arbitration cases than litigation cases.(51)  However, these members of the plaintiffs' bar do not want to try the case. If they think an early settlement is unlikely, they will not take the case. Thus, it is less likely that even less able lawyers will emerge as repeat players in employment arbitration cases. 

It is possible on the above analysis that only the most successful individual plaintiffs' lawyers will emerge as repeat players in employment arbitration, and they will do so generally representing highly compensated white collar employees who are willing to pay hourly rates, not blue and pink collar workers paying through contingent fees. However, some suggest that since arbitration costs less, attorneys eager for litigation experience may be more accessible for arbitration than for a courtroom trial. Moreover, as the legal profession becomes increasingly economically competitive, access to legal representation may improve with declining attorneys' fees.

Galanter observes:

The existence of a specialized bar on the OS side should overcome the gap in expertise, allow some economies of scale, provide for bargaining commitment and personal familiarity. But this is short of overcoming the fundamental strategic advantages of RPs--their capacity to structure the transaction, play the odds, and influence rule-development and enforcement policy.(52)
In employment arbitration, the ultimate question is an empirical one. It remains to be seen whether effective employee advocate repeat players will emerge over time. They have not emerged yet.

C. Employees as a Group as Repeat Players

It is conceivable that employees as a group might gain and pass on information about arbitrators from each other. To the extent that the process is used for grievances in an ongoing employment relationship, that is, grievances about issues other than dismissal from employment, it is possible that word of mouth at the workplace might serve as a mechanism for passing on information about arbitrators. Moreover, the outcomes of such grievances would likely affect others, and become widely known. However, this is less likely to be the case with employee dismissal. There are anecdotal accounts from dismissed employees to the effect that they become pariahs at the workplace; others do not wish to be associated with the one who is dismissed. Moreover, while a few loyal friends might follow the details of a case, the likelihood that it would become widely known would appear to be small. Most employees who have been fired are concerned about their reputation for purposes of future employment, and would not want to breach confidentiality. Most employment arbitration cases studied to date concern dismissal from employment. Moreover, a recent survey indicates that a significant minority of employers are limiting the arbitration remedy to dismissal cases.(53) However, the question certainly warrants research. One method might be a survey of employee complainants in arbitration asking what information they had about the arbitrator, and where that information came from. To the best of this author's knowledge, there is no currently available empirical evidence that employees as a group pass on information about arbitrators.

D. Repeat Player Employers

Using Galanter's typology, all employers are potentially repeat players. However, Galanter recognizes that there is a continuum between one-shotters and repeat players; he uses the dichotomous category for clarity in his analysis.(54) For purposes of measurement, it is logical to distinguish between those employers who are demonstrably repeat players in the employment arbitration system and those who use it less frequently. Generally, in employment arbitration the only repeat players observed to date are employers, with the exception of a single employee who was suspended, and in a later case, discharged. The empirical question is whether these repeat players have systematically superior outcomes in employment arbitration.

In labor arbitration, where both employer and union are repeat players, traditional outcome measures reveal a rough parity in outcomes. In a recent study of 3,949 public and private sector published labor arbitration cases, grievants won their cases in whole or in part 52% of the time, and in a subsample of 1,427 discipline cases, grievants won their cases in whole or in part approximately 57% of the time.(55) It should be noted that only about 10% of all labor arbitration awards are published, and the published sample is not random but instead the product of careful editorial selection. However, others have observed that labor arbitration cases generally split 50:50, with unions winning grievances in whole or in part about half the time, and these results are based on records including unpublished awards.(56) Professor Rau has noted that this aspect of labor arbitration, that the outcomes reflect a rough parity between the parties, may appropriately mirror the give and take of collective bargaining.(57)

II. Empirical Research on Employment Arbitration

The above discussion raises the empirical question do repeat players, as Galanter's analysis would predict, fare better in employment arbitration? There is substantial empirical research on outcomes in labor arbitration, including grievance or rights arbitration,(58) and interest arbitration,(59) but relatively little on employment arbitration. One common method is to examine labor arbitration outcomes in relation to a variable; for example, some researchers have found gender effects both in favor of and against female grievants in grievance labor arbitration.(60) There is also substantial research on grievance procedures and dispute resolution in employment.(61)

There is limited empirical research on employment arbitration. One study compared settled, litigated, and arbitrated employment discrimination case outcomes and surveyed counsel regarding case practices.(62)
Another recent study found that most employment arbitration procedures were developed in the last two years and that over 75 percent of the employers surveyed adopted the plans to reduce litigation costs, while only 15 percent of employers did so to improve employee relations or give employees a voice.(63) Most plans were developed by human resources staff and legal counsel without employee input. About 75 percent of the plans made arbitration a condition of employment. Ten percent of the plans prohibited employees from using outside counsel. Almost 15 percent of the plans provided for unilateral employer selection of the arbitrator. All of this provides empirical evidence for Galanter's proposition that repeat players have strategic superiority in that they can structure a transaction to their advantage.

On the other hand, a one-year sample of 1992 American Arbitration Association (AAA) Commercial Arbitration awards revealed no evidence of a systematic pro-employer bias.(64) That study examined arbitration outcomes in relation to the variables of who filed the claim (employer or employee) and whether the arbitrator received compensation (under prior rules some arbitrators served pro bono). Under the Commercial rules, either the employer or the employee could file a claim or demand for arbitration. Typically, employees filed claims arising out of dismissal, while employers sought to recover unearned commissions, cancel stock options, or enforce covenants not to compete. Employees had superior outcomes on their own claims compared to employers on theirs, both in terms of win rate and percentage of demand recovered as damages, whether or not the arbitrator was compensated. However, employees in Commercial cases tended to be highly compensated managerial or executive employees, represented by counsel. 

Effective January 1, 1993, the AAA implemented its Employment Dispute Resolution Rules (Employment rules); these were a new set of rules for the arbitration of nonunion employment disputes.(65) Nonunion employment disputes had been covered for many years primarily as a special area under the AAA's Commercial Rules,(66)
but the new rules were designed to meet the growing need for dispute resolution for employment cases generally. Since many parties incorporated the Commercial Rules into their contracts, the AAA continued to administer arbitrations under both rules. The Employment rules contained differences in sections regarding appointment of the arbitrator and disclosure of information about the arbitrator, but for the most part, the new Employment Rules were substantively the same as the previous Commercial Rules. An empirical study provided evidence that different populations of employees might be using the new AAA Employment Rules, and in particular, that there were repeat player employers using arbitration pursuant to the terms of unilaterally imposed personnel handbooks or policies.(67) These cases were more likely than commercial cases to involve unrepresented employees who lost a claim involving their dismissal from employment. The sample of employment cases available was too small (n=28) to do more than identify emerging concerns about due process.

III. Empirical Evidence of the Repeat Player Effect

The instant study examines a larger, two-year sample of employment arbitration awards. The above discussion leads to the following hypotheses:

H(1): Employees will win less frequently and/or will recover a lower proportion of their claims (have lower Outcome) in cases involving repeat player employers.

H(2): White collar workers will win more frequently and recover a greater proportion of their claims (have higher Outcome) than blue or pink collar workers in employment arbitration.

A. Method

This study represents a macrojustice assessment of employment arbitration; in other words, the study examines the overall pattern of outcomes produced by the dispute resolution process in a sample of actual cases, without controlling for the merits of the individual case.(68) It should be noted at the outset that this study examines a sample of cases decided before the existence of the Due Process Protocol for the Mediation and Arbitration of Statutory Employment Disputes, and before the AAA adopted the National Rules for the Resolution of Employment Disputes (effective June 1, 1996). Thus, the arbitration award decision patterns reflect the earliest period of employment arbitration, before new procedural protections were put into place. They are not necessarily representative of current decision patterns under newer AAA rules, and research into decision patterns under the newer rules is ongoing. 

This study examines a 270-case sample consisting of arbitration awards decided in 1993 under the AAA Commercial Arbitration Rules (n=186), and arbitration awards decided in 1993 and 1994 under the AAA Employment Dispute Resolution Rules (n=84). The total population of awards decided in 1993 under the Commercial Rules and in 1993-94 under the Employment Dispute Resolution Rules consisted of 330 cases, from which 55 cases were excluded to yield the total of 270 cases in the sample. The 55 cases were excluded on one of the following grounds: (1) essential information was missing, (2) the award represented a settlement or stipulated award, or (3) the case was not an employment dispute, but instead a partnership or real estate dispute. During the period from 1993 to 1994, there were no significant, substantive differences between the Commercial and Employment Dispute Resolution Rules, so it is appropriate to pool the cases for purposes of analysis. A multiple regression analysis using Outcome as the dependent variable and a categorical variable for Commercial or Employment Rules indicated that the rule set under which the case arose was not a significant factor in predicting Outcome, further justifying pooling the cases. Lastly, a case pool of exclusively 1993-1995 Employment Dispute Resolution Rules cases replicates the results of this study.(69)

For each case, the researcher examined, where available, the demand for arbitration, arbitrator's award, and AAA closing data sheet. Table 1 shows the characteristics of the sample of awards.

Table 1. Distribution of Employees and Claims in Repeat Player and Non-repeat Player Employment Arbitration Cases

Purchase Article
Commercial 9 26% 177 75%
Employment 26 74% 58 25%

Employee Collar Color
White 10 29% 170 72% 
Blue 15 43%  17 7% 
Pink 9 26%  16 7% 
Unknown  1 2%  32 14% 

Employee Gender
Male 30 86%  189 80% 
Female 5 14%  40 17% 
Unknown 0 6 3% 

Arbitrator Gender
Male 29 83% 191 81% 
Female 6 17%  36 15% 
Unknown 0 8 4% 

Independent variables include whether the case is an employer or an employee claim, and whether the case involves a Repeat Player employer, defined as an employer who uses arbitration more than once in the sample (coded 1 for yes and 0 for no). In addition, independent variables include job category of the employee, specifically whether the case involves a White, Blue or Pink collar employee. For the purpose of this study, white collar includes management, financial, real estate, sales, physicians, or similar highly compensated (greater than $40,000 annually in salary) workers; blue collar workers include manufacturing, food service, and transportation; pink collar workers include medical-technical workers, clerical or administrative employees, or other non-manufacturing employees who earn less than $40,000 annually. Information on job category was found on the demand for arbitration (which asks for job title and basis for claim), in the award of the arbitrator (which may discuss the employee's position and the basis for damages), together with the nature of the company and size of the damage award (if the latter is based on annual salary). 

Dependent variables include the dollar amount of damages, if any, awarded by the arbitrator (Damages Amount), and the percentage of claim recovered (a ratio of Damages Amount divided by Demand), or amount of money the claimant indicated was at stake. This ratio was used to create a variable named Outcome. On the descriptive tables (Tables 2 and 4), the Outcome variable may be multiplied by 100 to convert it into the percentage of the original demand that the claimant recovered. In addition, a dichotomous variable entitled Damages was created to indicate a win by the claiming party of damages any dollar amount. In general, there were few counterclaims, and these were not separately coded or considered as independent cases in data analysis. Instead, where a counterclaim was present upon which the arbitrator made an award, the net result of the two combined was reflected as the Damages Amount. 

B. Results

Table 2 shows descriptive statistics on the sample by Repeat Player and Non-repeat Player employer. Although the mean damage award for repeat player cases involving employee claims (EE Claims) is higher than for non-repeat player cases, this is due to a single outlier award. The median is the better measure of central tendency for the sample, and it reflects that employees generally win nothing in repeat player cases. In general, repeat player cases involve smaller stakes, and lower outcomes for employees.

Table 2. Descriptive Statistics on Repeat Player and Non-repeat Player Employment Arbitration Cases, All Cases

                                     Repeat Player (N=35)                 Non-Repeat Player (N=235)

Whose Claim
EE Claims 31 89%  201 86% 
ER Claims 4 11%  34 14% 
                                                Standard                                                                Standard
Demand                       Mean                 Deviation          Median           Mean                  Deviation           Mean
Overall $160,678 $327,472 $25,000 $249,669 $794,369 $50,000
EE Claims $199,526 $360,814 $25,000 $267,105 $840,982 $50,000
ER Claims $15,000 $7,071 $12,500 $153,771 $458,076 $22,400

Damages Amount
Overall $77,066 $261,524 $0 $69,963 $201,731 $11,027
EE Claims $87,178 $278,004 $0 $73,986 $213,749 $13,402
ER Claims $3,750 $7,500 $0 $46,530 $107,046 $8,438

Outcome (Damages Amount Divided By Demand)
Overall .13 .32 0 .47 .73 .28
EE Claims .11 .30 0 .48 .74 .28
EE Claims .25 .50 0 .44 .66 .06

Since the hypotheses are concerned with employee outcomes, analyses were performed on a sub-sample of the total sample, namely those cases where the employee is the claimant in arbitration (232/270 cases). Employers may be the claimant in arbitration when they seek to enforce a covenant not to compete, or to recover an overpaid advance on commissions which were not subsequently earned, for example. Employees were generally claimants in disputes concerning the termination of their employment, terminal compensation including stock options, and other compensation disputes. 

Table 3, Damages by Repeat Player, indicates the lower frequency with which arbitrators award damages in any amount in repeat player cases involving employee claims.

Table 3. Contingency Table Showing Damages (rows) by Repeat and Non-repeat Player Employer (columns), Employee Claims Only

                                            Repeat Player           Non-Repeat Player            Total
Damages 5 2%  142 61%  147 63% 
No Damages 26 11%  59 26%  85 37% 
Totals 31 13%  201 87%  232 100% 

Table 3 shows that employees lose significantly more frequently in cases involving Repeat Player employers, Pearson Chi Square = 34.39, DF 1, P<.001. 

Employees also have significantly lower Outcomes in cases involving Repeat Player employers. A one-way analysis of variance using Repeat Player as the independent variable and Outcome (the proportion of their original Demand that employees are awarded as damages), was significant, F=6.53 (DF 1, 213), P<.01. Table 2 shows that in repeat player cases, employees recover only 11% of what they demand, on average, while in cases involving non-repeat player employers, employees recover an average of 48%; the analysis of variance shows that this difference is statistically significant.

Table 4 shows descriptive statistics for White, Blue, and Pink Collar employees.

Table 4. Descriptive Statistics for White, Blue, and Pink Collar


                                              White                              Blue                             Pink                                 Unknown
All Claims                             N=180                             N=32                          N=25                                 N=33
Mean $313,806 $47,058 $22,684 $97,158
S.D. $893,581 $59,021 $11,899 $195,249
Median $67,000 $25,000 $25,000 $32,500

Damages Amount
Mean $101,843 $2,802 $7,851 $14,183
S.D. $250,620 $7,223 $14,138 $22,123
Median $16,242 $0 $0 $2,000

Mean .52 .10 .25 .40
S.D. .79 .25 .42 .57
Median .33 0 0 .07

Employee Claims                   N=151                             N=31                             N=21                              N=29

Mean $341,032 $48,861 $25,900 $107,566
S.D. $952,419 $60,324 $10,677 $208,147
Median $82,680 $25,000 $25,000 $32,500

Damages Amount
Mean $111,744 $2,547 $8,900 $13,059
S.D. $268,487 $7,212 $15,144 $20,233
Median $25,910 0 0 $2,000

Mean .54 .08 .25 .35
S.D. .82 .23 .41 .44
Median .35 0 0 .07

Employer Claims                     N=29                                 N=1                                N=4                                 N=4

Mean $177,679 $16,400 $12,234 $29,507
S.D. $496,302 . $10,428 $20,751
Median $19,800 $16,400 $11,967 $37,000

Damages Amount
Mean $51,314 $10,200 $2,344 $22,338
S.D. $114,881 . $10,428 $20,751
Median $7,500 $10,200 $0 $6,926

Mean .39 .62 .25 .71
S.D. .57 . .50 1.21
Median 0 .62 0 .16

Table 5 shows that White Collar employees win something in employment arbitration significantly more frequently than Blue or Pink Collar employees, Pearson Chi-Square = 47.40, DF 3, P< 0.001.

Table 5. Contingency Table showing Damages (rows) by White, Blue, or
Pink Collar Worker (columns), Employee Claims Only

                                                     White                  Blue                      Pink            Unknown                    Total
Damages 117 50%  5 2%  9 4%  16 7%  147 63% 
No Damages 34 15%  26 11%  12 5%  13 6%  85 37% 
Total 151 65%  31 13%  21 9%  29 13%  232 100%

White Collar employees also have significantly higher Outcome. A one-way analysis of variance using White, Blue or Pink Collar as the independent variable and Outcome as the dependent variable was significant, F= 4.11 (DF 3, 211), P<.01. The mean White Collar employee outcome was .54, the Blue Collar mean was .08, and the Pink Collar mean was .25. In other words, white collar employees recovered 54 percent of their claims, blue collar employees recovered 5 percent, and pink collar employees recovered 25 percent of the demand.

There are significantly more Repeat Players among the Employment cases than the Commercial cases, Pearson Chi-Square 34.97, DF 1, P<.001. Similarly, there are significantly more blue and pink collar employees in the Employment cases than in the Commercial cases, Pearson Chi-Square 27.94, DF 3, P<.001. Only 36% of employees in Repeat Player cases had counsel of record appearing in the case file documents.

IV. Discussion and Implications of the Repeat Player Effect for Employment Policy

Conflict resolution scholars have suggested that dispute resolution procedures should be evaluated from not only a microjustice perspective (i.e., how satisfied the participant is in a particular case), but also from a macrojustice perspective, that is, in terms of the patterns in the aggregate distribution of outcomes.(70)  In labor arbitration cases, grievants win something in 52% of all cases.(71) In employment arbitration, using the sample in this study, employees won something in 63% of the total employee claims sample (n=232), including repeat and non-repeat player employers (Table 3). However, looking at the subsample of repeat player cases (n=31), employees won something only 16% of the time. Employees dealing with non-repeat player employers recovered on average 48% of what they demanded, while employees dealing with repeat player employers recovered only 11% of what they demanded (Table 2). These differences in employee success are statistically significant. For lack of better terminology, let us call this the repeat player effect. 

There are several possible explanations for the repeat player effect, and research is ongoing. Galanter's analysis might suggest it is the whole combination of repeat player advantages previously discussed. Specific hypotheses to test are several. One possible explanation is unequal information in arbitrator selection. Another possible explanation is unequal representation in the hearing itself. Researchers have found that a party does better in labor arbitration when it has legal counsel and the other party does not.(72) There may, of course, be systematic differences in the merits of these cases, in that employees in repeat player cases may have weaker legal claims, while employees in the non-repeat player cases may have stronger legal claims. The preliminary evidence is that repeat player cases tend to involve personnel manuals, while non-repeat player cases involve express written contracts. Personnel manuals are more likely to be the product of a repeat player employer unilaterally structuring the employment relationship and dismissal transaction to its best advantage.

This in turn might reflect unequal bargaining power. Professor Marc Galanter noted that litigation requires information and skills, and that complaints come disproportionately from better educated, better informed, and more politically active households.(73) The fact that white collar employees have systematically superior outcomes to blue or pink collar employees is consistent with this theory (see Table 5). They win significantly more frequently than blue or pink collar workers, and they recover a larger percentage of what they demand.

The decision standard for the arbitrator may vary. Employers may get better at screening cases with repeat experience in the process, and learn to settle the cases they otherwise would lose. One theory is that arbitrators respond to the economic incentive of repeat business, but that incentive might be present with the first case an arbitrator hears for an employer. It has been suggested that arbitrators become more familiar with repeat players and/or their lawyers, and develop a relationship of trust.(74) Where the evidence is equally balanced, this may make a difference in the outcome. All of these hypotheses warrant investigation.

The above study does not establish a cause for the repeat player effect. It merely identifies its presence.

Since the time period of this study (1993-94), the AAA and others have signed A Due Process Protocol for Mediation and Arbitration of Statutory Disputes Arising Out of the Employment Relationship.(75) Unlike the previous Commercial or Employment Dispute Resolution Rules, this Protocol attempts to address how the dispute resolution process is structured by the employer. To that extent, it seeks to counterbalance the advantage the repeat player employer enjoys in structuring the employment relationship, which is one of the strategic advantages identified by Galanter.(76) The Protocol recommends freedom of choice of representative, adequate but limited pre-hearing discovery, providing the parties where possible with references from the arbitrators' most recent six cases, arbitrator training in employment law, an enhanced duty for the arbitrator to disclose any relationship that might reasonably constitute or be perceived as a conflict of interest, an arbitration award setting forth a summary of the issues, including the type of dispute and damages or other relief requested and/or awarded, a statement of the issues and the statutory claims, and joint selection and shared compensation of the arbitrator. These provisions were in turn incorporated into the National Rules for the Resolution of Employment Disputes.(77) A provision requires arbitrators to disclose "all information that might be relevant to the standards of neutrality set forth in this Section, including but not limited to service as a neutral in any past or pending case involving any of the parties, or that may prevent a prompt hearing."(78) These changes permit employees to identify when a particular employer is making repeat use of an arbitrator, and might alter the likelihood that the arbitrator would receive repeat appointments. Moreover, they guarantee a right to counsel. These changes might reduce the repeat player effect. The AAA National Rules for the Resolution of Employment Disputes have been modified more recently effective June 1, 1997, but the disclosure rules are unchanged, and the amendments are largely technical.(79) Future research will examine the effects of the Protocol and subsequent rules changes on arbitration outcomes.

The repeat player effect creates the perception of an uneven playing field. In dispute resolution, the perception of a fair process can be as important as the reality of impartiality. The procedural justice literature tells us that satisfaction with the fairness of the process is strongly correlated with satisfaction with the outcome of that process.(80) This is why it is not too early to address possible causes through appropriate reasonable changes in arbitration rules. While the AAA has adopted the Protocol and several changes in its rules in the interest of internal quality control, numerous other providers are competing in the dispute resolution marketplace. These providers may not have the same institutional capacity or procedural protections. What protections that directly address the repeat player effect should providers consider and parties look for in employment arbitration? What reforms might address the imbalance of information about arbitrators in the selection process? The labor and employment law community should explore alternatives to address both the perception of an unequal playing field, and any underlying reality. A few issues to begin examining are disclosure rules, random assignment to lists, and substitute repeat players.

A. The Arbitration Service Provider and Disclosure Rules

The AAA has already attempted to redress the imbalance of information by including new disclosure requirements in its employment arbitration rules. Arbitrators must disclose whether they have had a case with either party in the past or have any ongoing cases.(81) In addition, AAA will provide parties, to the extent possible, the names of other parties who have used the arbitrators' services.(82) This provides information to the employee when an employer is making repeat use of an arbitrator. Similar disclosure rules are required under state law in California.(83) Some might argue that it makes no difference whether an employer is making repeat use of an arbitrator or not, because the arbitrator knows that only the employer is capable of providing repeat business. Others point out that an employee may not know what use to make of the disclosed information. Moreover, disclosure rules as currently structured do not generally reach an arbitrator's record within a particular industry.(84) On the other hand, knowledge that the employer is making repeat use of an arbitrator gives the employee power to make a somewhat more informed choice, one that the employer already can make. The cost of the reform is only the time it takes arbitrators to prepare and make the disclosure.

Employment arbitration users might also inquire as to how many arbitrators are on the provider's panel. What is the likelihood that a single arbitrator will get listed repeatedly? Earlier research indicated that when the probability of getting a case was low in relation to the size of the arbitrator panel, there was no evidence of pro-employer bias in the arbitrated outcomes.(85) In other words, the larger the panel, the more diffuse the distribution of cases, and the lower the probability that any one arbitrator might get many of them. 

It is worthwhile to consider whether a small provider receives a disproportionate amount of its business from a few large employers.(86) Disclosure rules to date operate on the individual arbitrator level. While this makes sense for a large third party administrator with numerous users of its services, it may not provide sufficient information to the consumer dealing with a small provider. Another issue is whether the employer has established its own permanent panel, without resort to a third party administrator, or has sole power to select the arbitrator; Professor Rau notes that such a system is unlikely to be tolerated by the courts.(87) However, one survey found that almost 15 percent of employer plans had a provision for employer selection of the arbitrator.(88) Certainly, the existence of this relationship and the value of the contract to the arbitrator render the employer a repeat player by definition.

B. Random Selection of Arbitrators for Inclusion on Parties' Lists

Random assignment to lists reduces a repeat player employer's ability to select the same arbitrator for multiple cases. While this is possible for large institutional providers with larger rosters, those arbitration service providers with small panels may not have this capacity. Under the most recent version of the AAA Employment Rules, parties are given the whole roster of the local Employment Arbitration Panel from which to select a mutually agreeable arbitrator. The reason for distributing the whole roster includes considerations of diversity; random assignment might result in a list of entirely white, male arbitrators, for example. In labor arbitration, the AAA administrators prepare lists by examining the issue, the parties, and the experience of the potential panelists in an effort to come up with a list that will meet the parties' needs.

In employment arbitration, if the parties cannot reach agreement on an arbitrator, the AAA will provide them with a short list from which an arbitrator will be selected using the alternate strike method. The disclosure rule does not come into effect until after a neutral arbitrator has been appointed; then the AAA has the power to rule on a party's motion to disqualify the arbitrator, and its ruling is for the most part final. Courts are more willing to vacate an arbitrator's award where the arbitrator has failed to make the requisite disclosures, than when he or she discloses information and a motion for disqualification is overruled by the administrator.(89)

This system affords more opportunity for a repeat player to select the same arbitrator on a repeated basis than would the random assignment system. It shifts the burden onto the objecting party, for example, the employee, to move repeatedly to disqualify arbitrators after the employee receives information about the arbitrator's record, until such point as the administrator invokes the alternate strike method from a shortened list. Where an employee is not represented by counsel, as seems commonly to be the case in employment arbitration, this procedure is likely to be somewhat daunting. A random assignment system would limit a repeat player's ability to propose, for example, its top three choices from the entire local roster on an iterative basis in each employment arbitration case to which it is a party.

C. Substitute Repeat Players

Galanter points out the strategic advantages to one-shotters who pool their resources, organize, and act in a coordinated fashion.(90) A repeat player employer, or a member of an association of employers which pools information, may have more information about arbitrators than the non-repeat player employee. For those employees with counsel, it may be possible through the collective action of their lawyers to create an analogous institutional memory. The plaintiffs' bar can pool its resources. Just as sophisticated repeat players such as Kaiser Permanente have created data bases on arbitration awards in the medical malpractice arena, organizations such as the National Employment Lawyers Association (NELA) can create them for employment arbitration awards. The problem with this approach is that it only serves employees with counsel, and particularly, with counsel who are members of NELA, or the attorney organization maintaining the data base.

There are already businesses in the private sector which maintain information on labor arbitrators, and provide it for a fee to parties. They use publicly available labor arbitration reports to create their data base. Where the parties consent and the case is of interest, labor arbitration awards are published by the Bureau of National Affairs, Commerce Clearinghouse, and the American Arbitration Association. Fewer than 10 percent of all labor arbitration awards are published. At present, employment arbitration awards are rarely published. The tradition of confidentiality in commercial arbitration is strong. Thus, a private sector service could not do for employment arbitration what it does for labor arbitration unless there were some mechanism for getting awards. There is no particular incentive for parties to send copies of awards to private sector services. One possibility is that the arbitration service providers could make sanitized copies of cases available to private sector analysts. Either the names of the parties could be redacted from the award, or the awards could be provided under a confidentiality agreement. The private sector service could provide reports on the industries, issues and outcomes associated with a particular arbitrator, without naming the specific parties. However, it is unclear whether unrepresented employees would take advantage of such a service. These services generally market themselves only to lawyers.

Another possibility is for unions to provide a new service to a traditional constituency. Instead of limiting representation in arbitration to members of a traditional certified bargaining unit, they could provide a limited membership to employees at unrepresented work places. There is precedent for this in the area of veterans' affairs. Previously by law, a lawyer representing a client in an appeal concerning denial of compensation for a veteran's injury could receive no more than $10 in compensation.(91) As a result, individual lawyers generally did not take these cases. In response to demand for representation services, veterans' groups which are dues-paying membership organizations, began to provide counsel at no charge for these hearings. Veterans pay annual membership dues, and the representation is provided as one of many services offered to members. While this would put unions in competition with the bar for some employment arbitration cases, it is more likely that unions would represent employees who otherwise could not obtain legal counsel. 

There are many examples of successful membership organizations based upon common interests, such as AARP, the Sierra Club, and the myriad of professional organizations for employees in certain businesses. These entities, after all, copied the organizational model of the union. The advantages of such a system are many. It would provide experienced, quality representation to employees at a relatively modest cost. By adapting the model to this new constituency, unions could provide a base for future organizing efforts. And it would reinstate the union as a repeat player with institutional memory about arbitral records.


Employment arbitration is still in its infancy, but it has begun to walk. It is becoming more widespread, and the question is how the labor and employment relations community can influence its structure. The process has the potential to provide a meaningful opportunity to be heard by employees who otherwise may have no voice in their dismissal. However, there are emerging patterns of repeat player use and significantly different outcomes based on whether an employer is a repeat player in cases decided before the Due Process Protocol and the most recent AAA Rules for employment disputes. These patterns suggest that it was appropriate for the dispute resolution user and provider community to be proactive in developing the structures and institutions that will foster employment arbitration's growth, such as the Protocol. The Protocol is a major step forward in self-regulation, but not all providers have adopted the Protocol, and not all employers adopt plans that comport with basic notions of due process, such as the right to a jointly selected arbitrator.(92) Moreover, it remains to be seen how effectively signatories can enforce its terms. It is necessary to assess the employment arbitration process and its outcomes as we continue to move incrementally forward with reform. 

* Assistant Professor, Indiana University School of Public and Environmental Affairs, B.A. Smith College, J.D. University of Connecticut. The author is a member of the American Arbitration Association Labor Arbitrator Panel, the Federal Mediation and Conciliation Service Labor Arbitrator Panel, and the American Arbitration Association Employment Disputes Panel. She is also Co-Director, Indiana Conflict Resolution Center.

The author gratefully acknowledges the assistance of George Friedman, Richard Naimark, Barbara Brady, and Frank Zotto of the American Arbitration Association for making available the materials for this study, and for their helpful comments on the manuscript. In addition, the author would like to thank Prof. Alan Scott Rau for helpful discussions during preparation of the manuscript, and a thank you for the helpful comments of anonymous reviewers. The opinions expressed here are those of the authors and are not in any way representing the views of the AAA. This research was supported in part by a grant from Indiana University, Research and the University Graduate School. A portion of this article in an earlier form was presented at the 1997 Meeting of the Industrial Relations Research Association, Refereed Papers, Labor and Employment Law, New Orleans, LA, January 5, 1997, and appeared in the 49th Proceedings of the Industrial Relations Research Association at 201-11.

1.  See, e. g, Commission on the Future of Worker-Management Relations, U.S. Dep't of Labor and U.S. Dep't of Commerce, Report and Recommendations, GPO-CTLG, L1.2-F 98/2(1994); Terry A. Bethel, Wrongful Discharge: Litigation or Arbitration?, 1993 J. Disp. Resol. 289 (1993).

2.  Peter Feuille & Denise R. Chachere, Looking Fair Or Being Fair: Remedial Voice Procedures In Nonunion Workplaces, 21 J. Mgmt. 1, 27-42 (1995). 

3.  U. S. Gen. Acct. Off., Employment Discrimination-Most Private-Sector Employers Use Alternative Dispute Resolution, GAO/HEHS-95-150 (1995).

4.  See, e.g., Steven M. Kaufmann & John A. Chanin, Directing the Flood: The Arbitration of Employment Claims, 10 lab. Law. 217 (1994); Christopher C. Miller & Brian D. Poe, Arbitrating Employment Claims: the State of the Law, 46 Lab. L.J. 195, 195-204 (Apr. 1995); Jay S. Siegel, Changing Public Policy: Private Arbitration To Resolve Stat utory Employment Disputes, 13 Lab. Law. 87, 87-106 (1997) (advocating voluntary, not mandatory arbitration).

5.  Allan E. Lind & Tom R. Tyler, The Social Psychology of Procedural Justice (1988); Allan E. Lind, et al., In the Eye of the Beholder: Tort Litigants' Evaluations of Their Experiences in the Civil Justice System, 24 L. & Soc'y Rev. 953 (1990). 

6.  James N. Dertouzos, et al., The Legal and Economic Consequences of Wrongful Termination (1992); James N. Dertouzos & Lynn A. Karoly, Labor-market responses to employer liability (1988).

7.  Jean R. Sternlight, Panacea Or Corporate Tool?: Debunking The Supreme Court's Preference For Binding Arbitration, 74 Wash. U. LQ. 637 (1996).

8.  Margaret A. Jacobs, Some Workers Resist Mandatory Arbitration, Star Trib., Feb. 5, 1995 at 1J.

9.  See, e.g., Lewis Maltby, Paradise Lost -- How the Gilmer Court Lost the Opportunity for Alternative Dispute Resolution to Improve Civil Rights, 12 N.Y.L.Sch.J.Hum.Rts. 1 (1994); Robert J. Rabin, The Role of Unions in the Rights-based Workplace, 25 U.S.F.L.Rev. 169 (1991); James Wallihan, Too Little, Too Late: The Limits of "Stand-Alone" Arbitration in Discharge Cases, 21 Lab. Stud.J. 39 (1996).

10.  See, e.g., Reginald Alleyne, Statutory Discrimination Claims: Rights "Waived" and Lost in the Arbitration Forum, 13 Hofstra Lab. L. J. 381, 403, 426 (1996); Samuel Estreicher, Arbitration of Employment Disputes Without Unions, 66 Chi.-KentL.Rev 753 (1990); Julius G. Getman, Labor Arbitration and Dispute Resolution, 88 Yale L. J. 916, 929-930 (1979); Robert A. Gorman, The Gilmer Decision and the Private Arbitration of Public-Law Disputes, 1995 U. Ill. L. Rev. 635, 669 (1995); Maltby, supra note 9 at 4-5 (1994); Sternlight, supra note 7 at 685.

11.  Richard Edwards, Rights at Work: Employment Relations in the Post-Union Era 221 (1993); Tia Schneider Denenberg & Richard V. Denenberg, The Future Of The Workplace Dispute Resolver, 49 Disp. Resol. J. 48(1994).

12.  For excellent reviews of the case law on arbitrator impartiality, see, Stephen L. Hayford, Law in Disarray: Judicial Standards for Vacatur of Commercial Arbitration Awards, 30 Ga. L. Rev. 731 (1996); Alan S. Rau, Integrity in Private Judging, 38 S. Tex. L. Rev. 485 (1997); see generally, Ian R. Macneil, et al., Federal Arbitration Law (1994). 

13.  105 F.3d 1465 (D.C. Cir. 1997).

14.  See, e.g., Harry T. Edwards, Alternative Dispute Resolution: Panacea or Anathema?, 99 Harv. L. Rev. 668 (1986).

15.  Cole, 105 F.3d at 1470, citing 9 U.S. §§1-2 (1994).

16.  Sternlight, supra note 7, at 637-712.

17.  Cole, 105 F.3d at 1473-74.

18.  Id. at 1475, citing Bernard D. Meltzer, Ruminations About Ideology, Law, and Labor Arbitration, Proceedings of the 20th Annual Meeting of the National Academy of Arbitrators 1, 3-4 (1967); Getman, supra note 10, at 929-30.

19. Cole, 105 F.3d at 1476 (citations omitted).

20.  Id. at 1477-78 (citing Gilmer v Interstate/Johnson Lane Corp., 500 U.S. 20 (1991)).

21.  Available at

22.  Cole, 105 F.3d at 1485.

23. Id.

24.  Id. at 1485, n.17.

25.  Id. at 1475.

26.  Marc Galanter, Why the "Haves" Come Out Ahead: Speculations on the Limits of Legal Change, 9 L. & Soc'y Rev. 95 (1974).

27. Id. at 97.

28.  Id. at 96-103.

29.  Id. at 109-10.

30.  Id. at 110.

31.  Id.

32.  Lauren B. Edelman, et al., Professional Construction of the Legal Environment: The Inflated Threat of Wrongful Discharge, 9 L. & Soc'y Rev. 95 (1974).

33.  The term wrongful discharge is used to refer to causes of action sounding in tort or contract. Historically, all employment was terminable at will, that is, with or without cause in the discretion of the employer. Jay M. Feinman, The Development of the Employment at Will Rule, 20 Am. J.Legal Hist.118 (1978). Since the late 1970s, state courts have begun to limit the doctrine of employment at will through a civil cause of action for retaliatory or wrongful discharge. In general, an employee may recover damages if the employer terminated employment in retaliation for employee conduct that is protected by a substantial and important public policy. The employee has the burden of proving that the conduct is protected, that the public policy at stake is sufficiently important, and that the employer actually fired the employee in retaliation for the conduct. The burden of proof then shifts to the employer. The employer may avoid liability by proving a permissible ground for dismissal. What represents a sufficiently important public policy varies from state to state. The highest courts of more than two-thirds of the states have recognized the cause of action. See generally, Lex K. Larson, Unjust Dismissal §§2-18 (1992); Michael A. DiSabatino, Annotation, Modern Status of Rule That Employer May Discharge At-will Employee for Any Reason, 12 A.L.R. 544 (1982). For a review of major pieces in the literature, see Paul C. Weiler, Governing the Workplace, Harvard University Press, Cambridge Mass.(1990). See also, Lawrence Blades, Employment at Will vs. Individual Freedom: Limiting the Abusive Exercise of Employer Power, 67 Colum. L. Rev. 1404 (1967); Martin H. Malin, Protecting the Whistleblower from Retaliatory Discharge, 16 U. Mich. J.L. Ref. 277 (1983); Clyde Summers, Individual Protection Against Unjust Dismissal: Time for a Statute,62 Va. L. Rev. 481 (1976); Note, Protecting At Will Employees Against Wrongful Discharge: The Duty to Terminate Only in Good Faith, 93Harv. L. Rev. 1816 (1980).

34.  Dertouzos et al., supra note 6; Dertouzos & Kanoly, supra note 6.

35.  For a more complete discussion of the implications of employment arbitration for enforcement of public policy, see, Martin H. Malin & Robert F. Ladenson, Privatizing Justice: A Jurisprudential Perspective on Labor and Employment Arbitration from the Steelworkers Trilogy to Gilmer, 44 Hastings L.J. 1187 (1993).

36.  See generally Patrick J. Hardin, The Developing Labor Law 1409-88 (3d ed. 1992).

37.  David E. Bloom and Christopher L. Cavanagh, An Analysis of the Selection of Arbitrators, 76 Am.Econ.Rev. 408 (1986); Steven S. Briggs & John Anderson, An Empirical Investigation of Arbitrator Acceptability, 19 Indus. Rel. 163 (1980); Robert J. Thornton & Perry A. Zirkel, The Consistency and Predictability of Grievance Arbitration Awards, 43 Indus. & Lab. Rel. Rev. 294 (1990).

38.  Bloom, supra note 37, at 408-22; Thorton, supra note 37, at 294-307.

39.  Thornton, supra note 37, at 294-307.

40.  For a discussion of the empirical evidence for and against this proposition, see Henry S. Farber, Splitting-the-difference in Interest Arbitration, 35 Indus. & Lab. Rel. Rev. 70 (1981); Peter Feuille & Susan Schwochau, The Decisions of Interest Arbitrators, 43 Arb. J. 28 (1988); William W. Notz & Frederick A. Starke, Arbitration and Distributive Justice: Equity or Equality? 72 J. Applied Psychol. 359 (1987); Susan Schwochau & Peter Feuille, Interest Arbitrators and Their Decision Behavior, 27 Indus. Rel. 37 (1988).

41.  Galanter, supra note 26, at 114-19.

42.  Cole, 105 F.3d at 1485.

43.  Robert N. Block & Jack Stieber, The Impact of Attorneys and Arbitrators on Arbitration awards, 40 Indus. & Lab. Rel. Rev. 543 (1987); Stephen M. Crow & James W. Logan, Arbitrators' Characteristics and Decision-making Records, Gender of Arbitrators and Grievants, and the Presence of Legal Counsel as Predictors of Arbitral Outcomes, 7 Employee Resp. & Rts. J. 169 (1994); see Galanter, supra note 26, at 114, n.45.

44.  Lisa B. Bingham, Emerging Due Process Concerns in Employment Arbitration: A look at Actual Cases, 47 Lab. L.J. 108 (1996).

45.  Galanter, supra note 26, at 116.

46.  See, Id. at 117.

47.  See William M. Howard, Arbitrating Claims of Employment Discrimination: What Really Does Happen? What Really Should Happen? Disp. Resol. J., Oct.-Dec. 1995, at 40, 45 (detailing a study which found that mean and median jury verdicts in employment discrimination cases were at least three times higher than the comparable mean and median arbitration awards).

48.  Nevertheless, similar comparisons of outcomes have been done in the medical malpractice context for arbitrated and litigated cases. See, e.g., Thomas B. Metzloff, Comment, Alternative Dispute Resolution Strategies in Medical Malpractice, 9 Alaska L. Rev. 429, 439 n.43 (1992)(citing a study whcih found that plaintiffs prevailed in 22% of actually arbitrated cases, and 18% of actually litigated cases); Neil Vidmar, The Unfair Criticism of Medical Malpractice Juries, 76 Judicature 118, 123(1992) (differences in mean damage award in controlled experiment on same hypothetical case facts using jurors and arbitrators found to be statistically insignificant).

49.  See Mei L. Bickner, et al., Developments in Employment Arbitration, Disp. Resol. J., Jan. 1997, at 8.

50.  See, Howard, supra note 47, at 43.

51.  See Id. at 44 (observing that while defense counsel estimated 79% of discrimination cases designated for litigation settled, only 31% of the arbitration cases did).

52.  Galanter, supra note 26, at 118.

53.  See Bickner et al., supra note 49, at 79 (reporting that 25 percent of surveyed employers limited arbitration to discharge). 

54.  See Galanter, supra note 26, at 97-98.

55.  See Debra J. Mesch, Grievance Arbitration in the Public Sector: A Conceptual Framework and Empirical Analysis of Public and Private Sector Arbitration Cases, Rev. Pub. Personnel Admin., Fall 1995, at 22, Tables 1 and 3.

56.  See Peter Feuille, Dispute Resolution Frontiers in the Unionized Workplace in Workplace Dispute Resolution 17 (Sandra E. Gleason, ed. 1997).

57.  See Rau, supra note 12, at 523.

58.  For reviews of the literature, see Block & Stieber, supra note 41; Thornton & Zirkel, supra note 35. For additional literature review and model of factors at work in labor arbitrator decison-making, see Stephen M. Crow & James W. Logan, Arbitrators' Characteristics and Decision-Making Records, Gender of Arbitrators and Grievants, and the Presence of Legal Counsel as Predictors of Arbitral Outcomes, 7 Employee Resp. & Rts. J. 169 (1994).

59.  See Feuille & Schwochau, supra note 40; Schwochau & Feuille, supra note 40.

60.  Empirical research examining gender effects in labor arbitration has produced mixed results, some in favor of women, some finding adverse gender effects, and some finding no significant effects at all. It has been reported that women are less likely to be suspended or terminated in disciplinary processes for alcohol abuse, see Janice M. Beyer & Harrison M. Trice, A Field Study of the Use and Perceived Effects of Discipline in Controlling Work Performance, 27 Acad Mgmt. J. 743, 752 (1984). Others report women will be treated more favorably in the grievance process. See Dan R. Dalton & William D. Todor, Gender and Workplace Justice: A Field Assessment, 38 Personnel Psychol. 133, 145 (1985). Bemmels found women were treated more leniently in his field studies of actual arbitration cases. See Brian Bemmels, The Effect of Grievants' Gender and Arbitrator Characteristics on Arbitration Decisions, 15 Lab. Stud. J., Summer 1990, at 48, 61; Brian Bemmels, The Effect of Grievants' Gender on Arbitrator's Decisions, 41 Indus. & Lab. Rel. Rev. 251, 259 (1988); Brian Bemmels, Gender Effects in Discharge Arbitration, 42 Indus. & Lab. Rel. Rev. 63, 73 (1988); Brian Bemmels, Gender Effects in Discipline Arbitration: Evidence from British Columbia, 31 Acad. Mgmt. J. 699, 704 (1988). See also Allen Ponak, Discharge Arbitration and Reinstatement in the Province of Alberta, Arb. J., June 1987, at 39. Another researcher found that women were treated more leniently by student arbitrators considering a drug-testing case. See Sharon L. Oswald, Students as Arbitrators: An Empirical Investigation, 30 Indus. Rel. 286, 289-91 (1991).

However, Mesch analyzed cases taken to labor arbitration in the public sector and found a potential sex bias against women. See Debra J. Mesch, Arbitration and Gender: An Analysis of Cases Taken to Arbitration in the Public Sector, 24 J. Collective Negotiations Pub. Sector 207, 214 (1995). Her results indicate that women tend to lose more cases and receive fewer compromise outcomes -- even when controlling for the severity of the grievance. Other work reports no statistically significant effects as a function of gender at the grievance stage or at labor arbitration. See Block & Stieber, supra note 41, at 549-52; Crow & Logan, supra note 56, at 180-81; Dan R. Dalton, et al., The Iron Law of Paternalism in Organizational Justice Outcomes: Two Field Assessments, 46 Lab. L. J. 669, 673 (1995); Dan R. Dalton & William D. Todor, Composition of Dyads as a Factor in the Outcomes of Workplace Justice, 28 Acad. Mgmt. J. 704, 708 (1985); Sharon L. Oswald & Steven B. Caudill, Experimental Evidence of Gender Effects in Arbitration Decisions, 4 Employee Resp. & Rts. J. 271, 277 (1991); Clyde Scott & Elizabeth Shadoan, The Effect of Gender on Arbitration Decisions, 10 J. Lab. Res. 429, 433 (1989); K. Dow Scott & G. Stephen Taylor, An Aanalysis of Absenteeism Cases Taken to Arbitration: 1975-1981, Arb J., Sept. 1983, at 61, 69: Kenneth Wm. Thornicroft, Gender Effects in Grievance Arbitration...Revisited, Lab. Stud. J., Winter 1995, at 35, 43. None of the empirical works examine gender of the grievant as a factor in employment arbitation, although the popular press has speculated that the male-dominated panels would discriminate against women. See Judith P. Vladeck, Validity of ADR for Job Disputes: "Yellow Dog Contracts" Revisited, N.Y. L.J., July 24, 1995, at 7. Given that labor arbitration panels are also male-dominated, and the results of careful empirical research on gender effects in labor arbitration are so mixed, it seems at the very least premature to conclude there is gender bias in employment arbitration panels.

61.  For recent reviews of the literature, see Peter Feuille & Robert L. Hildebrand, Grievance Procedures and Dispute Resolution, in Handbook of Human Resource Management 330 (Gerald R. Ferris et al. Eds., 1995); Feuille, supra note 54. See David Lewin, Conflict Resolution and Management in Contemporary Work Organizations: Theoretical Perspectives and Empirical Evidence Research in the Sociology of Orgzanizations 12, 167-209. See also Richard B. Peterson & David Lewin, The Nonunion Grievance Procedure: A Viable System of Due Process?, 4 Employee Resp. & Rts. J. 1 (1990).

62.  See Howard, supra note 47.

63.  See Bickner et al., supra note 49, at 79. For a recent review of the evolution of the debate and survey of employer-promulgated plans, see Id. at 10-13. 

64.  Lisa B. Bingham, Is There a Bias in Arbitration of Nonunion Employment Disputes? An Analysis of Actual Cases and Outcomes, 6 Int'l J. Conflict Mgmt. 369, 384 (1995).

65.  American Arbitration Association, Employment Dispute Resolution Rules, Publication No. AAA121-10M-12/92 (1993).

66.  See American Arbitration Association, Commercial Arbitration Rules, Revised Rules and Fees for Cases Filed on or after May 1, 1992.

67.  See Bingham, supra note 44, at 115-17.

68.  See William D. Todor & Crystal L. Owen, Deriving Benefits from Conflict Resolution: A Macrojustice Assessment, 4 Employee Resp. & Rts. J. 37, 38-39 (1991).

69.  See Lisa B. Bingham, Unequal Bargaining Power: An AlternativeAccount for the Repeat Player Effect in Employment Arbitration (forthcoming 1998)(manuscript for presentation in the Industrial Relations Research Association Refereed Papers Competition in Labor and Employment Law)(on file with author). 

70.  See Philip Brickman et al., Microjustice and Macrojustice, in The Justice Motive in Social Behavior 173 (Melvin J. Lerner & Sally C. Lerner eds., 1981); Todor & Owen, supra note 68, at 38-39.

71.  See Mesch, supra note 55, at 30 tbl.1.

72.  See Block & Stieber, supra note 43, at 553; Crow & Logan, supra note 58, at 181-82.

73.  See Marc Galanter, Reading the Landscape of Disputes: What We Know and Don't Know (and Think We Know) About Our Allegedly Contentious and Litigious Society, 31 U.C.L.A. L. Rev. 4, 23 (1983).

74.  See Galanter, supra note 26, at 99 n.9. A related phenomenon has been called "agency capture theory" in the public administration literature; agency personnel become more familiar with the representatives of regulated entities, and take their concerns into account. See Paul J. Quirk, Industry Influence in Federal Regulatory Agencies (1981); Kay Lehman Schlozman & John T. Tierney, Organized Inerests and American Democracy 330-55 (1986); Paul Sabatier, Social Movements and Regulatory Agencies: Toward a More Adequate--And Less Pessimistic--Theory of "Clientele Capture," 6 Pol'y Sci. 301 (1975).

75.  See American Arbitration Association, Pub. No. AAA219-10M-8/96, Guide for Employment Arbitrators 1 (1996). Most recently, the American Bar Association has approved the Protocol. See ABA Adopts Due Process Protocol, Disp. Resol. Times (Am. Arb. Ass'n,New York, N.Y.), Summer 1997, at 7.

76.  See Galanter, supra note 26, at 98.

77.  Cf. American Arbitration Association, Pub. No. AAA121-20M-6/96, National Rules for the Resolution of Employment Disputes (1996).

78.  Id. at 15.

79.  Cf. Modifications Made in the AAA's Employment Rules, Disp. Resol. Times (Am. Arb. Ass'n, New York, N.Y.), Summer 1997, at 3 (not identifying a change in the disclosure rules).

80.  See Lind & Tyler, supra note 5, at 66-70.

81.  See American Arbitration Association, Pub. No. AAA121-20M-6/97, National Rules for the Resolution of Employment Disputes 15-16 (1997).

82.  See Id. at 16-17.

83.  See Cal. Civ. Code § 1281.9 (West Supp. 1997).

84.  See discussion in Rau, supra note 12, at 525-26 & n.148-50.

85.  See Bingham, supra note 64, at 382-83.

86.  Rau observes that having the administrator appoint the arbitrator does not entirely solve the problem of structural bias. The provider itself may have an interest in repeat business. See Rau, supra note 12, at 527-28.

87.  See Id. at 528 & n.154.

88.  See Bickner et al., supra note 49, at 80.

89.  See Rau, supra note 12, at 492-94.

90.  See Galanter, supra note 26, at 141-43.

91. See Walters v. National Ass'n of Radiation Survivors, 473 U.S. 305, 308 (1985)(citing 38 U.S.C. § 3404 to allow attorneys to charge reasonable attorney fees in such cases, and the current section is 38 U.S.C. § 5904 (1994).

92.  See Bickner et al., supra note 49, at 80. Purchase Article Back | Chicago-Kent Home | Journal Home