Long v. Trans World Airlines, Inc., 761 F. Supp. 1320 (N.D. Ill. 1991)

*1321 *1322 Samuel W. Witwer Jr., Richard F. Sarna, Witwer, Moran, Burlage & Witwer, Wayne B. Giampietro, Thomas R. Meites, Lynn Sara Frackman, Julia E. Getzels, Meites, Frackman & Mulder, Chicago, Ill., for plaintiffs.

Gordon B. Nash, Jr., Laurence A. Carton, Mark E. Furlane, Gardner, Carton & Douglas, Chicago, Ill., Michael A. Katz, Trans World Airlines, New York City, for defendants.

ROVNER, District Judge.

This is a class action lawsuit brought by approximately 3,000 flight attendants against their employer, Trans World Airlines, Inc. ("TWA"), challenging the company's refusal to provide them with designated rights letters pursuant to the Airline Deregulation Act, 49 U.S.C.App. § 1552, and its applicable regulations, 29 C.F.R. Pt. 220, after plaintiffs went on strike and were not re-hired. On August 18, 1988, the Court granted plaintiffs' motion for class certification pursuant to Fed.R.Civ.P. 23(b) (3), rejecting defendant's argument that an abundance of individual issues with respect to damages made a class action unwieldy. The Court stated that a number of methods were available to manage these issues, but it deferred consideration of which method should be adopted.

On January 26, 1989, the Court granted plaintiffs' motion for summary judgment as to liability. Long v. Trans World Airlines, Inc., 704 F. Supp. 847 (N.D.Ill.1989). That decision has been affirmed on appeal. Long v. Trans World Airlines, Inc., 913 F.2d 1262 (7th Cir. 1990). The Court must now determine how litigation of the damages issues shall proceed.

Relevant background facts may be found in the earlier opinions of this Court and the Seventh Circuit. An additional development is the completion of questionnaires by over 2,000 of the class members for use in another case in which they are also plaintiffs, Independent Federation of Flight Attendants v. Trans World Airlines, Inc., No. 86-6084-CV-SJ-6 (W.D.Mo.).^[1] Those "IFFA questionnaires" consist of six forms which request the following information:

1. Form A seeks 23 types of personnel information, such as name, address and marital status.

2. Form B seeks information concerning earnings and expenses during the period for which damages are sought in this case, information concerning other employment during that period, information relating to job search and employment expenses, information relating to job search activity, and information concerning any periods of inability to work.

3. Form C seeks information concerning pre-strike employment and use of child care services during the year preceding the strike.

4. Form D seeks information concerning medical and dental expenses.

5. Form E seeks information concerning quarterly travel pass usage.

6. Form F seeks information concerning insurance expenses.

The parties have submitted a number of briefs on the damages issues, addressing such issues as discovery, proximate cause, mitigation, manner of computing lost wages, and availability of compensatory damages.

Of foremost importance at this stage of the litigation are determinations regarding how discovery on the damages issue should proceed and whether defendant should be permitted to pursue a defense that plaintiffs failed to mitigate their damages.

Defendant seeks permission to serve interrogatories and discovery requests *1323 on all class members.^[2] Defendant suggests that it report to the Court after it has reviewed the responses to this discovery in order to discuss the necessity of additional discovery.

Plaintiffs request that discovery proceed on a "sampling" basis, with the initial step being the selection of a random sample of class members. Defendants would then be allowed to serve discovery only on the sample. Plaintiffs also suggest using affidavits from all class members to prove only incidental and consequential damages.^[3]

The parties do not dispute the appropriateness of discovery on the damages issues, and they do not dispute the necessity of an eventual hearing. They differ, however, as to the propriety of using a sampling method to limit discovery. This appears to be an issue of first impression in this context. Plaintiffs primarily argue that full-blown discovery is unnecessary and unduly burdensome, but they have provided no authority squarely authorizing sampling in this situation. Defendants argue that the presence of individual issues entitles them to discovery from each plaintiff, but they have provided no authority which prohibits using sampling to limit discovery. After reviewing the law concerning discovery and the law governing class actions, the Court agrees with plaintiffs that a sampling method should be used.

The issue may be broken down into two questions. First, is sampling ever appropriate over a party's objection? Second, if sampling is appropriate in some instances, is this case such an instance?

The Court finds initially that sampling is a permissible procedure. If sampling were universally prohibited, it could only be because a party had an absolute right to obtain discovery from each adverse party or to introduce evidence concerning each such party. Defendants have not identified any source for such a right. Fed.R.Civ.P. 26(c) provides district courts broad discretion to limit discovery when "justice requires to protect a party or person from annoyance, embarrassment, oppression, or undue burden or expense." See Seattle Times v. Rhinehart, 467 U.S. 20, 36, 104 S. Ct. 2199, 2209, 81 L. Ed. 2d 17 (1984). In other class actions, courts have not required absolute precision as to damages and have allowed damages to be proven by reference to a class as a whole, rather than by reference to each individual class member. In Stewart v. General Motors Corp., 542 F.2d 445 (7th Cir.1976), cert. denied, 433 U.S. 919, 97 S. Ct. 2995, 53 L. Ed. 2d 1105 (1977), the defendant was found to have committed discriminatory hiring and promotion practices in violation of Title VII of the Civil Rights Act of 1964. The court of *1324 appeals outlined the approach to be taken in determining the amount of backpay to be awarded. It "set down three general rules:"

(1) unrealistic exactitude is not required; (2) ambiguities in what an employee or group of employees would have earned but for discrimination should be resolved against the discriminating employer; (3) the district court, far closer to the facts of the case than we can ever be, must be granted wide discretion in resolving ambiguities.

542 F.2d at 452. In applying these principles to the Stewart case, the court recognized that the first issue:

is whether backpay should be awarded on an individualized basis to particular employees or on a classwide basis to be divided among the entire group which plaintiffs represent. Where possible, an individualized remedy should be utilized because it will best compensate the victims of discrimination without unfairly penalizing the employer.

Id. With respect to individuals who had been unfairly denied promotions, the court determined that damages should be computed on a classwide basis, even though that might create a windfall for some class members and undercompensate others. Id. The court then noted that a number of techniques are available for approximating damages in class actions. Id. at 453. Of these techniques, the court determined that a "test period" approach was preferable in the specific case before it:

The district court should trace over a period of time the history of a group of white hourly employees which is comparable to the group of black employees constituting the class receiving the backpay award. By determining which members of the control group achieved salaried status, how high within the salaried ranks they rose, and what their increases in salary were, it will be possible on remand to estimate for that period the increment in salary which the black hourly employees as a group would have earned had they been permitted to advance to salaried positions. This figure will serve as a benchmark for computation of the actual award.

Id. See also Liberles v. County of Cook, 709 F.2d 1122 (7th Cir.1983) (Stewart "does not stand for the proposition that individualized hearings are necessary;" affirming classwide determinations of relief in Title VII case); Bowe v. Colgate, Palmolive Co., 489 F.2d 896, 902-04 (7th Cir.1973) (using "test period" approach).

Commentator Herbert Newberg has discussed at length the propriety of aggregate proof of damages. "In appropriate circumstances, various individual issues can be eliminated by use of classwide proof of aggregate damages and by the employment of methods for distribution of aggregate recoveries by means other than individual proofs of damages suffered." 2 Newberg, Class Actions 332 (1985).

[T]he use of proofs of aggregate class recoveries, when feasible, eliminates the need for individual damage proofs at trial. Aggregate class recovery proofs are becoming a recognized means of proceeding in class adjudications, provided such means of proof are able to be developed in the particular circumstances and otherwise comply with applicable evidentiary rules.

Id. at 336. In determining what method to use,

the court must be careful to balance two basic yet opposing interests. The defendants must be given fair opportunity to contest the validity of individual claims and to present defenses unique to particular claims. At the same time, consistent with the judicial economy and access to judicial relief objecti[ve]s of class actions, `class members should be able to secure the relief to which they are entitled without expending more money and effort tha[n] is necessary.'

Id. at 337 (citation omitted). Newberg identifies sampling techniques in particular as an appropriate method for class proof of damages. Id. at 350-51, citing Manual for Complex Litigation 2.71 (5th ed. 1982); Fed.R.Evid. 703; Hart-Scott-Rodino Antitrust Improvements Act of 1976, Pub.L. No. 94-435, tit. III, § 301, 90 Stat. 1383, *1325 1395, enacting 15 U.S.C. § 15d; Developments In The Law Class Actions, 89 Harv.L.Rev. 1318, 1447 n. 275 (1976). Thus, to the extent defendant argues in this case that sampling is improper because it has an absolute right to individualized determinations of damages, its contention must be rejected as contrary to the case law and to the policies governing class actions.

Sampling itself has been used previously, although apparently not in the precise circumstances of this case. In Romasanta v. United Airlines, Inc., No. 70 C 1157 (N.D. Ill. Sept. 12, 1984), Judge Moran used a sampling method, with appropriate statistical models, to determine backpay in an employment discrimination case. None of the parties made an objection to the use of such a method, although they disagreed on particular elements of the models and particular figures to be used in the computations.

Sampling and statistical analysis were also approved for use in computing damages on a classwide basis in In re Coordinated Pretrial Proceedings in Antibiotics Antitrust Actions, 333 F. Supp. 278 (S.D.N. Y.), mandamus denied, 449 F.2d 119 (2d Cir.1971). In that case, the court rejected arguments very similar to those raised here:

[T]he defendants contended that individual proof of damages at trial by the class members would require so much time and create such complexities for the parties, courts and jury that the actions would be unmanageable. On the other hand, they contended that any other method of proving this issue at trial would substantially prejudice them and violate their rights to a jury trial and due process.

333 F. Supp. at 287. The court rejected these arguments and certified the case as a class action. The court found that it was appropriate to determine damages by using as a basis the representative plaintiffs or by using sampling to intensively study a representative group. Id. at 287-88. The court also described the irony in the defendants' opposition to use of statistical modelling:

Most important management decisions in the business world in which these defendants operate are made through the intelligent application of statistical and computer techniques and these class members should be entitled to use the same techniques in proving the elements of their cause of action. The court is confident that they can be successfully utilized in the courtroom and that their application will allow the consumers to protect their rights while freeing the court and the defendants of the specter of unmanageability. In these circumstances the court cannot conclude that the defendants are constitutionally entitled to compel a parade of individual plaintiffs to establish damages.

Id. at 289. See also Catlett v. Missouri Highway and Transportation Comm'n, 828 F.2d 1260, 1267 (8th Cir.1987) (backpay determined on classwide basis in Title VII case), cert. denied, 485 U.S. 1021, 108 S. Ct. 1574, 99 L. Ed. 2d 889 (1988); Segar v. Smith, 738 F.2d 1249, 1291 (D.C.Cir.1984) (although individual hearings were ordered with respect to certain claimants, they were not necessary with respect to damages for discrimination in promotion decisions affecting other claimants), cert. denied, 471 U.S. 1115, 105 S. Ct. 2357, 86 L. Ed. 2d 258 (1985); Green v. United States Steel Corp., 640 F. Supp. 1521, 1539 (E.D. Pa.1986) (in Title VII case, it was "appropriate to address mitigation by estimating outside wages on a class-wide basis" using profiles of the class members to estimate earnings and "accounting in some reasonable manner for persons who made no attempt to find [comparable] work"), aff'd in part, vacated in part on other grounds, 843 F.2d 1511 (3d Cir.1988), vacated on other grounds, 490 U.S. 1103, 109 S. Ct. 3151, 104 L. Ed. 2d 1015 (1989); Thomas v. City of Evanston, 610 F. Supp. 422, 435-36 (N.D.Ill.1985) (in Title VII case, class-wide backpay award based on statistical computations was appropriate despite ignorance as to which class members would have been hired); E.E.O.C. v. Chicago Miniature Lamp Works, 640 F. Supp. 1291, *1326 1298-1300 (N.D.Ill.1986) (in Title VII case, individual hearings not necessary to determine entitlement to backpay; classwide award would be determined on statistical basis).^[4]

Thus, courts have approved various methods of discovering and determining damages in class actions on the basis of classwide, rather than individualized proof of damages, and the use of statistics and representative samples are one such legitimate method. The next question, then, is whether such a method is appropriate in the instant case.

Plaintiffs' primary concern is the burden which classwide damages discovery would impose on the class members and their counsel. They point out that conscientious attorneys do not merely forward discovery requests to their clients for completion; rather, they draft the answers themselves after consulting with their clients. Plaintiffs estimate that such consultation and drafting of answers would consume at least eight hours per class member, or over 24,000 attorney hours total. Plaintiffs thus contend that individual discovery requests propounded to each class member would constitute an impracticable method of proceeding. They suggest that a statistical model, based on discovery from a random sample, can adequately address the various damages issues.

Defendant contends that it is entitled to serve discovery on each class member for two principle reasons.^[5] First, defendant argues that plaintiffs have not yet provided a model which may be utilized to compute damages. It is premature, however, to expect plaintiffs to present an actual model before the Court has even ruled that a statistical model will be employed. The size of the sample must be determined in light of sound statistics principles, and the elements of the model may depend on information gathered in discovery.

Second, defendant contends that individualized discovery is necessary because each class member may be differently situated with respect to damages. There are individual issues with respect to injury in fact, proximate cause,^[6] qualification for other employment, desirability of other employment, and mitigation of damages. However, the very purpose of a statistical model is to account for individual issues to take into account, for instance, what percentage of class members would have been able to secure alternative employment at another air carrier. Defendant offers no reason why a model cannot be constructed which accounts for the individual issues it identifies. Accordingly, the presence of individual issues does not in itself render the use of a model inappropriate.

Furthermore, defendant's suggestion that discovery from all class members is mandated simply by virtue of the existence of individual issues is contrary to Clark v. Universal Builders, Inc., 501 F.2d 324 (7th Cir.), cert. denied, 419 U.S. 1070, 95 S. Ct. 657, 42 L. Ed. 2d 666 (1974). In Clark, a *1327 housing discrimination case, the district court dismissed those class members who failed to answer interrogatories or appear for depositions. The court of appeals reversed, stating that although discovery from absent class members is appropriate in special circumstances, the defendants had not met their burden of demonstrating the necessity of the discovery. 501 F.2d at 340-41.

Although defendant does not make this argument clearly, it appears also to argue that a sampling method is inappropriateeven a denial of due process because it cannot be 100 percent accurate. As described above, however,^[7] damages in class actions (like other cases) need not be 100 percent accurate. As Newberg states,

Aggregate computation of class monetary relief is lawful and proper. Challenges that such aggregate proof affects substantive law and otherwise violates the defendant's due process or jury trial rights to contest each members claim individually, will not withstand analysis. ... Virtually all circuits that have considered this issue ... have expressly condoned aggregate proofs of damages in other contexts [than antitrust].

Id. at 352-53, citing Simer v. Rios, 661 F.2d 655 (7th Cir.1981), cert. denied, 456 U.S. 917, 102 S. Ct. 1773, 72 L. Ed. 2d 177 (1982); Roper v. Consurve, Inc., 578 F.2d 1106 (5th Cir.1978), aff'd on other grounds, 445 U.S. 326, 100 S. Ct. 1166, 63 L. Ed. 2d 427 (1980); Van Gemert v. Boeing Co., 553 F.2d 812, 815 (2d Cir.1977); Samuel v. University of Pittsburgh, 538 F.2d 991, 997, 999 (3d Cir.1976); Partain v. First Nat'l Bank, 59 F.R.D. 56 (M.D.Ala. 1973). Even when an aggregate method is used,

the defendant has full jury trial rights to contest ... proofs of aggregate assessment of monetary relief by the class. If such aggregate assessments are otherwise proper in class actions, the Constitution does not guarantee the defendants a jury trial to contest each class member's damage claim just as if the defendant were involved in a traditional, nonclass suit.

Id. at 355. Defendant has no "right" to an individualized determination of damages for each plaintiff; the desire for accuracy must be balanced against other factors such as the burdens of discovery in relation to the size of the individual claims.

Although there exist some cases which provide superficial support for defendant, those cases present fundamentally different circumstances which made classwide determination of damages inappropriate. For instance, in In re Fibreboard Corp., 893 F.2d 706 (5th Cir.1990), the district court had consolidated for trial approximately 3,000 asbestos cases and had decided that classwide damages would be determined on the basis of full trials for eleven plaintiffs and such evidence as the parties wished to submit from 30 other illustrative plaintiffs, fifteen of which would be chosen by each side. The jury was to decide the percentage of plaintiffs exposed to each defendant's products and the percentage whose claims were barred by various defenses, and then compute total damages based on the evidence concerning the 41 plaintiffs and expert testimony concerning the computation of classwide damages.

On appeal, the court granted the defendants' petition for a writ of mandamus, holding that the method proposed by the district court would violate Texas law, which governed the actions. The court emphasized that, at least in the absence of proper class certification, Texas law required individualized proof of damages. 893 F.2d at 711-12. Fibreboard differs from this case in that the court expressly noted that individual issues were sufficiently predominant that class certification was not available. Id. at 712. Those issues included differences as to the nature of the diseases suffered by the plaintiffs, differences in the manner and extent to which they were exposed to asbestos, and differences in the likelihood that the illnesses were caused by asbestos. Id. Here, in contrast, the class members were all subjected to exactly the same conduct, and the *1328 individual issues concerning damages do not predominate. For that reason, the Court has already held class certification to be appropriate. Because it is a class action, individualized proof of damages is not mandated.^[8]

Another case which provides superficial support for defendant's position is Windham v. American Brands, Inc., 565 F.2d 59 (4th Cir.1977), cert. denied, 435 U.S. 968, 98 S. Ct. 1605, 56 L. Ed. 2d 58 (1978). In Windham, the court held that the trial court did not abuse its discretion in denying class certification in an antitrust action brought by tobacco growers against tobacco companies on the basis that the damages issues were too complex to be consolidated in one case. The court noted that damages were not amenable to determination from a classwide formula, and that they would require separate mini-trials which could consume ten years of trial time. 565 F.2d at 68-70. The court also emphasized that at least some of the plaintiffs were financially able and willing to continue their claims in the absence of class certification in light of the size of the possible recovery. Id. at 69. In this case, although there are a number of variables which any model of damages must take into account, the case is not so complex as to make a model unfeasible. Furthermore, the individual claims are too small to realistically justify individual lawsuits.

Defendant cites various other cases concerning individual determination of issues such as mitigation of damages, but in none of those cases was a court asked to allow a classwide determination of damages based on extrapolation from a representative sample. Those cases, therefore, provide little guidance in the present context. For instance, in Int'l Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S. Ct. 1843, 52 L. Ed. 2d 396 (1977), the Court stated, "[w]hen the Government seeks individual relief for the victims of the discriminatory practice, a district court must usually conduct additional proceedings after the liability phase of the trial to determine the scope of individual relief." 431 U.S. at 361, 97 S. Ct. at 1868. The Court also stated that "the Government must carry its burden of proof, with respect to each specific individual, at the remedial hearings to be conducted by the District Court on remand." Id. at 371, 97 S. Ct. at 1873. Courts have rejected the argument that this language indicates a universal requirement of individual damage hearings:

Though Teamsters certainly raises a presumption in favor of individualized hearings, the case should not be read as an unyielding limit on a court's equitable power to fashion effective relief for proven discrimination. The language of Teamsters is not inflexible; after stating that individual hearings are "usually" required, Teamsters, supra, 431 U.S. at 361, 97 S. Ct. at 1867 [sic: 1868], the Court went on to note that "[i]n determining the specific remedies to be afforded, a district court is `to fashion such relief as the particular circumstances of a case may require to effect restitution.'" Id. at 364, 97 S. Ct. at 1869, quoting Franks [v. Bowman Transportation Co., 424 U.S. 747, 764, 96 S. Ct. 1251, 1264[, 47 L. Ed. 2d 444] (1976).] Later courts have often faced situations in which the Teamsters hearing preference had to blend to accommodate Title VII's remedial purposes.

Segar, 738 F.2d at 1289-90.

Defendant also cites N.L.R.B. v. Madison Courier, Inc., 472 F.2d 1307 (D.C.Cir. 1972):

Where an employer with a back pay liability contends that the discriminatees in question did not all make the required effort to mitigate their damages, the willful idleness issue must "be determined with respect to each employee considering the record as a whole. N.L.R.B. v. Rice Lake Creamery Co., 124 U.S.App. *1329 D.C. 355, 361, 365 F.2d 888, 894 (1966) (emphasis supplied). The particular facts concerning each separate individual must be considered by the Board. See United States Air Conditioning Corp., 141 NLRB 1278, 1280 (1963), enfd., 336 F.2d 275 (6th Cir.1964). This individualized, rather than group, approach is dictated by the nature of the mitigation rule which is generally recognized today. The merit of an individualized approach in a judicial matter is self-evident.

472 F.2d at 1318. Madison Courier, however, arose in a context very different from that of the present case. While this case is a class action involving some 3,000 plaintiffs, Madison Courier involved only thirteen claimants. Furthermore, the agency in Madison Courier had treated all of the claimants the same. Id. at 1322. Here, no one is suggesting that all 3,000 class members should be treated identically; rather, the suggestion is that sampling will enable an approximate determination of the circumstances of the individual class members.

Defendant further cites United States v. Lee Way Corp., 625 F.2d 918 (10th Cir. 1979), in which damages were awarded for a violation of Title VII. The district court had determined mitigation issues on an individual basis. On appeal, the court rejected the defendant's contention that a "best man" or "average man" approach should have been used. Under the former, every claimant would be held to a standard of mitigation determined by the claimant who gained the most interim earnings. Under the latter, every claimant would be held to a standard of mitigation determined by the average of the claimants' interim earnings. The court held that individual determination of mitigation was more equitable, and that the district court had not abused its discretion. In this case, in contrast, no presumptions would be applied concerning reasonable mitigation. Indeed, a sampling method would further, rather than contravene, the Lee Way court's interest in allowing for differences among claimants by permitting an efficient approximation of those differences.

The Court finds that the appropriateness of limiting discovery to a representative sample should be determined by weighing the relative benefits and burdens of such a method in comparison with full discovery. In this case, the burdens of full discovery are substantial,^[9] especially in light of the modest size of the individual claims involved. Those burdens can be reduced considerably by limiting discovery to a representative sample of class members.

The benefits of full discovery, in contrast, are minimal. Although defendant argues at length about the existence of individual issues, it has presented no argument that, because of those issues, extrapolation from a representative sample is less likely than is full discovery to produce an accurate reflection of total damages.^[10] Through the use of a proper model and an appropriate sample size, any such deviation should be minimized. Furthermore, there is no way of knowing whether the total amount of damages assessed on the basis of such extrapolation would be higher or lower than the amount assessed based on full discovery and individual hearings, and it would seem that the likelihood that extrapolation would produce a lower total for damages is the same as the likelihood that it would produce a higher total. In the absence of any evidence that defendant *1330 would fare worse under a sampling method, the Court cannot justify the expense^[11] of full discovery in this case. There simply is no substantial benefit to full discovery that offsets the added burdens. This is particularly true in light of the existence of some 2,000 questionnaires which class members have already completed in other litigation and which provide evidence relevant to this case.^[12]

Accordingly, the Court will enter a protective order pursuant to Rule 26 limiting discovery of information from plaintiffs to a representative sample of class members.^[13] The court shall first require expert testimony, however, as to the sample size necessary in order to allow a reasonably accurate determination. Such testimony may be provided initially by way of affidavits, to be submitted by April 24, 1991. However, pursuant to local rule 12(k), the parties are directed to consult first in an attempt to agree on the size of the sample and the method of its selection.^[14]

The majority of the IFFA questionnaires are in the possession of IFFA's counsel in the Kansas City case. Although TWA has not yet subpoenaed the questionnaires, a dispute has arisen in this case concerning defendant's entitlement to those questionnaires. Defendant has indicated an intent to seek the questionnaires, and plaintiffs have stated that they will move for a protective order. Both parties addressed the merits of the issue in a hearing before the Court on January 10, 1991, and it is therefore ripe for resolution. *1331 Plaintiffs argue that discovery of the questionnaires should be limited to a random sample just like other discovery from class members. Defendant argues that discovery of the questionnaires is different from other discovery because production of the existing questionnaires would pose no burden to anyone other than defendant itself.

The Court agrees with defendant that it should be allowed to review all of the questionnaires. The questionnaires clearly contain information which is relevant to the damages issue, and because the information has already been compiled, there is no substantial burden imposed by ordering their production.

Plaintiffs also object that defendants will try to use the questionnaires to impeach the evidence gained from the random sample. The Court does not view this concern as warranting preclusion of the discovery. Just as the Court has held that the class need not submit all of its members for discovery and may present its damages claim on the basis of a random sample, the Court also holds that defendant is entitled to use relevant and available evidence to oppose plaintiffs' damages claim. In the abstract, it would seem counterproductive for defendant to burden the factfinder with extensive particularized evidence from the questionnaires. However, if the questionnaires indicate a flaw in plaintiffs' theory of damages, defendant should not be precluded from gathering or introducing evidence of that flaw in response to plaintiffs' proof of damages. Accordingly, the Court denies plaintiffs' motion for a protective order limiting defendant's access to the IFFA questionnaires.^[15]

Plaintiffs argue that defendant, because of its conduct, should be estopped from arguing that plaintiffs have failed to mitigate damages. Plaintiffs contend that they should not be penalized for any failure to mitigate damages because defendant improperly discouraged them from taking actions to find comparable employment.^[16]See Walaschek & Associates v. Crow, 733 *1332 F.2d 51, 54 (7th Cir.1984) ("Estoppel arises when one party's conduct leads another to believe that a right will not be enforced and causes the one so misled to act to his detriment in reliance upon his beliefs."). See also Phelps v. Federal Emergency Management Agency, 785 F.2d 13, 16 (1st Cir.1986) (but for the fact that suit was against the government, estoppel would be available based on representations to insured that no written proof of loss need be filed; "[e]quitable estoppel is a judicially-devised doctrine which precludes a party to a lawsuit, because of some improper conduct on that party's part, from asserting a claim or defense, regardless of its substantive validity").

Plaintiffs base this argument on three acts by defendant. First, they point to a written final settlement proposal which defendant issued on May 14, 1986. In this proposal, defendant stated that anyone "accepting employment within the field of aviation without the prior written consent of TWA shall be deemed ineligible for rehire and shall be removed from the Preferential Rehire List." Plaintiffs argue that this sentence was contrary to the provisions of the Airline Deregulation Act^[17] and that it improperly discouraged plaintiffs from seeking jobs with other air carriers. Second, plaintiffs rely on defendant's refusal to issue the designated rights letter. As this Court has already held, with the Seventh Circuit's subsequent agreement, this refusal violated the Airline Deregulation Act. Third, plaintiffs emphasize defendant's conduct when it finally, during the pendency of this lawsuit, issued the designated rights letter. In the cover letter, dated February 13, 1987, defendant stated: "The contents of this letter and its enclosure should not be viewed as reflecting TWA's position as to your employment status in any context other than that relating to the airline employee protection program."

Defendant argues that plaintiffs have failed to establish the three prerequisites of equitable estoppel (1) a misrepresentation (2) on which plaintiff reasonably relied (3) to plaintiff's detriment. See Matter of Larson, 862 F.2d 112, 115 (7th Cir. 1988). Plaintiffs respond that this three-part test for estoppel is not relevant here, that they seek estoppel not in the traditional sense but rather in the sense that a claimant's obligation of diligence or mitigation may be suspended by virtue of a defendant's conduct. Plaintiffs, however, have provided no authority for the proposition that the usual estoppel principles are suspended in this context. Furthermore, estoppel is generally a question of fact, and a court may find estoppel as a matter of law only if the facts allow just one reasonable conclusion. Sawyer v. County of Sonoma, 719 F.2d 1001, 1006 n. 12 (9th Cir. 1983); Travelers Indemnity Co. v. Swanson, 662 F.2d 1098, 1102 n. 2 (5th Cir.1981); Ehret Co. v. Eaton, Yale & Towne, Inc., 523 F.2d 280, 283 (7th Cir.1975), cert. denied, 425 U.S. 943, 96 S. Ct. 1683, 48 L. Ed. 2d 186 (1976).

Here, the Court cannot conclude as a matter of law that defendants are estopped from arguing mitigation. Certainly the arguments made by plaintiffs are relevant, and if a class member failed to take steps to mitigate damages, it would be relevant to know whether a reason for that failure was a belief, based on improper statements by defendant, that efforts to secure comparable employment would be fruitless or counterproductive.^[18] At this point in the proceedings, plaintiffs have not shown an entitlement to estoppel as a matter of law. See Premier Electrical Construction Co. v. U.S. Fidelity & Guaranty Co., No. 80 C 6689, 1987 WL 12936 (N.D.Ill. June 19, *1333 1987) (rejecting argument that defendant should be estopped as a matter of law from invoking mitigation defense and leaving issue for trial).

Discovery of absent class members shall proceed only on a random sample basis. The parties are directed to submit memoranda concerning the size and determination of a random sample by April 24, 1991. The Court denies plaintiffs' motion for a protective order limiting defendant's access to the IFFA questionnaires. Plaintiffs' request that defendant be precluded from presenting a mitigation defense also is denied.

[1] See Independent Federation of Flight Attendants v. Trans World Airlines, Inc., 126 F.R.D. 560 (W.D.Mo.1989) (union has standing to seek backpay for individual employees).

[2] The proposed interrogatories include twelve questions concerning background information, a multi-part question concerning job search activity, a multi-part question concerning other employment, a multi-part question concerning use of the designated rights letter which was eventually sent to class members, a multi-part question concerning knowledge of job openings at other air carriers, a question concerning periods of inability to work, two questions concerning job-related expenses, a question concerning medical insurance, and two questions concerning unemployment compensation.

The document requests seek production of documents reviewed in preparing answers to the interrogatories, documents relating to job searches, copies of income tax returns, documents relating to employment income, documents relating to the Airline Deregulation Act and related regulations, and documents relating to reinstatement, recall, or seniority rights of any permanently replaced TWA striker.

[3] Plaintiffs summarize their position as follows: "Plaintiffs believe that to a large extent the amount due the class can be determined in a single classwide jury proceeding. In preparation for the hearing, a random sampling of the class will be generated, with defendant permitted to obtain through reasonable discovery the relevant facts as to sample members' experiences during the class period. Using this data, the parties' experts can then project what the class has lost in the way of wages and benefits, offset by amounts that were actually earned or through the exercise of reasonable diligence could have been earned. Plaintiffs also believe that the jury in the omnibus hearing can award, if it deems appropriate, a classwide amount for intangible injuries such as pain and suffering. Incidental and consequential losses (e.g., job search costs and relocation expenses) can then be determined using either a model or an affidavit process, in which each class member submits his or her costs, subject to audit and challenge by defendant as it deems appropriate." (Class' Submission Regarding Damages Determination at 1-2.)

[4] See also Chicago Miniature, 668 F. Supp. 1150, 1153-55 (N.D.Ill.1987) (classwide award would be reduced by a pro rata share of each individual's earnings).

[5] As defendant correctly notes, absent class members are not immune from discovery simply by virtue of being unnamed plaintiffs. See Brennan v. Midwestern United Life Ins. Co., 450 F.2d 999, 1004 (7th Cir.1971) ("absent class members may, under certain circumstances, be required to submit to discovery"), cert. denied, 405 U.S. 921, 92 S. Ct. 957, 30 L. Ed. 2d 792 (1972).

[6] Defendant argues (TWA's Submission Regarding Damages, filed Mar. 23, 1989, at 3-4) that plaintiffs have skipped the step of proximate cause by assuming that class members were in fact injured by defendant's failure to provide designated rights letters. If defendant is expressing a concern that perhaps the class members were not actually injured, this concern relates to the existence of damages and shall be litigated as part of the other damages issues. If defendant is stating that any injuries which the class members suffered may have been caused by the class members' failure to apply for other positions rather than by defendant's conduct, that issue is covered by defendant's mitigation defense. If defendant is stating that other air carriers would not have hired plaintiffs even if the designated rights letters had been issued, the Court has reservations about the legitimacy of this defense. As plaintiffs point out, the other air carriers were statutorily obligated to honor the letters. However, the Court shall defer ultimate resolution of this issue until the parties have completed discovery of other air carriers.

[7] See supra at 1323-1324.

[8] See supra at 1323-1325. Another flaw in the procedure proposed by the district court in Fibreboard was the manner in which the illustrative plaintiffs were chosen, with half to be selected by each side. 893 F.2d at 709. If classwide damages are to be determined on the basis of a sample, the sample should be selected at random in order to improve the accuracy of the final award.

[9] The Court also must consider the substantial burden on the judicial system of requiring 3,000 separate mini-trials a burden which defendant emphasized when it opposed class certification on the basis that individual damages issues would make a class action unmanageable.

[10] "Aggregate proof of the defendant's monetary liability is no more unfair than class treatment of other elements of liability. Under these circumstances, there is no room for the defendant to complain of any litigation unfairness by use of aggregate proofs of the defendant's monetary liability to the class. On the contrary, it is not unusual, and probably more common than otherwise in many types of cases, that aggregate evidence of the defendant's liability is more accurate and precise than would be so with individual proofs of loss." Newberg at 349. Here, certain factors relevant to damages may be analyzed even more accurately on a classwide basis than on an individual basis, such as the number of openings that were available at other airlines.

[11] Indeed, the Court is somewhat surprised that defendant has apparently not considered the expense of its own attorneys' fees in conducting full discovery an expense which could alone be larger than the monetary benefit to defendant of conducting full discovery as opposed to relying on a representative sample.

[12] See infra at 1330-1331.

[13] The parties disagree as to whether discovery is appropriate as to consequential damages, including costs occurred in seeking alternative employment, and non-economic damages such as mental suffering. Plaintiff argues that these damages are generally available in contract and employment cases and should therefore be available in this case as well. See Webb v. City of Chester, 813 F.2d 824, 836 (7th Cir.1987) (damages for embarrassment and humiliation available in § 1983 employment discrimination case); Muldrew v. Anheuser-Busch, Inc., 728 F.2d 989, 992-93 (8th Cir.1984) (damages for mental and emotional distress available in § 1981 employment discrimination case); Carter v. Duncan-Huggins, Ltd., 727 F.2d 1225, 1238 (D.C.Cir.1984) (damages for humiliation available in § 1981 employment discrimination case); McGee v. South Pemiscot School District, 712 F.2d 339, 344 (8th Cir.1983) (consequential damages and damages for mental anguish and loss of reputation available where plaintiff fired in violation of First Amendment). Plaintiffs suggest that these damages can be determined on the basis of affidavits submitted by each class member.

Defendant responds that in the employment context, such damages are not allowed. See Smith v. Office of Personnel Management, 778 F.2d 258 (5th Cir.1985) (Age Discrimination in Employment Act), cert. denied, 476 U.S. 1105, 106 S. Ct. 1949, 90 L. Ed. 2d 358 (1986); Graves Trucking, Inc. v. N.L.R.B., 692 F.2d 470, 475 (7th Cir.1982) (National Labor Relations Act); Shuttleworth v. Broward County, 649 F. Supp. 35 (S.D.Fla.1986) (Rehabilitation Act). As plaintiff notes, however, the cases cited by defendants arise in statutory contexts where reinstatement is an available remedy. They also rest on specific statutory language allowing equitable relief and, by implication, disallowing compensatory relief. See also Patzer v. Board of Regents of University of Wisconsin, 763 F.2d 851, 854 n. 2 (7th Cir.1985); Muldrew, 728 F.2d at 992 n. 2; Walker v. Ford Motor Co., 684 F.2d 1355, 1364 (11th Cir.1982). Here, where a private cause of action has been judicially implied, Long, 913 F.2d at 1267, and where the statute thus reflects no intent to limit the nature of damages, the Court shall look to principles of general tort and employment law for guidance as to available damages rather than looking to statutory schemes which specify certain types of remedies. Accordingly, the Court finds that plaintiffs are entitled to consequential and non-economic compensatory damages in addition to back pay.

[14] Although this opinion allows plaintiffs to limit the discovery served on class members, plaintiffs retain their usual burden of proof with respect to damages. It is they who shall choose what evidence to submit, and it is they who shall bear the risks of a sampling method which does not adequately prove the damages they seek. See Newberg at 349 ("Aggregate damage proofs are valid and proper only when proved and determined by the same standards by which the propriety of individual awards are decided. In all cases, the proofs must be sufficiently reliable to determine the defendant's liability justly.").

[15] The Court finds it necessary for the record to correct a misstatement in defendant's brief on this issue. During a settlement conference, the Court indicated that it was going to proceed to decide the pending discovery issue, but it wanted submissions from the parties first concerning the effect of the questionnaires on discovery in this case. Defendant apparently misunderstood the Court's request, stating:

At a meeting on December 7, 1990, TWA protested to this Court that it would be unfair for the Court to make any rulings on individual discovery until TWA had an opportunity to review the IFFA Questionnaires already completed by over 2,000 of the class members. In response to TWA's protest, this Court ordered the parties to file Briefs on the information obtainable from the IFFA Questionnaires that would be relevant to this case.

(Defendant's Brief on Information, filed Dec. 14, 1990, at 2.) The Court did not request the briefs in response to any concern expressed by defendant that it needed an opportunity to review the questionnaires; it requested briefs in order that the Court would have before it the information necessary in order to resolve the pending discovery issues.

[16] The Court agrees with plaintiffs that they were not required to mitigate damages by accepting any employment they could find, but rather by finding substantially equivalent employment. See Ford Motor Co. v. E.E.O.C., 458 U.S. 219, 231-32, 102 S. Ct. 3057, 3065-66, 73 L. Ed. 2d 721 (1982) (in Title VII case, "[a]lthough the unemployed or underemployed claimant need not go into another line of work, accept a demotion, or take a demeaning position, he forfeits his right to backpay if he refuses a job substantially equivalent to the one he was denied"); Hybert v. Hearst Corp., 900 F.2d 1050, 1054 n. 7 (7th Cir.1990) (in ADEA case, mitigation involves finding "similar," i.e., "substantially equivalent" employment). To hold otherwise would be to essentially reward defendant for its wrongdoing.

In support of a broader standard, defendant argues that the Court should look to the Employee Protection Program enacted as part of the Airline Deregulation Act. 49 U.S.C.App. § 1552. This program (which ultimately was not funded) provided for monthly assistance payments from the government to workers who lost jobs as a result of deregulation. The section provides for reduction of payments to an employee who does not accept "reasonably comparable" employment which has been offered, 49 U.S.C.App. § 1552(b) (2) a standard which defendant argues is broader than the "substantially equivalent" standard. The Court cannot agree that it should look to the Employee Protection Program for mitigation principles to apply in this case the government often sets higher standards for recovery from the public fisc than those applicable to private wrongdoers.

[17] See 49 U.S.C.App. § 1552(d) (1); 49 C.F.R. § 220.11(b).

[18] Plaintiffs do represent that they are prepared to make such an argument at the trial on damages: "[B]ecause defendant's conduct chilled many class members from applying by threatening loss of recall rights if successful, the applicant rate does not fully delimit the extent of the injury. The class thus anticipates that its experts will attempt to adjust their models for this chilling factor, with defendant of course free to present counter evidence." (Reply Mem. Regarding Damages Claim, filed Apr. 6, 1989, at 7 n. 4.)