Shannon v. Pay'N Save Corp.

104 Wn.2d 722 (1985)

709 P.2d 799

IRIS SHANNON, Respondent, v. PAY 'N SAVE CORPORATION, Appellant.

No. 51117-9.

The Supreme Court of Washington, En Banc.

November 21, 1985.

Ryan, Swanson, Hendel & Cleveland, by Michael R. Rayton and Thao Tiedt, for appellant.

Salter, McKeehen & Rabine, by James R.B. Salter and Carleton H.A. Taber, for respondent.

DURHAM, J.

Pay 'n Save Corporation appeals from a judgment entered upon a jury verdict finding it liable, pursuant to RCW 49.60, for discriminating against Iris Shannon in employment because of her gender. Pay 'n Save contends that the trial court erred by: (1) submitting Shannon's theory of disparate impact to the jury, (2) allowing Shannon to introduce statistical evidence at trial, (3) refusing to find that a juror's statements constituted misconduct requiring a new trial, and (4) awarding Shannon $40,267 in attorney fees. Alternatively, Pay 'n Save argues that even if the trial court was justified in submitting the disparate impact theory to the jury, it misstated the theory in its jury instructions. Shannon cross-appeals contending that the trial court erred by denying her the injunctive relief of *724 reinstatement and by refusing to allow her to recover expert witness fees. We agree with Pay 'n Save that the trial court's instructions misstated the law of disparate impact, and we reverse and remand for a new trial.

In 1966, Shannon began working as a receiving clerk for Ernst Hardware, a division of Pay 'n Save. Eight years later, she became the first woman promoted to the position of assistant manager at an Ernst store.[1] Between 1974 and 1979, Shannon served as assistant manager at four different Ernst stores. During her tenure as assistant manager, Shannon was evaluated annually by her store manager. Beginning in 1976, Ernst's store managers utilized a standardized form for evaluating assistant managers. Managers rated assistant managers on a scale from one to eight for qualities perceived by Ernst to be relevant to an employee's suitability for promotion. Those qualities evaluated ranged from involvement in community activities to proficiency in determining when to cut off inventory. In 1976, 1977 and 1978, Shannon received high ratings. In 1979, however, her rating plummeted and Shannon was informed by her district manager that she would never become a store manager. Upset by this evaluation, Shannon requested a transfer to another position. For lack of another opening close to her home, Shannon accepted a position as a receiving clerk.

The standardized evaluation form used to evaluate Shannon was developed pursuant to a consent decree entered into by Pay 'n Save in 1976. The consent decree evolved out of litigation instituted by a class of female Pay 'n Save employees who claimed that Pay 'n Save's promotion policies discriminated against them in violation of Title VII of the Civil Rights Act of 1964. The decree required Pay 'n Save to develop formal criteria for promotion to the department head category. Accordingly, Pay 'n *725 Save developed a standardized employee evaluation form which it uses not only to evaluate an employee's suitability to become department head but also to evaluate his or her suitability to become store manager or assistant manager.

Pay 'n Save, however, does not rely solely on the evaluation form in determining promotions to the store manager level. After the evaluation form stage, the store manager meets with a district and a division manager to discuss the evaluation. All district managers then confer, revise the store manager's evaluations, and from these revised evaluations, identify candidates for promotion.

In 1980, Iris Shannon filed the action that is the subject of this appeal, claiming that Pay 'n Save violated RCW 49.60[2] by refusing to promote her to the position of store manager because of her gender. At trial, Shannon employed an expert in statistics to testify. The expert introduced statistical evidence to support Shannon's contention that there is little correlation between the result of a store manager's evaluation and an offer of promotion. The expert maintained that after the consent decree, Pay 'n Save no longer discriminated against women at the department head level, but had raised the plateau and now discriminated against women at the assistant and store manager level.

Although it did not employ an expert, Pay 'n Save introduced evidence that Shannon could not get along with subordinates, was dictatorial in her managerial style, and had to be suspended for a week in 1977 because of misconduct. In short, Pay 'n Save maintained that Shannon was not *726 suitable to be a store manager.

The jury returned a verdict which awarded Shannon $20,000 in damages. The trial court then granted her $40,267 in attorney fees and $1,000 in punitive damages.

Both Pay 'n Save and Shannon made post-trial motions. Pay 'n Save moved for a new trial based upon juror misconduct. In Shannon's post-trial motion, she moved to amend her pleading to include the equitable relief of reinstatement. The trial court ordered Pay 'n Save to consider Shannon for promotion, but refused to order reinstatement. The court also denied Pay 'n Save's motion for a new trial. From the judgment entered by the trial court, Pay 'n Save appeals and Shannon cross-appeals.

Prior to evaluating Pay 'n Save's contentions on appeal, we first review some well established principles of employment discrimination law.

Two theories, "disparate impact" and "disparate treatment," are available to a plaintiff attempting to prove discrimination in employment pursuant to RCW 49.60. Fahn v. Cowlitz Cy., 93 Wn.2d 368, 378, 610 P.2d 857, 621 P.2d 1293 (1980); see also Furnco Constr. Corp. v. Waters, 438 U.S 567, 57 L. Ed. 2d 957, 98 S. Ct. 2943 (1978). "`Disparate treatment' ... is the most easily understood type of discrimination. The employer simply treats some people less favorably than others because of their race, color, religion, sex, or national origin." International Bhd. of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 52 L. Ed. 2d 396, 97 S. Ct. 1843 (1977). To establish a prima facie case of employment discrimination under a "disparate treatment" theory, the plaintiff must show that: (1) he belongs to a protected class, (2) he applied and was qualified for a job for which the employer was seeking applicants, (3) he was rejected for the position, and (4) after his rejection the position remained open and the employer continued to seek qualified applicants. McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802, 36 L. Ed. 2d 668, 93 S. Ct. 1817 (1973).

[T]he prima facie case "raises an inference of discrimination only because we presume these acts, if otherwise *727 unexplained, are more likely than not based on the consideration of impermissible factors."

Texas Dep't of Comm'ty Affairs v. Burdine, 450 U.S. 248, 254, 67 L. Ed. 2d 207, 101 S. Ct. 1089 (1981) (quoting Furnco, 438 U.S. at 577).

Once the plaintiff establishes a prima facie case, the defendant must produce some evidence that the rejection of the plaintiff's application was due to a legitimate nondiscriminatory reason. Texas Dep't of Comm'ty Affairs v. Burdine, supra. The plaintiff then must be afforded the opportunity to show that the defendant's asserted justification was a mere pretext for discrimination. McDonnell Douglas Corp., 411 U.S. at 804. The plaintiff in a disparate treatment case retains the burden of proof throughout the trial. United States Postal Serv. Bd. of Governors v. Aikens, 460 U.S. 711, 75 L. Ed. 2d 403, 103 S. Ct. 1478 (1983). He succeeds in meeting this burden by showing that a discriminatory reason more likely than not motivated the employer. Aikens, at 716.

Unlike disparate treatment, the disparate impact theory enables a plaintiff to address the consequences of seemingly objective employment practices by allowing the plaintiff to prevail in an employment discrimination suit without establishing discriminatory motive. Griggs v. Duke Power Co., 401 U.S. 424, 432, 28 L. Ed. 2d 158, 91 S. Ct. 849 (1971). Thus, a disparate impact analysis requires a plaintiff to prove: (1) a facially neutral employment practice, (2) falls more harshly on a protected class. International Bhd. of Teamsters, 431 U.S. at 336. Once the plaintiff establishes the prima facie case, the burden shifts to the defendant to show that the challenged requirement has a "manifest relationship" to the position in question. Griggs, at 432. If the employer meets this burden, the plaintiff may still prevail by showing that other less discriminatory alternatives can equally serve the employer's legitimate business requirements. Albemarle Paper Co. v. Moody, 422 U.S. 405, 425, 45 L. Ed. 2d 280, 95 S. Ct. 2362 (1975); Fahn v. Cowlitz Cy., 93 Wn.2d 368, 380, 610 P.2d 857, 621 P.2d 1293 (1980).

*728 Pay 'n Save maintains that the trial court misstated the law of disparate impact by presenting instructions 17 and 18 to the jury. We agree. Instruction 17 states:

In determining whether plaintiff has met her burden of proof of showing that defendant's promotional procedures disparately impact women Assistant Managers and favors male Assistant Managers for promotion to Store Manager, you are instructed that if the selection rate for women to Store Manager is less than four-fifths (4/5) or eighty percent (80%) of the rate for men, it is evidence that women are disparately impacted because of sex unless defendant can prove that promotional procedures are job related.

(Italics ours.)

[1] This instruction is based on 29 C.F.R. § 1607.4(D) (1984), an Equal Employment Opportunity Commission (EEOC) Guideline common referred to as the "four-fifths rule." Although EEOC Guidelines are generally entitled to deference, we refuse to incorporate this rule as a substantive part of discrimination law in Washington for two independent reasons.

First, the presence of discrimination is ultimately a factual question. See Page v. U.S. Indus., Inc., 726 F.2d 1038, 1044 (5th Cir.1984); Pullman-Standard v. Swint, 456 U.S. 273, 72 L. Ed. 2d 66, 102 S. Ct. 1781 (1982).

The right of trial by jury entitles every party to the judgment of the jury as to the ultimate facts upon which liability rests, and it is essential that the jury be left perfectly free to form and declare from the evidence their opinion upon the questions of fact.

(Italics ours.) Webb v. Seattle, 22 Wn.2d 596, 611, 157 P.2d 312, 158 A.L.R. 810 (1945). Instead of freeing the jury to determine whether discrimination exists, the four-fifths rule binds it to a rigid mechanistic formulation. In addition, the formulation itself has been criticized for having neither a substantive nor statistical basis. Cormier v. P.P.G. Indus., Inc., 519 F. Supp. 211, 255 (W.D. La. 1981), aff'd, 702 F.2d 567 (5th Cir.1983). Indeed, the EEOC clearly states that the four-fifths rule is "not a legal definition of discrimination, *729 rather it is a practical device to keep the attention of enforcement agencies on serious discrepancies ..." Jones v. First Fed. Sav. & Loan Ass'n, 546 F. Supp. 762, 778 (M.D.N.C. 1982) (quoting 43 Fed. Reg. 38,291 (1978)). We refuse to elevate a guideline, the purpose of which is to ensure the uniform exercise of prosecutorial discretion, into a rule of law which would change the determination of discrimination into an arbitrary numbers game.

Second, the four-fifths rule is also objectionable because it focuses the jury's attention on one specific type of statistical comparison. An instruction based on the rule, therefore, violates what

has, for some years, been the policy of our Washington system of jurisprudence, in regard to the instruction of juries, to avoid instructions which emphasize certain aspects of the case and which might subject the trial judge to the charge of commenting on the evidence ...

Laudermilk v. Carpenter, 78 Wn.2d 92, 100, 457 P.2d 1004, 469 P.2d 547 (1969). Accordingly, we conclude that the trial court erred by giving instruction 17.

Pay 'n Save also contends that the trial court misstated the law of disparate impact in jury instruction 18.[3] Instruction 18 defines the burden a defendant must bear once a plaintiff establishes a prima facie case of disparate impact. Although we have never specifically addressed the extent of a defendant's burden in a disparate impact case, we are not without guidance. Federal cases decided under Title VII provide precedent in construing RCW 49.60. Stieler v. Spokane Sch. Dist. 81, 88 Wn.2d 68, 73, 558 P.2d 198 (1977). However, as we noted in Fahn v. Cowlitz Cy., at *730 380-81, different federal circuits have placed varying burdens on employers in establishing a business necessity defense. Here, the instruction utilized by the trial court is derived from Robinson v. Lorillard Corp., 444 F.2d 791, 798 (4th Cir.), cert. dismissed, 404 U.S. 1006 (1971). Essentially, the Lorillard instruction requires the employer to prove three elements to establish a business necessity defense: (1) the employment or hiring practice is necessary to the safe and efficient operation of the business, (2) the practice carries out the alleged business purpose, and (3) no less discriminatory alternatives exist.

The first part of the Lorillard test requires an employer to do more than establish a direct correlation between the test and job performance. The employer must show that the test is necessary to his business. Thus, if the results of the test do not create a sexually and racially balanced work force, the employer must not utilize the test unless it significantly affects a core purpose of the business.

[2] In Contreras v. Los Angeles, 656 F.2d 1267 (9th Cir.1981), the court refused to place such a stringent burden on an employer. The court reasoned that "[d]iscriminatory preference for any group, minority or majority, is precisely and only what Congress [in enacting Title VII] has proscribed." Contreras, at 1278 (quoting Griggs v. Duke Power Co., 401 U.S. 424, 431, 28 L. Ed. 2d 158, 91 S. Ct. 849 (1971)). The court held that if an employer can demonstrate by professionally accepted methods that the employment or hiring practice accurately predicts or significantly correlates with a candidate's job performance, the employer has established a business necessity defense. Contreras, at 1280. To require more of an employer would go beyond prohibiting discriminatory preference and would impose an affirmative duty to hire a balanced work force if an objective standard for evaluation is used. Contreras, at 1278. The Ninth Circuit found that the imposition of such a duty is not consistent with either congressional intent in passing, or Supreme Court cases in construing, Title VII. Contreras, at 1279-80.

*731 We agree with the Ninth Circuit's interpretation of the underlying purpose of the laws against discrimination in employment. We also agree that the Ninth Circuit's standard furthers the basic tenet of the disparate impact theory by reaffirming that discriminatory preference occurs when an objective measurement unrelated to job performance results in the underrepresentation of a protected class. See Griggs v. Duke Power Co., 401 U.S. at 431-34.

Furthermore, the Lorillard instruction has an additional flaw. The instruction apparently requires an employer to prove, in establishing his business necessity defense, that no less discriminatory alternatives to his present policies exist. The Supreme Court, however, has clearly indicated that the burden of proving less discriminatory alternatives falls on the plaintiff once the defendant establishes his defense. Albemarle Paper Co. v. Moody, 422 U.S. 405, 45 L. Ed. 2d 280, 95 S. Ct. 2362 (1975).

We, therefore, hold that to establish a business necessity defense an employer must prove by professionally accepted measures that the hiring or promotion test utilized accurately predicts or significantly correlates with the fundamental requirements of job performance. The burden then shifts to the plaintiff to show that other less discriminatory alternatives equally serve the employer's legitimate business requirements. Because instruction 18 misstated the burden an employer must meet and misallocated the burden of proof, the trial court erred by giving it.

Having concluded that the inaccurate statements of law contained in instructions 17 and 18 are dispositive of this appeal, we proceed to address only those issues that must be resolved to provide guidance to the trial court upon retrial.

Pay 'n Save contends that not only did the trial court misstate the theory of disparate impact, but the court erred by even presenting the theory to the jury. Pay 'n Save argues that because their evaluation process involves subjective appraisals of employee performance, the process cannot be considered a facially neutral practice. Thus, Pay *732 'n Save reasons that the disparate impact theory is inapplicable here.

[3] Pay 'n Save's employee evaluation process does call for the employee's immediate supervisor to make subjective judgments regarding the employee's performance. This process can be distinguished from the typical disparate impact case which involves a facially objective evaluation process. See, e.g., Griggs v. Duke Power Co., supra (intelligence tests); Albemarle Paper Co. v. Moody, supra (intelligence tests and high school diplomas); Dothard v. Rawlinson, 433 U.S. 321, 53 L. Ed. 2d 786, 97 S. Ct. 2720 (1977) (minimum height requirement). Nonetheless, a plaintiff is not precluded from proving a case of disparate impact simply because a portion of an employee evaluation process involves subjective judgments. Federal courts have consistently stated that either disparate impact or treatment may be applicable to the same set of facts. See, e.g., Page v. U.S. Indus., Inc., 726 F.2d 1038, 1045 (5th Cir.1984); International Bhd. of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 52 L. Ed. 2d 396, 97 S. Ct. 1843 (1977). The federal courts, however, do not agree on whether a solely subjective evaluation process can be analyzed under the disparate impact model. Not only are the circuit courts divided, but cases within circuit courts come to opposite conclusions. Compare Walls v. Mississippi Dep't of Pub. Welfare, 730 F.2d 306, 321 (5th Cir.1984) and Heagney v. UW, 642 F.2d 1157, 1163 (9th Cir.1981) (subjective hiring criteria cannot be analyzed as a facially neutral practice) with Page, at 1046 and Hung Ping Wang v. Hoffman, 694 F.2d 1146 (9th Cir.1982) (subjective criteria analyzed as a facially neutral practice). Those courts that allow a disparate impact theory to be applied to a subjective evaluation process reason that the theory can be applied whenever a plaintiff attacks a generalized employment policy. See, e.g., Peters v. Lieuallen, 693 F.2d 966, 969 (9th Cir.1982). We disagree with this approach.

The rationale underlying the development of the disparate impact model provides the basis for our resolution of *733 Pay 'n Save's contention. Because the traditional disparate treatment model requires proof of discriminatory motive, a plaintiff will have great difficulty, under that theory, attacking practices fair in form but discriminatory in operation. Thus, the disparate impact model was developed to enable a plaintiff to utilize other elements of proof in place of discriminatory motivation when attacking facially neutral practices that have discriminatory consequences. See Griggs, at 431. However, discriminatory motivation can be readily inferred when solely subjective employment evaluations result in the exclusion of a protected class. Thus, because motive can be inferred when solely subjective criteria are being attacked, the rationale for enabling a plaintiff to utilize elements of proof that exclude motive disappears. Pegues v. Mississippi Empl. Serv., 699 F.2d 760, 765 (5th Cir.), cert. denied, 464 U.S. 991, 78 L. Ed. 2d 679, 104 S. Ct. 482 (1983). We, therefore, conclude that to establish a prima facie case of disparate impact, a plaintiff must demonstrate that he is attacking an employment practice that includes objective features.

The establishment of a prima facie case of disparate impact is a question of proof. Thus, on retrial if Shannon can demonstrate that she is attacking features of an evaluation process that do not turn on discretionary decisions, she is entitled to a disparate impact instruction.

Pay 'n Save also argues that even if it utilizes objective criteria in its promotional policies, it should nonetheless be immune from a disparate impact claim. Pay 'n Save concedes that Shannon can argue that she, personally, was discriminated against because of her gender, but maintains that she should be precluded from arguing that Pay 'n Save policies discriminate against women generally. Pay 'n Save argues that because its employment policy was developed pursuant to a court approved consent decree, the policy should not be subject to attack. Thus, Pay 'n Save reasons that because its policy cannot be attacked, only a disparate treatment theory is applicable. We disagree.

Only those policies of Pay 'n Save that were actually *734 addressed in the consent decree are immune from attack. See United States v. International Bldg. Co., 345 U.S. 502, 506, 97 L. Ed. 1182, 73 S. Ct. 807 (1953). Paragraph 8 of the consent decree does contain general language that if viewed in isolation would support Pay 'n Save's contention:

This Decree fully and finally determines all issues raised in this consolidated action. The Company's undertakings and commitments in this Decree settle and satisfy all claims of sex discrimination in recruitment, hiring, training, assignment, transfer, promotion, job classification, termination, pay, benefits and/or terms and claims for damages, back-pay, punitive damages, benefits, injunctive, declaratory or other relief for past or present sex discrimination against named plaintiffs and members of the class. This Decree binds named plaintiffs and class members, other than those who have filed formal requests for exclusion pursuant to the court's Order of November 17, 1975.

However, the remedial provisions of the decree were not as inclusive as the above language would indicate. These provisions were narrowly tailored to address the underrepresentation of women in the department head category only. Paragraph 19 of the decree provides, in pertinent part:

In connection with the Ernst-Malmo Division the Company will substantially increase its efforts to place qualified female personnel in department head managerial positions, but in no event shall the Company make to qualified females job offers as described herein at a percentage less than 20% of all department head managerial positions (lumber, hardware, sporting goods and nursery department heads) made in the first year following the entry of this decree, 30% of such offers in the second year and 40% in the third, fourth and fifth years.

(Italics ours.) Thus, when read as a whole, the consent decree actually addressed discriminatory practices only at this one level, and Shannon can challenge Pay 'n Save's policies of promotion to other positions.

Pay 'n Save next contends that not only did the trial court err in its presentation of the disparate impact theory *735 to the jury, it also erred by allowing Shannon to introduce statistical evidence in support of her disparate treatment theory. Pay 'n Save cites Stieler v. Spokane Sch. Dist. 81, 88 Wn.2d 68, 558 P.2d 198 (1977) for the proposition that statistics are inadmissible under a disparate treatment theory. Stieler, however, does not support Pay 'n Save's contention.

[4] The Stieler court held that when an individual plaintiff presents a disparate treatment theory, he cannot establish a prima facie case with statistics the plaintiff must show he was qualified for the position he sought. Stieler, at 74. Here, Shannon did not attempt to substitute statistical evidence for an element of a prima facie case of disparate treatment. Instead, she introduced statistical evidence to prove that Pay 'n Save's assertion that it refused to promote her for legitimate nondiscriminatory reasons was a mere pretext for discrimination. It is settled law in the federal courts that in a disparate treatment case "[s]tatistics showing a general pattern of discrimination are probative on the question of whether the reasons given for a particular action are pretextual." Bauer v. Bailar, 647 F.2d 1037, 1045 (10th Cir.1981); see also Talley v. United States Postal Serv., 720 F.2d 505, 507 (8th Cir.1983), cert. denied, 466 U.S. 952, 80 L. Ed. 2d 541, 104 S. Ct. 2155 (1984); McDonnell Douglas Corp. v. Green, 411 U.S. 792, 805, 36 L. Ed. 2d 668, 93 S. Ct. 1817 (1973). Stieler does not prohibit the use of statistics in this context, and we agree with the federal courts that statistics are relevant to show that an employer's asserted justification for not hiring or promoting the plaintiff is a facade for discrimination.

Pay 'n Save next contends that the trial court erred by admitting statistical evidence of discrimination prior to 1976. Pay 'n Save contends that any discrimination prior to that year was remedied by the consent decree. The trial court agreed with Pay 'n Save's position, but admitted the evidence for illustrative purpose only. The court allowed Shannon to introduce the statistics to illustrate her theory that after 1976 Pay 'n Save simply elevated the plateau *736 above which women could not advance.

[5] Pay 'n Save maintains that evidence of discrimination prior to the consent decree is inadmissible because it is equivalent to evidence of discrimination prior to the passage of Title VII. Even if this contention is correct, the trial court did not err in admitting the evidence for illustrative purposes. Although it has no present legal consequences,

a discriminatory act which occurred before the statute was passed ... may constitute relevant background evidence in a proceeding in which the status of a current practice is at issue ...

United Air Lines, Inc. v. Evans, 431 U.S. 553, 558, 52 L. Ed. 2d 571, 97 S. Ct. 1885 (1977). Here, the trial court refused to admit the statistics to support a finding that Pay 'n Save discriminated prior to 1976. It did allow the statistics to be admitted solely for background purposes to support a claim that Pay 'n Save presently discriminated. In so ruling, the trial court acted properly.

[6] Because the resolution of the other assignments of error raised by Pay 'n Save are neither necessary to the disposition of this appeal nor needed to provide guidance to the trial court on retrial, we do not address them here. We must, however, address one issue raised by Shannon on cross appeal. Shannon contends that the trial court erred by finding that expert witness fees were not a "cost of suit" under RCW 49.60.030(2). We disagree. Washington courts have long held that "costs" under RCW 4.84.030 do not include compensation paid to the expert witness in excess of ordinary witness fees. See 2 L. Orland, Wash. Prac., Trial Practice § 422 (3d ed. 1972); Nelson v. Industrial Ins. Dep't, 104 Wash. 204, 176 P. 15 (1918). There is no dispositive distinction between "costs" as used in RCW 4.84.030 and "cost of suit" as used in RCW 49.60.030(2). Accordingly, a plaintiff cannot recover expert witness fees in excess of compensation paid to an ordinary witness.

Because the four-fifths rule and the business necessity defense instructions misstated the law of disparate impact, *737 we reverse the judgment and remand for a new trial.

DOLLIVER, C.J., DORE, CALLOW, and GOODLOE, JJ., and HAMILTON, J. Pro Tem., concur. UTTER, J. (concurring in part, dissenting in part)

It is inappropriate to rely on federal law which limits rights of plaintiffs when federal law, unlike Washington law, contains no directive to liberally construe the antidiscrimination statute.[4] The express directive of the Legislature should not be ignored. The majority's refusal to allow this plaintiff the opportunity to use a disparate impact analysis further limits rights of all plaintiffs and is inconsistent with the express legislative intent of RCW 49.60. Adoption by the majority of the "job related" definition least burdensome to the employer while, at the same time, placing the burden of proof with respect to less discriminatory alternatives to the employee is a policy decision also contrary to the explicit legislative intent of RCW 49.60. Moreover, it is inconsistent with our own decisions.

DISPARATE IMPACT

If selection criteria are subjective and a plaintiff is unable to prove disparate treatment, the majority leaves a plaintiff who is a victim of discrimination with no legal recourse. The dissent, on the other hand, recognizes that a plaintiff, while unable to prove intentional discrimination required by a disparate treatment analysis, may be able to show through a disparate impact analysis that solely subjective criteria impact a protected group. Under traditional discrimination law, both theories may be tried at the same time. Page v. U.S. Indus., Inc., 726 F.2d 1038 (5th Cir. *738 1984); Peters v. Lieuallen, 693 F.2d 966 (9th Cir.1982). And both may rely on statistics to prove discrimination. Page v. U.S. Indus., Inc., supra. There is no reason to foreclose a disparate impact analysis simply because only subjective criteria are involved. Moreover, the majority's per se rule immunizes upper level management policies from scrutiny and unfairly burdens employers who use objective as opposed to subjective criteria.

I believe, however, the dissent's formulation at page 742 which appears to require a plaintiff to choose an impact analysis to challenge the use of selection criteria and a treatment analysis for manner of application of selection criteria may be slightly confusing. I would simply permit use of a disparate impact analysis to prove discrimination when it might be difficult or impossible to prove intent to discriminate.

BUSINESS NECESSITY

The majority's business necessity test relies on Title VII, congressional intent, and the United States Supreme Court interpretation of Title VII. However, this court is interpreting RCW 49.60 and need not be bound by those considerations. If this court chooses to place the burden of proof on the employer once a prima facie case of discrimination has been proven, it may do so regardless of federal law.

The majority objects to the business necessity test definition of "job related" in Robinson v. Lorillard Corp., 444 F.2d 791 (4th Cir.), cert. dismissed, 404 U.S. 1006, 30 L. Ed. 2d 655, 92 S. Ct. 573 (1971). Majority, at 730-32. However, we have already adopted the federal allocation of burden of proof, placing on the employer the burden to show that the practice in question is necessary to the successful performance of the job. Fahn v. Cowlitz Cy., 93 Wn.2d 368, 384 n. 5, 610 P.2d 857, 621 P.2d 1293 (1980). This requirement is similar to the Lorillard test rejected by the majority. The majority's rejection of the Lorillard "job related" definition and imposition of the burden on the employer to *739 prove that no less discriminatory alternative exists to carry out its business narrowly, rather than liberally, construes RCW 49.60.

For these reasons I join Justice Brachtenbach's opinion.

PEARSON, J., concurs with UTTER, J. BRACHTENBACH, J. (concurring in part, dissenting in part)

I concur in the result because I agree that instruction 17, the statement of the "four-fifths" rule, is an impermissible comment on the evidence. However, I disagree with the majority's treatment of disparate impact analysis and the payment of expert witness costs.

Underlying my disagreement with the majority is a basic difference between the majority's and my understanding of the purpose of the discrimination statute. I believe the majority inaccurately dichotomizes potential interpretations of the statute; the majority endorses an interpretation of the statute which is solely concerned with eliminating intentional discrimination; it rejects, as do I, the opposite extreme: an interpretation of the statute which requires an employer to hire a racially and sexually balanced work force. However, the majority overlooks the middle ground. It does not acknowledge an interpretation of the statute which creates a duty in employers to examine their employment practices and attempt to eliminate those aspects of hiring and promotion systems which invidiously discriminate against minorities and women. International Bhd. of Teamsters v. United States, 431 U.S. 324, 364, 52 L. Ed. 2d 396, 97 S. Ct. 1843 (1977). Discrimination is too well rooted in our society to be eliminated merely by forbidding intentional bias. If the law is to have any meaning for those groups against whom discrimination has been aimed, the law must prohibit those methods of discrimination which have become institutionalized and no longer require the active bias of individuals.

First, I object to the majority's restriction of disparate impact analysis to cases in which objective selection criteria *740 are challenged. The use of certain selection criteria, whether they are subjective or objective, may have a disparate impact on a protected group. In a case involving hiring or promotions, disparate impact analysis should apply whenever a selection criterion is challenged. Disparate treatment analysis is required when the challenge is to the way in which selection criteria are applied. The majority's distinction between objective and subjective selection criteria is based on faulty reasoning in several federal cases. While we are guided by federal precedent in interpreting our discrimination statute, we should not be blind to their lapses in logic.

Disparate impact analysis allows a plaintiff to challenge facially neutral employment practices which have a disparate impact on a protected group. Griggs v. Duke Power Co., 401 U.S. 424, 430, 28 L. Ed. 2d 158, 91 S. Ct. 849 (1971). Unlike disparate treatment analysis, an employer's intent to discriminate is irrelevant to establishing a disparate impact claim. Griggs, at 432. A plaintiff may not simply point to underrepresentation of a protected group in an employer's work force and claim that the employer's general hiring or promotion system has a disparate impact, however. The plaintiff must identify a specific aspect of the selection process and prove that that facially neutral practice has a disparate impact. Spaulding v. UW, 740 F.2d 686, 707-08 (9th Cir.), cert. denied, ___ U.S. ___, 83 L. Ed. 2d 401, 105 S. Ct. 511 (1984); Pouncy v. Prudential Ins. Co. of Am., 668 F.2d 795, 801 (5th Cir.1982); Equal Empl. Opportunity Comm'n v. Federal Reserve Bank, 698 F.2d 633, 638-39 (4th Cir.1983), rev'd on other grounds sub nom. Cooper v. Federal Reserve Bank, 467 U.S. 867, 81 L. Ed. 2d 718, 104 S. Ct. 2794 (1984). Nor may a plaintiff identify as a facially neutral practice the fact that an employer utilizes subjective selection criteria. The use of subjective selection criteria is an accepted and sometimes necessary method of selecting employees. As a matter of policy, courts are unwilling to hold that their use alone is prima facie evidence of employment discrimination. Heagney *741 v. UW, 642 F.2d 1157, 1163 (9th Cir.1981); Harris v. Ford Motor Co., 651 F.2d 609, 611 (8th Cir.1981).

Thus, disparate treatment analysis is utilized to challenge alleged discriminatory bias for which the use of subjective criteria provides the opportunity. However, from this premise some federal courts incorrectly reason that only disparate treatment analysis should be utilized whenever a subjective selection criterion is challenged. Walls v. Mississippi Dep't of Pub. Welfare, 730 F.2d 306, 321 (5th Cir.1984); Spaulding v. UW, supra at 709. In rejecting the use of disparate impact analysis when a particular subjective criterion is challenged, these courts, as does the majority, fail to distinguish between a challenge to the use of subjective criteria in general and a challenge to the use of a particular subjective criterion. Similar to an objective selection criterion, a subjective selection criterion may have a disparate impact on a protected group even though the criterion is applied without bias. The only way to demonstrate the discriminatory effects of such a selection criterion, whether the criterion is subjective or objective, is through disparate impact analysis.

For example, prospective applicants may be subjectively evaluated as to how well they would "fit in" with other employees and customers. Having never worked with women or minorities, an employer's otherwise white male employees might not be as comfortable with women or minority applicants as with white male applicants. An impartial evaluator might correctly conclude that a woman or a minority would not "fit in" as well as other white male applicants. Or as here, an applicant may be subjectively evaluated as to her level of community service. Traditionally responsible for more household duties, working women may have less time for community service. An impartial evaluator might correctly determine that a woman applicant rates lower than a male applicant on this criterion. In both cases any discriminatory effect is caused by use of the criterion, not the subjective bias of the evaluator. See, e.g., Peters v. Lieuallen, 693 F.2d 966, 970 n. 2 (9th Cir.1982).

*742 When a selection criterion is challenged, it would be inappropriate to use disparate treatment analysis to analyze the discriminatory effect because there need not be an intent on the part of the evaluator to discriminate. In addition, it simply does not make sense to analyze a challenge to a selection criterion under disparate treatment analysis. In a disparate treatment suit, an essential element of the plaintiff's case is that the plaintiff is qualified, i.e., meets the employer's selection criteria. Yet it is the legitimacy of the selection criteria that the plaintiff is challenging, whether the challenged criterion is subjective or objective. Hung Ping Wang v. Hoffman, 694 F.2d 1146, 1148 (9th Cir.1982). Disparate impact analysis is the appropriate way to evaluate such selection criteria because it focuses inquiry on the effect of using the criteria and the employer's need to use the criteria. Thus, in determining whether to apply disparate impact or disparate treatment analysis, I would not distinguish between subjective and objective selection criteria. Instead, I would determine which method of analysis to apply to a particular case based on whether a challenge is made to the use of particular selection criteria or a challenge is made to the manner in which the selection criteria are applied.

I also take issue with the majority's analysis of the business necessity defense. Once a plaintiff has proven that a particular employment practice has a disparate impact on a protected class, an employer may justify the use of that practice by showing that the employment practice has a "manifest relationship to the employment in question." Griggs v. Duke Power Co., supra at 432. The federal courts have developed a variety of tests to determine whether a "manifest relationship" has been established. E.g., Robinson v. Lorillard Corp., 444 F.2d 791, 798 (4th Cir.) (practice must be "necessary to the safe and efficient operation of the business"), cert. dismissed, 404 U.S. 1006, 30 L. Ed. 2d 655, 92 S. Ct. 573 (1971); United States v. Jacksonville Terminal Co., 451 F.2d 418, 451 (5th Cir.1971) (practice must be essential to the safe and efficient operation of the business); *743 Contreras v. Los Angeles, 656 F.2d 1267, 1278 (9th Cir.1981) (practice accurately predicts or significantly correlates with job performance).

The majority rejects the use of the test set out in Robinson v. Lorillard Corp., supra at 798, which requires an employer to show that a practice with a disparate impact is necessary to the safe and efficient operation of the business. In the face of a statutory directive to construe RCW 49.60 liberally to achieve its purpose, RCW 49.60.020, the majority has chosen a very lenient test of business necessity. The majority follows Contreras v. Los Angeles, supra, and requires only that an employer show that a challenged employment practice "accurately predicts or significantly correlates" with job performance.

In defense of this choice, the majority asserts that to require a more compelling justification for the use of an employment practice already proven to have a disparate impact "would impose an affirmative duty to hire a balanced work force". Majority, at 730. This is simply not the case. Disparate impact analysis focuses on the effects of the challenged selection criteria, not on the composition of the work force. A more stringent test for business necessity does not create a duty to hire particular applicants. At most, it would make it somewhat more difficult for employers to justify the use of selection criteria shown to have a disparate impact on protected groups and would more often require employers to substitute another selection criterion. Even if an employer's work force is "balanced," a particular employment practice with a disparate impact is subject to challenge. Connecticut v. Teal, 457 U.S. 440, 449-51, 73 L. Ed. 2d 130, 102 S. Ct. 2525 (1982).

The test chosen by the majority does very little to further the discrimination statute's purpose of eliminating all forms of discrimination in employment. A selection criterion may be said to "accurately predict" or "significantly correlate" with job performance if there is sufficient evidence to indicate that the apparent statistical relationship between a selection criterion and job performance is not *744 due to chance. This test says nothing about the strength or importance of the relationship between the selection criteria and job performance. A 2 percent correlation is significant if a statistician's sample size is large enough to allow the statistician to estimate with confidence such a small correlation. In setting forth this test, the majority ignores the important difference between the reliability of a particular selection criterion as a predictor of a sought-after skill or quality in an employee and the importance of the skill or quality sought to job performance.

In addition, I have practical objections to the use of the test established by the majority as a jury instruction. The test relies too heavily on statistical terminology. Presumably each side will present experts who will testify that a particular selection criterion does or does not significantly correlate with job performance, leaving the jury to determine only which expert to believe. I prefer a jury instruction which establishes a nontechnical standard against which the jury may independently evaluate the experts' statistical evidence. Such an instruction also allows for those cases in which evidence is not technical.

Once a plaintiff has proven that a selection criterion has a disparate impact on a protected group, I would require the employer to identify the skill or characteristic which the challenged selection criterion is alleged to select. The employer would then have to show (1) that the possession of this skill or characteristic by the employee is necessary to the safe or efficient operation of the business and (2) that the selection criterion is a reliable method of determining if the applicant possesses the skill or characteristic.

Finally, I disagree with the majority's refusal to allow Shannon to recover expert witness fees. I agree that expert witness fees are not a "cost of suit". However, in addition to allowing a plaintiff to recover actual damages, attorney's fees and costs, RCW 49.60.030(2) authorizes the award of "any other remedy authorized by this chapter or the United States Civil Rights Act of 1964", 42 U.S.C. § 2000(a) et seq. The federal courts have long included the recovery of *745 expert witness fees among the remedies available to a plaintiff prevailing in a Title VII suit. E.g., Thornberry v. Delta Air Lines, Inc., 676 F.2d 1240, 1244-45 (9th Cir.1982). Our statute should also be interpreted to allow a plaintiff to recover expert witness fees. To hold otherwise would conflict with the intent of the statute to allow a plaintiff to recover all losses resulting from an employer's discrimination. Such an award is, of course, within the discretion of the trial court. The fees must be reasonable and should be awarded only if an expert witness was necessary to prosecute the case.

UTTER and PEARSON, JJ., concur with BRACHTENBACH, J.

NOTES

[1] The assistant manager position is the second tier in Ernst's 5-tier management structure. The entry level management position is department head. Assistant managers oversee department heads and are, in turn, supervised by store managers. Store managers are supervised by district managers who are accountable to the highest tier of the management structure, division managers.

[2] RCW 49.60.030 provides in pertinent part:

"(1) The right to be free from discrimination because of race, creed, color, national origin, sex, or the presence of any sensory, mental, or physical handicap is recognized as and declared to be a civil right. This right shall include, but not be limited to:

"(a) The right to obtain and hold employment without discrimination ...

"(2) Any person deeming himself injured by any act in violation of this chapter shall have a civil action in a court of competent jurisdiction to enjoin further violations, to recover the actual damages sustained by him, or both, together with the cost of suit including a reasonable attorney's fees or any other remedy authorized by this chapter or the United States Civil Rights Act of 1964 ..."

[3] The contested portion of instruction 18 states:

"The test is not merely whether there exists a business purpose for adhering to a challenged practice. The test is whether there exists an overriding legitimate business purpose such that the practice is necessary to the safe and efficient operation of the business. The challenged procedure must effectively carry out the business purpose it is alleged to serve, and there must be no acceptable alternative policies or practices which would better accomplish the business purpose advanced, or accomplish it equally well with a lesser differential of sexual impact."

[4] "The legislature hereby finds and declares that practices of discrimination against any of its inhabitants because of ... sex ... are a matter of state concern, that such discrimination threatens not only the rights and proper privileges of its inhabitants but menaces the institutions and foundation of a free democratic state." RCW 49.60.010.

"The provisions of this chapter shall be construed liberally for the accomplishment of the purposes thereof." RCW 49.60.020.