The appeal in this race discrimination case concerns disparate treatment and disparate impact under the W.Va. Human Rights Act, W.Va.Code, 5-11-1 [1967] et seq. Rose Marie Bradsher is a black woman. In August 1974, Ms. Bradsher applied to Guyan Valley Hospital to work as a nurse’s aide. There were no job openings at the time. In July 1976, there was an opening, for which Ms. Bradsher filed a new application. She was not hired. The job went to Rose Marie Wade, a white woman. Ms. Bradsher was qualified for the job by two years of college and two months experience as a nurse’s aide. Ms. *253Bradsher had listed as a reference on her application the Logan Day Care Center, where she had worked before. Logan gave her a poor recommendation. The job application also asked for personal references, including any friends or relatives already working at Guyan Valley Hospital.
Ms. Bradsher brought this case initially before the W.Va. Human Rights Commission. The commission found that the Hospital had violated Ms. Bradsher’s rights under W.Va.Code, 5-11-9 [1981] and ordered back pay and damages for mental suffering. The Hospital appealed to the Circuit Court of Kanawha County, which reversed the commission’s ruling. This appeal followed.
This Court has recognized a cause of action, under W. Va. Code, 5-11-9 [1981], for “disparate treatment,” that is, deliberate discrimination. We have not considered before whether there is a cause of action for “disparate impact.” “Disparate impact” is when employers do not deliberately discriminate, but their hiring practices have the effect of disproportionately excluding persons on the basis of race, age, and so forth. We now hold that there is a cause of action for “disparate impact” under W.Va.Code, 5-11-9 [1981].
W Va. Code, 5-11-9 [1981] prohibits “discrimination” on the basis of race. “Discrimination” means deliberately treating individuals differently because of different individual traits. Examples of such traits are an ability to read, possession of a professional degree, physical strength, or experience in a given line of work. “Illegal discrimination” means treating individuals differently because of some individual trait that the law says can’t legitimately be considered. Examples of such traits are race, age, sex, and handicap.
If a plaintiff cannot prove intentional discrimination under the disparate treatment theory, he may proceed under a disparate impact theory. This is appropriate when a facially-neutral hiring policy has the effect of irrationally excluding persons in a protected class.
I. Disparate Treatment
This Court held in Syl.Pt. 3, Conaway v. Eastern Assoc. Coal Corp., 178 W.Va. 164, 358 S.E.2d 423 (1986):
In order to make a prima facie case of employment discrimination under the West Virginia Human Rights Act, W. Va. Code, 5-11-1 et seq. (1979) the plaintiff must offer proof of the following:
(1) That the plaintiff is a member of a protected class.
(2) That the employer made an adverse decision concerning the plaintiff.
(3) But for the plaintiff’s protected status, the adverse decision would not have been made.
Or, more clearly put, “I was [not hired] because I was [black].” Id. 178 W.Va. at 170, 358 S.E.2d at 429, n. 5.
We cannot find substantial evidence on the record before us to support the commission’s finding of deliberate discrimination. However, because deliberate discrimination is difficult to prove directly, we give the Commission the benefit of the doubt and assume arguendo that Ms. Bradsher made out a prima facie case.
This Court held in Conaway, supra:
Once the plaintiff has met this burden, the burden would shift to the employer to show some non-discriminatory reason for the decision. The reason need not be a particularly good one.... The reason can be any other reason except that the plaintiff was a member of a protected class. If the fact finder believes that the proffered reason was the true reason for the decision, then the employer ... is not guilty of discrimination.
Id. 178 W.Va. at 171, 358 S.E.2d at 430.
The Hospital showed that it refused to hire Ms. Bradsher because she had unfavorable references from a past job in the same field. The Commission considered this and held in its Conclusions of Law, “Even the most casual examination of the facts of this case reveal[s] that the complainant has established her prima facie case and the respondent has failed to rebut it.” This conclusion was clearly wrong as a matter of law.
*254Nor did Ms. Bradsher offer any substantial evidence to prove that her poor recommendation was a mere pretext for a racial motive. On the record, then, there is simply no way the Commission could rationally have found an instance of explicit racial discrimination in the Hospital’s refusal to hire Ms. Bradsher. However, saying in this case that there was no malicious racial motive does not necessarily dispose of the matter if the totality of the hospital’s hiring practices had the effect of systematically excluding blacks.
II. Disparate Impact
A plaintiff makes out a case of disparate impact by identifying a particular hiring practice that has caused statistical under-representation of a given group within the relevant labor market. The plaintiff must then also demonstrate that the challenged practice is not fairly linked to job performance. See Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 109 S.Ct. 2115, 104 L.Ed.2d 733 (1989).
Ms. Bradsher claims that the Hospital’s use of job references has a disparate impact on black applicants. Her theory is that using personal references, such as friends who have worked with the employer, tends to exclude those without friends or relatives already working there. This is perfectly reasonable. If an employer has hired only whites in the past, undue reliance on references from friends and relatives already working for the employer will tend to perpetuate the existing racial composition.
Ms. Bradsher also claims that personal references as such (not just from friends already working for the employer) are not fairly linked to job performance. This is not so. Anyone who knows a job applicant well may be able to address such intangibles as honesty, friendliness, industry, and the like, even though the person providing the reference knows nothing at all about the job applicant’s professional qualifications or job experience. Because intangible, non-professional qualities are important in job performance, personal references as such are valuable to employers and legitimate to consider.
Proof of disparate impact is ordinarily by statistics. Interpreting Title VII of the Civil Rights Act of 1964, in the recent case of Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 109 S.Ct. 2115, 104 L.Ed.2d 733 (1989), the U.S. Supreme Court held that the proper comparison is “between the racial composition of the qualified persons in the labor market and the persons holding at-issue jobs.” Id. 490 U.S. at 650, 109 S.Ct. at 2121; Accord, Hazelwood School Dist. v. United States, 433 U.S. 299, 308, 97 S.Ct. 2736, 2741, 53 L.Ed.2d 768 (1977).
The job of nurse’s aid at Guyan Hospital is unskilled. A high-school diploma is not required, and training is on-the-job. Thus the relevant labor market is all of Logan County. According to the 1980 census, Logan County is 95% white and 4% black. The work force at Guyan Hospital is 9% black. During the relevant period in this case, August 1974 through July 1976, the Hospital hired fourteen people as nurse’s aides. Eleven were white, two were black, and one was oriental. Over 14% of those hired for the at-issue jobs during the relevant period were black, in a labor force that was 4% black.
Thus, because there was no statistical disparity in the at-issue jobs, there can perforce be no disparate impact claim. Accordingly, the judgment of the Circuit Court of Kanawha County is affirmed.
Affirmed.