Jacqueline Piva brought this action against her former employer, Xerox Corporation, under Title VII of the Equal Employment Opportunity Act, 42 U.S.C. § 2000e. She sought individual relief, alleging that Xerox discriminated against her on the basis of sex in compensating her, in evaluating her job performance, and later in discharging her. She also sought class action relief, alleging that Xerox engaged in a pattern and practice of employment discrimination against female applicants and employees. After a trial.on the merits, the district court found in favor of Xerox on all claims. We affirm.

Background

Piva began working for Xerox in 1964 as a customer representative. In 1966 she started working in sales, and in 1967 she was promoted to a top-level sales position. She was discharged in June of 1970. Her individual sex discrimination claim is based primarily on her assertions that she was paid less than males in comparable sales positions, and that her discharge was the result of sex discrimination.

Piva’s class action claim is based on her assertion that Xerox has engaged in a pattern and practice of discrimination against women in hiring and compensation. Piva relies primarily on statistical evidence to support her class action claims.

The district court ordered a bifurcated trial of the individual and class claims, certifying a class of plaintiffs for the purpose of granting injunctive relief only. Piva v. Xerox Corp., 70 F.R.D. 378 (N.D.Cal.1975). After a trial on the class claims the district court issued a initial memorandum of decision, finding that Xerox had discriminated in hiring during 1971 and 1972. After motions by both parties for reconsideration of its decision, the district court issued a supplemental memorandum reversing its findings concerning the claim of hiring discrimination during the years 1971 and 1972. Thus, the court denied relief with respect to all class claims.

Following a separate trial on Piva’s individual claims, the district court found that Piva had established a prima facie case of discrimination, but that Xerox had rebutted it by articulating legitimate reasons for its actions, and that a preponderance of the evidence had established that the reasons were not pretextual. The court also concluded that, even if Xerox had violated Title VII, Piva was not entitled to monetary relief because her losses were caused, not by her termination, but by her subsequent withdrawal from the job market.

Discussion

A. Standard of Review

Both Piva’s individual and class action claims are based on a disparate treatment theory. Thus, proof of discriminatory intent on the part of Xerox is a necessary element of Piva’s case. Texas Dept, of Community Affairs v. Burdine, 450 U.S. 248, 253 , 101 S.Ct. 1089, 1093 , 67 L.Ed.2d 207, 215 (1981); Golden v. Local 55, Etc., Ass’n of Firefighters, 633 F.2d 817, 821 (9th Cir. 1980). A Title VII claim brought under a disparate treatment theory is distinguishable from a claim brought under a disparate impact theory, in that in the latter case proof of discriminatory intent is not required. See Griggs v. Duke Power Co., 401 U.S. 424, 432 , 91 S.Ct. 849, 854 , 28 L.Ed.2d 158 (1971).

The plaintiff in a Title VII action bears the initial burden of establishing a prima facie case of employment discrimination. McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802 , 93 S.Ct. 1817, 1824 , 36 L.Ed.2d 668 (1973). Once a prima facie case is established, the burden shifts to the defendant to articulate legitimate reasons for its action and to produce evidence “which would allow the trier of fact rationally to conclude that the employment decision had not been motivated by discriminatory animus.” Texas Department of Community Affairs v. Burdine, 450 U.S. at 257 , 101 S.Ct. at 1096 , 67 L.Ed.2d at 218 . Finally, a plaintiff is given an opportunity to demonstrate that the proffered reasons are pretextual. McDonnell Douglas Corp. v. Green, 411 U.S. at 804 ,. 93 S.Ct. at 1825 .

In this circuit, the trial court’s conclusions regarding the success or failure of the plaintiff and defendant in meeting these burdens are reviewed under the clearly erroneous test. Golden v. Local 55, Etc., Ass’n of Firefighters, 633 F.2d at 820 . This standard of review is particularly appropriate where, as in this case, the underlying facts are in dispute and the validity of proffered statistical evidence is in question. See White v. City of San Diego, 605 F.2d 455, 459-60 (9th Cir. 1979).

B. The Class Action Claims

As noted above, the trial court certified a class of plaintiffs with respect to injunctive relief only. The class action claim was based on allegations that Xerox had discriminated against women in hiring, compensation, training, rewards, and termination policies. The evidence introduced at the trial related to the years 1965 — 76. The district court found for Xerox on all claims. Piva appeals only those holdings relating to discrimination in hiring and compensation for the years 1970-76. We first examine Piva’s assertion that the trial court erred in concluding that Xerox had not discriminated against women in hiring.

In attempting to establish a prima facie case of hiring discrimination, Piva relied upon statistical evidence rather than specific incidents of discriminatory conduct. For the years 1966-72 Piva provided a statistical comparison of the women employed by Xerox in its sales force with the women in the relevant available labor force. Piva did not introduce statistics comparing the number of women applicants with the number of women hired, because Xerox was unable to provide such records for these years. The district court assumed that Piva’s statistical evidence was sufficient to establish a prima facie case for the years 1965-1972.

The trial court went on to determine, however, that there was substantial evidence that there was an insignificant number of female applicants through the year 1971. Thus, the court held that Piva’s prima facie case has been rebutted on the basis that the lack of applicants provided the reason for the small number of women in the sales force. The trial court’s determination was based upon the testimony of numerous persons involved in the hiring process at Xerox who testified as to the lack of female applicants and the absence of a company policy of discriminating against any women who did apply.

As to the year 1970, the first year at issue here, we conclude that the trial court’s finding that Piva’s prima facie case was rebutted by evidence that there was a lack of female applicants is not clearly erroneous. White v. City of San Diego, 605 F.2d at 459-60 . The testimony of Xerox employees regarding the number of female applicants and the official employment policies at Xerox provide sufficient support for the finding.

As to the year 1971, the trial court initially found that Xerox had not rebutted Piva’s prima facie case. This conclusion was evidently based on a projection back to the year 1971 from an applicant flow chart provided by Xerox for the years 1973-75, which indicated to the trial court that a significant number of women began applying for sales positions at Xerox in 1971. However, after a motion for reconsideration, which cast doubt upon the mathematical accuracy of the projection, the trial court reversed this determination, stating that upon reviewing the evidence it appeared that the number of women applicants in 1971 remained insignificant. This conclusion apparently was based on the testimony of the original witnesses regarding the number of female applicants. This was substantial, believable evidence. This determination of the trial court that Xerox did rebut Piva’s prima facie case we also uphold under the clearly erroneous standard, in light of the testimony of these witnesses.

The trial court also initially found that Xerox had not rebutted Piva’s prima facie case as to the year 1972, but reversed its conclusion following motions for reconsideration. This change in the trial court’s conclusion was based on its consideration of Xerox’s Exhibit Q, which consisted of work sheets and a summary of applicant flow relating to 1972. The trial court found that Exhibit Q reflected that women “fared better than men” in obtaining employment with Xerox during that year.

Piva argues that Exhibit Q, which was prepared by Xerox in response to Piva’s interrogatories, should not have been admitted because it was hearsay. Xerox contends that the exhibit was admissible under Fed.R.Evid. 803(24), the “catch-all” exception for trustworthy hearsay that does not fall within any of the established exceptions. To be admissible under this rule, the proffered evidence must have “circumstantial guarantees of trustworthiness” equivalent to those of the other hearsay exceptions. Rule 803(24) requires the proponent to notify the adverse party or his or her intention to offer the evidence under the rule sufficiently in advance of the trial or hearing to provide the adverse party with a fair opportunity to meet the evidence. It also requires the trial court to make certain findings regarding the trustworthiness of and the necessity for the evidence. Xerox apparently did not comply fully with the notice requirements of Rule 803(24). In the circumstances here, however, we conclude that even if there was error it was harmless.

The purpose of the notice requirement is to give adverse parties an opportunity to attack the trustworthiness of the evidence. Piva had ample opportunity to do so, and the record shows that her attorney extensively cross-examined the preparers of the exhibit concerning the original data on which it was based and the actual construction of the exhibit. In addition, Piva had more than a year after the admission of the evidence until the end of the trial on her individual claims to move to strike the exhibit or to rebut it with additional evidence. See United States v. Leslie, 542 F.2d 285, 291 (5th Cir. 1976) (failure to satisfy notice requirements of Rule 803(24) is harmless error where adverse party had fair opportunity to meet the challenged statements). Moreover, we note that in admitting the evidence the trial court specifically stated its belief that the exhibit was a reliable summary of the original employment documents.

Because Exhibit Q provided substantial evidence supporting the trial court’s determination that Xerox rebutted the prima facie case as to 1972, we hold that the trial court’s finding is not clearly erroneous.

The trial court found that Piva failed to establish a prima facie case of discrimination in hiring for the years 1973-1976. The court considered substantial statistical evidence in reaching this conclusion, including records kept by Xerox relating to the number of female sales employees for the years 1973-76, Xerox’s applicant records for these years, and Bureau of Labor Statistics indicating the percentage of women employed in various occupational fields.

When an employment position involves skills that many persons possess or can easily learn, statistical comparisons between the composition of the defendant’s employees in the position and the composition of the general work force can be highly probative. Hazelwood School District v. United States, 433 U.S. 299 , 308 n. 13, 97 S.Ct. 2736, 2742,n. 13 , 53 L.Ed.2d 768 (1977); International Brotherhood of Teamsters v. United States, 431 U.S. 324, 337-39 , 97 S.Ct. 1843, 1855-56 , 52 L.Ed.2d 396 (1977). Gross statistical disparities between the composition of an employer’s work force and the composition of the general population in a proper case may constitute, by themselves, prima facie proof of a pattern or practice of discrimination. Hazelwood, 433 U.S. at 307-08 , 97 S.Ct. at 2741-42 ; Teamsters, 431 U.S. at 339 , 97 S.Ct. at 1856 . On the other hand, “[w]hen special qualifications are required to fill particular jobs, comparisons to the general population . .. may have little probative value.” Hazelwood, 433 U.S. at 308 n. 13, 97 S.Ct. at 2741-2742 n. 13.

The disparity between the percentage of female Western Region salespersons and the percentage of women in the general work force is statistically significant. The probative value of the disparity in proving discrimination, however, depends in large part upon whether special qualifications are required to do the job.

Xerox introduced evidence that the job of its Western Region sales personnel required advanced sales skills and that the failure rate among Western Region salespersons was relatively high. The district court implicitly found that Piva’s statistical comparison with the general work force was not highly probative. Because there was no evidence that most workers would be qualified to be Xerox salespersons, the district court did not err in finding the statistical disparity unpersuasive. Cf. EEOC v. United Virginia Bank/Seaboard National, 615 F.2d 147 (4th Cir. 1980) (comparison between composition of general work force and composition of bank clerical staff, including tellers, not probative enough to establish prima facie case of discrimination).

The representation of women in Xerox’s Western Region sales force increased dramatically during the time period in question. The percentage of women in sales increased from 4.6% in 1973 to 11.1% in 1976. Xerox introduced evidence that in 1973 Xerox’s chairman of the board notified branch managers that they would be held accountable for any failure to enforce equal employment opportunity policies. In light of this evidence and the absence of specific alleged incidents of discrimination, we hold that in finding that Piva failed to establish a prima facie case of a pattern and practice of hiring discrimination for these years the district court was not clearly erroneous.

We turn next to Piva’s assertion that the trial court erred in concluding that Xerox had not discriminated against women with regard to compensation. The evidence offered by Piva related to the time periods 1970-72 and 1973-76.

The evidence relating to 1970 — 72 consisted primarily of (1) the testimony by an employee of the California State Division of Industrial Welfare who supervised an investigation of Xerox following a complaint from Piva, (2) the study produced during that investigation, and (3) reference to an “ASR plan” instituted by the San Francisco office, pursuant to which some women were channeled into low-level positions. The district court concluded that this evidence was insufficient to establish a prima facie case of class-wide discrimination in matters of compensation. We conclude that this finding was not clearly erroneous. The state study related only to Piva’s individual claim, and no attempt was made to conduct an investigation which extended beyond the San Francisco office. Similarly, the “ASR plan” was experimental and was in effect only for a time in the San Francisco office after which it was phased out at the.request of the personnel office. This evidence is thus insufficient to make out a prima facie case that Xerox engaged in a pattern or policy of sex discrimination over the ten-state area covered by the Western sales region.

The evidence produced concerning 1973-76 consisted of Xerox records regarding the average salaries of male and female employees. The district court concluded that this evidence was insubstantial and misleading, in that it failed to take into account different locations, levels of experience, or comparability of position, and found the evidence insufficient to make put a prima facie case of discrimination. We conclude that this determination was not clearly erroneous.

C. The Individual Claims

Piva began working for Xerox in 1964 as a customer representative, a service position. In 1966 she was promoted to a low-level sales position, and in 1967 she received a promotion to a top-level sales position.

Jerry Lewis became Piva’s manager in 1969. He introduced a policy in Piva’s sales unit that reduced the importance of revenue production and increased the importance of new orders and machine installations. In April, 1969 Lewis warned Piva in a memorandum that her work for the previous fifteen months had been inadequate in terms of orders for new units. He gave her further warnings in July and September of that year.

In a September memorandum, Lewis informed Piva that she would be transferred to a different sales position. When Piva learned that the reassignment would involve a substantial salary reduction, she refused to accept it. In response, Lewis informed Piva that the fourth quarter of 1969 would be an unofficial trial period for her.

In November, 1969 Ron Roberts replaced Lewis as Piva’s manager. He reviewed Piva’s performance and offered Piva the choice of accepting a new position or going on probation for sixty days. Piva chose probation and Roberts set up sales goals that Piva was to meet. The goal for December, 1969 was less than Piva’s normal quota in order to give her time to reach the expected sales level. Piva did not meet the goals for December, 1969 or January, 1970. Roberts allowed her a second probationary period. The period was extended until the end of April, 1970 because Piva became ill. On April 22, Roberts concluded that Piva could not reach her probation sales targets and recommended that she be terminated as an employee. She was terminated in June, 1970.

From 1968 until Piva’s employment was terminated in June, 1970, she was paid less than many male Xerox salespersons. According to the report by the California Division of Industrial Welfare, Xerox violated California law during this period by paying Piva less than male salespersons performing the same basic duties.

Piva’s individual claim is based upon allegations that she was discriminated against in compensation, in the application of performance standards, and in her termination. The trial court determined that

Piva contends that the district court erred in finding that Xerox rebutted her prima facie case of discrimination. As previously stated, we overturn such a finding only if clearly erroneous. Golden v. Local 55, Firefighters, 633 F.2d at 820 .

With respect to the claim of discrimination in Piva’s performance goals and discharge, the district court found that Xerox’s treatment of Piva was not related to her sex. This finding was not clearly erroneous. Several male salespersons also received critical evaluations and were placed on probation and fired for failing to attain sales goals. For example, Roberts was equally critical of a male co-worker of Piva’s, who became an account representative at the same time she did. This coworker was also discharged by Xerox, without being offered another position or the opportunity to prove himself during a probationary period. Further, there was evidence that similarly situated female salespersons succeeded in sales and received promotions. Piva’s introduction of evidence regarding the so-called “permanent ASR plan” pursuant to which some women were channeled into low-level positions, which is referred to above in connection with the class claim, is not significantly probative on the issue of whether Piva was discriminated against, as that plan did not involve her.

The district court also found that Xerox had rebutted Piva’s prima facie showing of pay discrimination. In a typical wage discrimination case, the plaintiff bears the initial burden to show that: (1) the work of the employees of one sex required the exercise of substantially equal skill, effort, and responsibility and was performed under working conditions similar to that of employees of the opposite sex; and (2) the pay to men and women was unequal. Corning Glass Works v. Brennan, 417 U.S. 188, 195 , 94 S.Ct. 2223, 2228 , 41 L.Ed.2d 1 (1974). The burden then shifts to the em ployer to show that the unequal pay results from a seniority system, a merit system, a system tying earnings to productivity, or any factor other than sex. 42 U.S.C. § 2000e-2(h); Corning Glass Works, 417 U.S. at 196 , 94 S.Ct. at 2229 .

The district court stated that the finding by the California Division of Industrial Welfare that Xerox discriminated against Piva in pay was not determinative, because the agency failed to consider job-related factors such as education and experience. Piva contends that the male salespersons receiving higher salaries than she received did not have more job-related education and experience than she had. The record refutes that contention. Unlike Piva, many of the men whose salaries were compared to Piva’s had degrees in business or management. Others had substantially more business or sales experience than Piva. The district court did not err in determining that these factors are related to high-level sales positions at Xerox. The court’s determination that Xerox had rebutted Piva’s prima facie case of discrimination in compensation is thus not clearly erroneous.

Because we affirm the district court finding that Xerox did not discriminate against Piva, we need not review the court’s holding that Piva failed to mitigate damages.

Ill

Conclusion

The district court’s findings of fact will not be disturbed on appeal unless they are clearly erroneous. We hold that there is sufficient evidence to support the findings of the district court and the judgment is AFFIRMED.

. We note that the Fifth Circuit concluded otherwise in Burdine v. Texas Dept, of Community Affairs, 608 F.2d 563, 566 (5th Cir. 1979), reversed, 450 U.S. 248 , 101 S.Ct. 1089 , 67 L.Ed.2d 207 (1981) and stated that: [a] lthough discrimination is a question of fact, it is also the ultimate issue for resolution in a Title VII case. Therefore, we as an appellate court must independently determine the merits of plaintiffs allegations, but we are bound by findings of subsidiary facts (evidentiary facts) that are not clearly erroneous. We must also determine whether the ultimate finding is based on requisite subsidiary facts. (citations omitted). The Supreme Court, in expressly declining to rule on this issue, stated that [b] ecause the Court of Appeals applied the wrong legal standard to the evidence, we have no occasion to decide whether it erred in not reviewing the District Court’s finding of no intentional discrimination under the “clearly erroneous” standard of Federal Rule of Civil Procedure 52(a). Addressing this issue in this case would be inappropriate because the District Court made no findings on the intermediate questions posed by McDonnell Douglas. 450 U.S. at 260, n. 12 , 101 S.Ct. at 1097, n. 12 , 67 L.Ed.2d at 219, n. 12 . Our holding in Golden is thus unaffected. . Piva has not raised an issue regarding this determination, so the question of its propriety is not now before us. The question whether Piva is a proper class representative is also not before us. We do note, however, that Piva was employed by Xerox for only four months during 1970-76, the time period at issue in her class action claim. Further, Piva purported to represent women who allegedly suffered discrimination in that they were not hired by Xerox. Piva had been hired by Xerox and this is not the basis of her individual claim. . We recognize that two circuits have strictly applied the procedural requirements of Rule 803(24). See United States v. Guevara, 598 F.2d 1094, 1100 (7th Cir. 1979); United States v. Ruffin, 575 F.2d 346, 358 (2d Cir. 1978). In neither case, however, did the adverse party have so great an opportunity to attack the trustworthiness of the offered statements. Piva also contends that Exhibit Q is inadmissible under the best evidence rule. Fed.R.Evid. 1004(1). We believe that Xerox made an adequate showing that the original documents were not destroyed in bad faith. . We note that the district court appears to have erroneously placed upon Piva the burden to take into account employee productivity. Had Piva established her prima facie case, it would have been Xerox’s burden to explain the wage disparity by reference to productivity or other legitimate factors as part of its rebuttal. Corning Glass Works v. Brennan, 417 U.S. 188, 196 , 94 S.Ct. 2223, 2229 , 41 L.Ed.2d 1 (1974). In light of the inadequacies in Piva’s proof of a prima facie case, however, the district court’s error was harmless. . In Corning Glass Works the claim was brought under the Equal Pay Act. When a Title VII claim alleges that the plaintiffs have been denied equal pay for equal work, Equal Pay Act standards apply. Gunther v. County of Washington, 623 F.2d 1303, 1318 (9th Cir. 1979), aff’d, - U.S. -, 101 S.Ct. 2242 , 68 L.Ed.2d 751 (1981). The Supreme Court in Gunther approved the Ninth Circuit’s holding that a claim of sex-based discrimination in compensation will lie under Title VII even when the jobs at issue are not substantially equal, provided that the challenged rate is not based on merit, seniority, quality of production, or any factor other than sex. Id. at-, 101 S.Ct. at 2247. Piva’s claim was brought under the theory that her job was equal to that of the male salespersons.