POLITZ, Circuit Judge:

Fired at the age of 47, James L. Stendebach, a managerial supervisor, invoked the Age Discrimination in Employment Act, 29 U.S.C. §§ 621- 634, against his employer, CPC International, Inc. (CPC). After both sides had rested their case, the district court directed a verdict for defendant. On appeal, Stendebach challenged the directed verdict and various evidentiary rulings. Finding no merit in any error assigned, we affirm.

During the latter part of 1977, the management of CPC, in an effort to revive and save its plant at Corpus Christi, Texas, instituted a plan to reduce and restructure the work force and otherwise effect production economies. The Corpus Christi facility, engaged in the labor intensive corn wet milling business, had been in a state of economic extremis for several years. Survival required a dramatic reduction in production costs, by a decrease in the labor force, an increase in unit production or both.

Subject to a general policy outlined by corporate headquarters, the reorganization became the charge of six senior managers in the Corpus Christi facility. Jobs were merged, functions were streamlined, performance minimums were markedly increased. Resulting jobs took on a qualitatively different character. To select those employees who would constitute the reduced labor force, qualified and capable of performing in the new, demanding employment structure, the select committee prepared a list of job qualifications for each of the new positions, together with a comprehensive list of criteria for rating the candidates for each job.

The rating process was the result of both individual and group action. All candidates, including all salaried employees such as Stendebach, considered arguably capable of performing the tasks demanded in new jobs, were individually graded by each of the six committee members who assigned numerical ratings. Thereafter, the committee discussed the ratings and attempted to achieve a consensus. The numbers were tabulated and the candidates with the highest totals got the job assignment.

During the first six months of 1978 the salaried work force was reduced from 76 to 37. Stendebach was one of the 39 terminated. The average age of the salaried work force prior to the reduction was 43.8 years; the average age post-reduction was 45.9 years.

The operation was a success but the patient expired. The gargantuan efforts improved efficiency and reduced production costs but economic conditions were such that the plant could not continue as a viable business entity. By October of 1979, the activity in the Corpus Christi facility was stilled, and all but three of the remaining work force were terminated.

To prevail in an age discrimination case, as presented here, a plaintiff must establish that the defendant-employer either refused to consider his retention or relocation because of his age, or else regarded age as a negative factor. Williams v. General Motors, Inc., 656 F.2d 120 (5th Cir. 1981). The directed verdict was grounded on the finding by the district judge that this element of Stendebach's case was wanting.

To uphold this finding we must be persuaded that no reasonable jury could have concluded otherwise. Hedrick v. Hercules, Inc., 658 F.2d 1088 (5th Cir. 1981). The standard in this circuit, against which we measure a directed verdict, was announced in our en banc decision in Boeing Company v. Shipman, 411 F.2d 365, 374 (5th Cir. 1969):

On motions for directed verdict and for judgment notwithstanding the verdict the Court should consider all of the evidence — not just that evidence which supports the non-mover's case — but in the light and with all reasonable inferences most favorable to the party opposed to the motion. If the facts and inferences point so strongly and overwhelmingly in favor of one party that the Court believes that reasonable men could not arrive at a contrary verdict, granting of the motions is proper. On the other hand, if there is substantial evidence opposed to the motions, that is, evidence of such quality and weight that reasonable and fair-minded men in the exercise of impartial judgment might reach different conclusions, the motions should be denied, and the case submitted to the jury. A mere scintilla of evidence is insufficient to present a question for the jury.

Thus charged, we consider the evidence presented "in the light and with all reasonable inferences most favorable to" Stendebach. The totality of his evidence is composed of: (1) a list containing the ages of employees, (2) the statement by one committee member that someone at corporate headquarters had stated that age might be a relevant factor in the reorganization effort, and (3) an assortment of statistical data. Viewing same in its entirety, we are convinced that this evidence is, at best, a mere scintilla which is insufficient to present a jury question.

Accepting as true that someone in corporate management had told a committee member that age might be relevant, the exhaustive test eventually adopted is devoid of any reference to age. No numerical rating based on age was included. The tests, varied and tailored for the various jobs, and containing as many as 18 criteria, were strictly adhered to by all members of the committee and the final reduction-inforce decisions were based exclusively on the scores received. In this regard we note that the record is devoid of any suggestion that any representative of corporate management attempted to influence the members of the committee in their final decision as to which employees would be retained and which would be released.

Stendebach makes much of a list of employees prepared in October 1977 by Charles Shoemate, a corporate vice-president who had been plant manager in Corpus Christi from February 1974 until June 1976. The hand-written Shoemate list was not prepared in conjunction with the preparation or implementation of the reorganization plan. This list, containing names, ages, years of service, and grade and performance evaluations taken from the latest performance appraisal forms, was prepared for another purpose. There is no evidence that this list was used by the committee members, nor is there evidence that the contents or existence of the list in any way influenced their decisions. There is no evidence that Shoemate was involved in committee deliberations, directly or indirectly, and no suggestion that he was inclined to discriminate against older employees.

The only credible evidence respecting age supports CPC. Upon completion of the committee's efforts, its product was evaluated at corporate headquarters for the express purpose of assuring that the committee had not inadvertently discriminated against older employees.

Plaintiff's evidentiary base is thus reduced to the statistical data offered. We recognize the value of statistical evidence and we have and will dispose of a proper case on that basis. Harrell v. Northern Electric Company, 672 F.2d 444 (5th Cir. 1982). But we are cautious in our use of statistics drawn from small samples such as those involved in the instant case. The reason for our hesitance is obvious: "the smaller the sample size, the greater the likelihood that the underrepresentation reflects chance rather than discriminatory practices." Williams v. Tallahassee Motors, Inc., 607 F.2d 689, 693 (5th Cir. 1979). See Mayor of the City of Philadelphia v. Educational Equality League, 415 U.S. 605, 94 S.Ct. 1323, 39 L.Ed.2d 630 (1974).

Thus cautioned, we examine Stendebach's statistical offerings and find them totally unpersuasive of a discriminatory intent by CPC. By the end of 1977, there were 39 employees in the production, technical, and mechanical divisions of the Corpus Christi plant. It was against these workers that Stendebach competed. Of those between the ages of 40 and 70, the group protected by the ADEA, approximately 36% were laid off. Of those under the age of 40, 55% lost their jobs. This reduction effected an increase in the average age of the work force in these departments from 47.06 to 47.35 years of age. And, as earlier noted, after the reduction the average age of the salaried work force increased from 43.8 years to 45.9 years.

We note finally that appellant complains of the admission in evidence of the worksheets used in evaluating the employees. Stendebach maintains that these documents are self-serving and hearsay. The former is inconsequent; the latter is incorrect. The documents were introduced to show the methodology employed by the committee in rating the employees. They were not offered to prove the truth of but merely the fact of their contents. As such, the instruments are not hearsay.

The district court is AFFIRMED.