Brent Ryan Wilson, along with his parents, Susan and Ted Wilson, brought a diversity suit against Merrell Dow Pharmaceuticals Inc. (Merrell Dow), alleging that Bendectin, a drug manufactured by Merrell Dow’s predecessor Richardson-Merrell, Inc.1 and prescribed for Mrs. Wilson to alleviate morning sickness during her pregnancy with Brent, caused him to be born missing one finger on each hand. The Wil-sons alleged claims of products liability, fraud and misrepresentation, breach of express and implied warranty, strict liability, and negligence. After a three-week trial in which both parties presented expert witnesses, the jury returned a general verdict in favor of Merrell Dow.
The Wilsons now appeal the district court’s2 judgment entered on that verdict. They contend that the district court erred in (1) declining to give a jury instruction noting the failure of defense counsel to call an expert witness who had been expected to testify for Merrell Dow; (2) allowing defense counsel to tell the jury that the absent witness was equally available to the plaintiffs; (3) admitting into evidence Mer-rell Dow’s “sales charts,” which compared the rate of birth defects in the general population with the number of Bendectin tablets distributed and the number of Ben-dectin new therapy starts; and (4) failing to grant the Wilsons’ motion for judgment notwithstanding the verdict or a new trial. We affirm.
I.
Merrell Dow’s counsel told the jury during his opening statement that Dr. Burhan Say, a geneticist, would testify that Brent’s birth defect was a genetic condition and was thus not produced by Mrs. Wilson’s ingestion of Bendectin. After Merrell Dow completed presenting all of its evidence without calling Dr. Say to the witness stand, the Wilsons’ counsel requested that the district court give a missing witness instruction regarding Dr. Say. The district court declined to do so, suggesting instead that the plaintiffs’ attorney simply point out in his closing argument that Merrell Dow had failed to call Dr. Say as promised. The Wilsons’ counsel followed this suggestion and argued to the jury that Dr. Say’s nonappearance indicated that his testimony would have been adverse to Merrell Dow. Merrell Dow’s attorney then responded in his closing argument by asserting that the Wilsons also could have called Dr. Say as an expert witness but had not even attempted to subpoena him, and that therefore nothing should be inferred from Dr. Say’s failure to appear.
In their appeal the Wilsons contend that the district court should have given a jury instruction noting the failure of Dr. Say to testify. They also claim that it was improper for the district court to allow Merrell Dow’s attorney to argue to the jury that Dr. Say was equally available to testify for the Wilsons.
In reviewing the district court’s decisions to not give the requested missing witness instruction and to allow defense counsel’s comment during closing argument, this court recognizes that those decisions rest largely within the trial judge’s discretion. See, e.g., United States v. Sutton, 732 F.2d 1483, 1492 (10th Cir.1984) (“A trial court has discretion to give or refuse to give a missing witness instruction.”), cert. denied, 469 U.S. 1157, 105 S.Ct. 903, 83 L.Ed.2d 919 (1985); Chicago College of Osteopathic Medicine v. George A. Fuller Co., 719 F.2d 1335, 1352-53 (7th Cir.1983) (“Questions as to the propriety of comment by counsel in [closing] argument upon the failure to produce a witness rest largely in the discretion of the trial court.”).
Courts have recognized four factors that must be present before a jury can be instructed to infer that a missing witness would have testified adversely to a party: (1) the party must have the power to pro*1151duce the witness, see, e.g., Sutton, 732 F.2d at 1492; 2 J. Wigmore, Evidence in Trials at Common Law § 286 (J. Chadbourn rev. ed. 1979 & Supp.1989); (2) the witness must not be one who would ordinarily be expected to be biased against the party, see id. § 287, at 202 & n. 1; (3) the witness’s testimony must not be “comparatively unimportant, or cumulative, or inferior to what is already utilized” in the trial, see id. § 287, at 202-03 (emphasis omitted); and (4) the witness must not be equally available to testify for either side, see, e.g., Sutton, 732 F.2d at 1492; Quad Constr., Inc. v. William A. Smith Contracting Co., 534 F.2d 1391, 1394 (10th Cir.1976); 2 J. Wigmore, supra, at § 288.3 The party requesting a missing witness instruction adverse to the other side has the burden to demonstrate that these criteria are satisfied. See, e.g., Sutton, 732 F.2d at 1492 (criminal defendant has burden to show that there are missing government witnesses); Jones v. Otis Elevator Co., 861 F.2d 655, 659-60 (11th Cir.1988) (requesting party must establish that potential witness is unavailable and that potential testimony is relevant and noncumulative).
Factors one and two appear to be present in this case. Merrell Dow’s counsel’s comment in his opening statement about Dr. Say indicated that Merrell Dow had the power to call Dr. Say to testify and that this testimony was not expected to be biased against the defendant. It is not clear from the record or briefs, however, whether criteria three and four were met here.
The Wilsons point out that without Dr. Say’s testimony, Merrell Dow had no geneticist to testify. The plaintiffs, in contrast, called three geneticists, all of whom concluded that the birth defect was not genetically induced. Thus, one might argue that Dr. Say’s testimony was comparatively important because it was the only testimony from a geneticist available to Merrell Dow to rebut the conclusions of the Wilsons’ geneticists. On the other hand, one might reasonably determine that Dr. Say’s testimony was not so essential for Merrell Dow’s defense that Dr. Say’s failure to testify should be accorded any evi-dentiary significance. Merrell Dow presented experts from other scientific fields, whose testimony will be discussed later in this opinion, who testified that Ben-dectin did not cause birth defects. The Wilsons’ geneticists themselves acknowledged that many birth defects are genetic in nature and that Brent’s birth defect was of a type that had occurred in the human population long before the introduction of Bendectin into the market. In light of the other evidence supporting Merrell Dow’s position, it was within the discretion of the district court to conclude that Dr. Say’s testimony was cumulative and had relatively insignificant probative value.
Moreover, the district court acted within its discretion in determining that Dr. Say was equally available to both parties. The Wilsons argue that it was impossible for them to call Dr. Say as a witness because of Dr. Say’s relationship with Merrell Dow. That may be true, but the Wilsons did not even attempt to subpoena Dr. Say, nor did their counsel explain adequately in their appellate brief how Dr. Say’s relationship with Merrell Dow prevented the Wilsons from calling Dr. Say to testify. The district court could have properly concluded that the plaintiffs did not meet their burden of demonstrating that Dr. Say was either legally or practically unavailable to testify for them. See United States v. Montoya, 676 F.2d 428, 431 (10th Cir.1982) (holding that district court did not abuse its discretion in refusing criminal defendant’s request for a missing witness instruction regarding a government informant when defense counsel made only one attempt to meet with informant and did not attempt to *1152subpoena him), cert. denied, 459 U.S. 856, 103 S.Ct. 124, 74 L.Ed.2d 108 (1983).
We thus conclude that the district court did not abuse its discretion in declining to give a missing witness instruction. Similarly, it was permissible for the district court to allow defense counsel’s comment in closing argument that Dr. Say was equally available to testify for the Wilsons. When an absent witness is equally available to both parties, either party is open to the inference that the missing testimony would have been adverse to it. See 2 J. Wigmore, supra, at § 288, at 208 & n. 4 (listing cases supporting the “more logical view ... that the failure to produce [a witness equally available to both sides] is open to an inference against both parties, the particular strength of the inference against either depending on the circumstances”) (emphasis in original). In these circumstances, comment by both sides in closing argument regarding the missing witness is appropriate. See C. McCormick, McCormick’s Handbook of the Law of Evidence § 272, at 657-58 (E. Cleary ed. 1972). If one side makes “an argument on failure to produce [a witness that] is fallacious, the remedy is the usual one, namely the answering argument and the jury’s good sense.” Id. § 272, at 659.
II.
Next we consider whether the district court properly admitted into evidence Merrell Dow’s two sales charts containing graphs plotting the rate of birth defects in the general population, the number of Bendectin tablets distributed, and the number of Bendectin new therapy starts.4
Merrell Dow introduced the first sales chart, Exhibit 2288, into evidence during the testimony of Dr. James Lee Goddard, a former Commissioner of the Pood and Drug Administration (FDA). This chart, which was prepared by Dr. Goddard, compared the number of Bendectin tablets distributed with the rate of limb reduction birth defects for the years 1970-1984. Dr. Goddard calculated the number of Bendec-tin tablets distributed using data from annual reports that each pharmaceutical company was required to submit to the FDA. His sources for the rate of limb defects were annual reports of the Federal Center for Disease Control (CDC) in Atlanta, Georgia. The chart showed an increase and then decrease to zero for the number of Bendectin tablets distributed from 1970 to 1984, along with what Dr. Goddard called “a remarkably constant rate of limb defects” during the same period. Relying upon this chart and other evidence, Dr. Goddard concluded that Bendectin does not cause birth defects in general or limb reduction defects in particular.
The second chart, Exhibit 2280, was prepared by Dr. Steven H. Lamm, Merrell Dow’s epidemiological expert, and was introduced into evidence during his testimony. This chart compared the number of Bendectin new therapy starts with the rate of limb reduction birth defects for the years 1970-1984. Like Dr. Goddard, Dr. Lamm obtained data for the rate of birth defects from the CDC. Dr. Lamm’s source for the number of new therapy starts was Richard P. Smith, manager of new products and marketing research at Merrell *1153Dow, who derived this data from information supplied by all pharmaceutical companies to independent market research firms. Exhibit 2280 showed the number of new therapy starts fluctuating up and down before declining to zero during a period when birth defects remained constant. Based upon this chart and other studies, Dr. Lamm agreed with Dr. Goddard’s testimony that Bendectin does not cause birth defects.
The district court ruled that the charts were admissible as facts or data forming the basis of expert testimony under Federal Rule of Evidence 703. We review this decision under the abuse of discretion standard. See Marsee v. United States Tobacco Co., 866 F.2d 319, 321 (10th Cir. 1989).
The Wilsons contend that the sales charts were hearsay that should have been excluded pursuant to Federal Rule of Evidence 802. We agree that the charts were hearsay, but this determination does not end our analysis. Federal Rule of Evidence 703 allows an expert witness to base his testimony upon facts or data that are hearsay, provided that those facts or data are “of a type reasonably relied upon by experts in the particular field in forming opinions or inferences upon the subject.” We have interpreted Rule 703 as allowing an expert to reveal the basis of his testimony during direct examination, even if this basis is hearsay, provided that the facts or data underlying his conclusions are of a type reasonably relied upon by others in his field of expertise. See United States v. Affleck, 776 F.2d 1451, 1457-58 (10th Cir. 1985). The hearsay is admitted for the limited purpose of informing the jury of the basis of the expert’s opinion and not for proving the truth of the matter asserted. See id. at 1457.
After examining the trial transcript and sales charts, we conclude that the district court acted within its discretion in determining that the charts were of a type reasonably relied upon by experts who study birth defects and were therefore admissible under Rule 703.5 Dr. Goddard testified that the method he used in preparing Exhibit 2280 — comparing the rate of birth defects with the distribution of a drug in order to determine whether the drug produced birth defects — is known as pharma-coepidimiology and is becoming accepted in the research community. He noted that a 1980 study in Northern Ireland utilized a similar technique by comparing the number of Bendectin prescriptions over a seven-year period with the rate of birth defects and found that while the amount of Ben-dectin prescribed increased very steadily, the rate of birth defects declined during the same period. Based upon this finding, the study determined that Bendectin does not cause birth defects. Dr. Goddard also discussed a study conducted by teratolo-gists James Wilson and Clarke Fraser that compared the total distribution of the drug Contergan in West Germany with the rate of birth defects in that country from 1958 through 1962. According to Dr. Goddard, the West German study showed the rate of birth defects increased “almost in an identical fashion” with the sales of Contergan, which was later proven to cause birth defects. Dr. Goddard’s testimony regarding these previous studies was sufficient for the district court to determine that both Exhibits 2280 and 2288, which also utilized a pharmacoepidimiologieal technique, were of a type reasonably relied upon by birth defect experts.
The Wilsons contest the admissibility of the sales charts by arguing that the charts did not take into account when the Bendec-tin was actually consumed. They point out that in order for the drug to have caused a birth defect, a pregnant woman would have had to have ingested it during the early part of her pregnancy, when the fetus’s limbs were being formed. See Hull v. Merrell Dow Pharmaceuticals, Inc., 700 *1154F.Supp. 28, 29-30 (S.D.Fla.1988) (granting Merrell Dow’s motion for summary judgment because Bendectin was not taken during time of limb development). Because the charts did not distinguish those sales of Bendectin that occurred during the limb development period from those that did not, the Wilsons contend that the charts were misleading. We agree that this failure of the charts to take into account when the Bendectin was consumed may weaken their value as the basis for expert testimony, but this failure affects the weight, not the admissibility, of the charts under Rule 703. Cf. Bazemore v. Friday, 478 U.S. 385, 400, 106 S.Ct. 3000, 3009, 92 L.Ed.2d 315 (1986) (noting that the failure of a regression analysis to include other variables that may affect the salary level of a Title VII plaintiff “[njormally ... affect[s] the analysis’ probativeness, not its admissibility”). The Wilsons’ counsel had ample opportunity, which he utilized, to cross-examine Doctors Goddard and Lamm regarding the data and methods used to prepare the sales charts. This questioning was sufficient to bring to the jury’s attention any alleged inadequacies of the information contained in the charts.6
III.
Finally we consider the Wilsons’ contention that the district court erred in denying their motion for judgment notwithstanding the verdict or a new trial.
Judgment n.o.v. is warranted when the evidence is insufficient to support the verdict, that is, when “the evidence points but one way and is susceptible to no reasonable inferences sustaining the position of the party against whom the motion is made.” Cooper v. Asplundh Tree Expert Co., 836 F.2d 1544, 1547 (10th Cir.1988). A new trial should be granted if the verdict is “ ‘clearly, decidedly, or overwhelmingly against the weight of the evidence.’ ” Farmers Ins. Co. v. Hubbard, 869 F.2d 565, 571 (10th Cir.1989) (quoting Champion Home Builders v. Shumate, 388 F.2d 806, 808 (10th Cir.1967)). We will not reverse the district court’s denial of a motion for a new trial based upon the weight of the evidence unless there was a “manifest abuse of discretion.” Id.
Our review of the evidence leads to the conclusion that the district court’s ruling on the Wilsons’ motion was proper. Mer-rell Dow presented expert testimony, which was not contradicted by the Wilsons’ experts, that of the approximately forty epidemiological studies of Bendectin, none has shown a statistically significant association between ingestion of the drug and incidence of birth defects generally or limb defects in particular. This lack of epidemiological proof for the Wilsons’ claims is particularly significant in light of recent decisions of federal courts of appeals granting judgment n.o.v. for Merrell Dow based upon the absence of epidemiological evidence showing a causal relationship between Bendectin use and birth defects. See Brock v. Merrell Dow Pharmaceuticals, Inc., 874 F.2d 307 (5th Cir.), modified, 884 F.2d 166 (5th Cir.), reh’g en banc denied, 884 F.2d 167 (5th Cir.1989); Richardson v. Richardson-Merrell, Inc., 857 F.2d 823 (D.C.Cir.1988), cert. denied, - U.S. -, 110 S.Ct. 218, 107 L.Ed.2d 171 (1989); Lynch v. Merrell-National Laboratories, 830 F.2d 1190 (1st Cir 1987). Other evidence supporting the verdict included (1) testimony of Doctors Goddard and Lamm that from the 1960s through the early 1980s Bendectin new therapy starts and *1155overall sales gradually increased and then decreased to zero while the rate of birth defects remained constant; (2) testimony that animal studies of Bendectin, taken as a whole, indicated that the drug does not cause birth defects; and (3) testimony that a 1980 expert advisory committee to the FDA found no link between Bendectin use and birth defects.
Although the Wilsons called several experts who testified in support of their claims, Merrell Dow presented at least sufficient expert testimony to create a conflict in the evidence, and perhaps even enough to sustain a directed verdict under the reasoning of Brock, Richardson and Lynch. When the evidence is in conflict, the jury alone has the power to weigh that evidence and assess the credibility of witnesses, and we will not retry the facts. See Rogers v. Hyatt, 697 F.2d 899, 905 (10th Cir.1983).
Thus, we conclude that there was sufficient evidence for the jury to return a verdict in Merrell Dow’s favor and that it was within the district court’s discretion to determine that the weight of the evidence supported the verdict as well.
IV.
We find no error in the district court’s decisions to decline to give a missing witness instruction, to allow Merrell Dow’s counsel to comment that the missing witness was equally available to testify for the Wilsons, to admit Merrell Dow’s sales charts into evidence, and to deny the Wil-sons’ motion for judgment n.o.v. or a new trial. Accordingly, the district court’s judgment entered on the jury verdict in favor of Merrell Dow is affirmed.
. Over the years the entity now known as Mer-rell Dow Pharmaceuticals Inc. has been known as Richardson-Merrell, Inc., Merrell National-Laboratories and The Wm. S. Merrell Company. Merrell Dow is a wholly owned subsidiary of The Dow Chemical Company.
. The Honorable James O. Ellison, United States District Judge, Northern District of Oklahoma.
. For the purpose of our analysis, we assume that the missing witness inference may be made in a diversity suit in federal district court, provided that the foregoing four criteria are satisfied. Thus, we do not address the questions whether state or federal law governs the giving of the missing witness instruction in a diversity suit and whether federal law permits the missing witness inference to be made. See generally Jones v. Otis Elevator Co., 861 F.2d 655, 659 n. 4 (11th Cir.1988) (questioning the application of a missing witness charge under the Federal Rules of Evidence).
. In their appellate brief, the Wilsons' attorneys listed Exhibits 2257-A, 2258-A, 2258-E and 2258-F as the sales charts that were admitted over their objections at trial. Subsequently, an addendum to the Wilsons’ brief was submitted which stated that Exhibits 2289 and 2299 were erroneously cited as Exhibits 2257-A and 2258-A in the brief. Merrell Dow’s attorneys then made a motion to strike the Wilsons’ addendum on the grounds that neither Exhibit 2289 nor Exhibit 2299 was offered or admitted at trial. After reviewing the trial transcript, we agree that Exhibits 2289 and 2299 were not admitted at trial, and therefore we grant Merrell Dow’s motion to strike the addendum. We also observe that although Richard Smith, one of Merrell Dow’s witnesses, referred to Exhibits 2258-E and 2258-F during his testimony, neither of those exhibits was offered into evidence. We assume that the Wilsons’ attorneys meant to list Exhibits 2288 and 2280 as being the objectionable charts. Merrell Dow’s counsel pointed out that Exhibits 2288 and 2280 contain information that is similar to that in Exhibits 2289 and 2299, and expressly waived any objection to this court’s consideration of the admissibility of Exhibits 2288 and 2280. We thus confine our discussion to the admissibility of Exhibits 2288 and 2280.
. Although the sales charts were admissible under Rule 703 for the limited purpose of providing the basis for expert testimony, the district court did not err by not giving a jury instruction noting the limited admissibility of the charts. Under Federal Rule of Evidence 105, the opponent of the evidence has the burden of requesting that a limiting instruction be given. See, e.g., United States v. Regner, 677 F.2d 754, 757 (9th Cir.), cert. denied, 459 U.S. 911, 103 S.Ct. 220, 74 L.Ed.2d 175 (1982). The Wilsons did not request such an instruction.
. Because we hold that the sales charts were admissible as the basis of expert testimony under Rule 703, we need not decide whether the charts were admissible under any of the hearsay exceptions contained in Federal Rule of Evidence 803. We do note, however, that the information used to create the graphs on the charts came from sources that are recognized as trustworthy under Rule 803. Dr. Goddard determined the number of Bendectin tablets distributed using FDA data that appears to be admissible under the public records exception of Rule 803(8). Doctors Goddard and Lamm calculated the rate of birth defects using information from the CDC, whose data has been ruled to be admissible under Rule 803(8) as well. See, e.g., Ellis v. International Playtex, Inc., 745 F.2d 292, 300-04 (4th Cir. 1984); Kehm v. Proctor & Gamble Mfg. Co., 724 F.2d 613, 618-20 (8th Cir. 1983). Also, the number of new therapy starts, utilized by Dr. Lamm and compiled by Mr. Smith from independent market research firms, appears to be admissible under the market reports exception of Rule 803(17).