OPINION
SEERDEN, Chief Justice.This is a product liability case. Appellees, Marilyn and Ernest Havner, sued Merrell Dow Pharmaceuticals, Inc. (“Merrell”) on behalf of their minor daughter, Kelly. Kelly was bom with no fingers or thumb on her right hand. Appellees alleged, and the jury found, that these birth defects were caused by Mrs. Havner taking the drug, Bendectin, a morning sickness drug manufactured by appellant. Merrell appeals from the judgment of actual and exemplary damages entered in favor of the Havners. Four issues are presented for review: first, whether there is legally or factually sufficient evidence of causation to support the verdict; second, whether the court erred in accepting a verdict when the same ten jurors did not concur on all material issues; third, whether legally and factually sufficient evidence supports the award of punitive damages, and if so, whether such an award violates Merrell’s due process rights under the United States Constitution and the Texas Constitution; and fourth, whether the trial court reversibly erred in admitting various items into evidence. The Havners, by way of cross points, complain of the trial court’s reduction in the damage award and the failure to award prejudgment interest. We reverse and render.
FACTS AND PROCEDURAL HISTORY
Kelly Havner was bom with a limb reduction defect consisting of missing fingers and thumb on her right hand. During the pregnancy with Kelly, Marilyn Havner took the anti-nausea prescription medication, Bendec-tin, which is composed of doxylamine succi-nate (an antihistamine) and pyridoxine hydrochloride (vitamin B-6). This drug was manufactured in the United States from 1956 to 1983 by Merrell.
On Kelly’s behalf, the Havners filed a product liability action against Merrell asserting that the birth defect was caused by Mrs. Havner’s ingestion of Bendectin during pregnancy. The Havners advanced liability theories based on negligence, defective design, and defective marketing. The trial was bifurcated with liability and compensatory damage issues being resolved by the jury prior to the presentation of any evidence on punitive damages. Both phases of the proceeding were presented to the same twelve jurors. Ten jurors assessed liability and actual damages in phase one. In phase two, a different ten of the same twelve jurors found Merrell grossly negligent and assessed punitive damages. Merrell moved for judgment notwithstanding the verdict, to set aside the jury findings, and for a new trial. Those motions were denied. The trial court reduced the punitive damage award to four times the compensation award and entered judgment for the Havners.
“NO EVIDENCE” POINT Liability Standard
In product cases involving either alleged negligence or design and marketing defects, a plaintiff must prove that the defective product was a producing cause of the *540injuries suffered. Morgan v. Compugraphic Corp., 675 S.W.2d 729, 732 (Tex.1984); American Cyanamid Co. v. Frankson, 732 S.W.2d 648, 657 (Tex.App.—Corpus Christi 1987, writ ref'd n.r.e.). Producing cause has been properly defined as an efficient, exciting, contributing cause which, in a natural sequence, produced the injuries complained of. Rourke v. Garza, 530 S.W.2d 794, 801 (Tex.1975); American Cyanamid Co., 732 S.W.2d at 657.
In its first point of error, Merrell asserts that there is no evidence of probative value on the essential element of producing cause. Merrell contends that the case presented by the Havners is fatally defective because there is no evidence that the drug, Bendectin, causes human birth defects in general, or limb reduction defects in particular. Further, no evidence was offered, asserts Mer-rell, that Mrs. Havner’s ingestion of Bendec-tin during her pregnancy was the producing cause of Kelly’s deformed right hand.
Standard of Review
In reviewing a “no evidence” point concerning producing cause of injury, only the evidence and inferences which tend to support the finding of the jury will be considered. All evidence or inferences contrary to the jury finding will be disregarded. Havner v. E-Z Mart Stores, Inc., 825 S.W.2d 456, 458 (Tex.1992); Garza v. Alviar, 395 S.W.2d 821, 823 (Tex.1965). See also Transportation Ins. Co. v. Moriel, 879 S.W.2d 10, 20 (1994). In this ease, therefore, only the evidence and inferences which support the jury’s finding that Bendectin caused Kelly Havner’s birth defect will be considered. If there is any evidence of probative force to support the finding of the jury, the “no evidence” point must be overruled. In re King’s Estate, 150 Tex. 662, 244 S.W.2d 660, 661 (1951). However, “when the evidence offered to prove a vital fact is so weak as to do no more than create a mere surmise or suspicion of its existence, the evidence is no more than a scintilla and, in legal effect, is no evidence.” Browning-Ferris, Inc. v. Reyna, 865 S.W.2d 925, 927 n. 3 (Tex.1993).
Expert Testimony Required
The issue of causation was submitted to the jury, over appellant’s objection. The trier of fact is generally allowed to decide the issue of causation in the following cases: (1) when general experience and common sense will enable a layman to fairly determine the causal relationship between the event and the condition or injury; (2) when there is a scientific principle, a sharp categorical natural law, which theorizes that a result is always directly traceable back to a particular cause;1 and, (3) when the reasonable probability of a causal relationship is shown by expert testimony. Lenger v. Physician’s Gen. Hosp., Inc., 455 S.W.2d 703, 706 (Tex.1970); American Cyanamid Co., 732 S.W.2d at 657. It was undisputed that in a majority of limb reduction birth defects, isolation of the cause has eluded the determined efforts of even the brightest scientific minds. Therefore, in this ease, a lay jury is ill-equipped to determine causation armed only with general experience and common sense. So too, birth defects of this type are not the subject of a sharp categorical natural law, nor are they usually traceable back to a particular cause. Neither is this the usual drug liability case where the product is known to produce certain effects under particular circumstances, and the inquiry is merely whether those circumstances are *541present. Nor is this an instance where the injurious nature of the product has become apparent after the plaintiffs use and the inquiry is whether the manufacturer knew or should have known of the risk at an earlier date. No causal link has been firmly established between Bendectin and human birth defects. Applicable here is the statement made by the United States Court of Appeals for the Second Circuit when it affirmed a summary judgment for the herbicide manufacturer in the Agent Orange litigation:
This is not a case in which a hazard is known to have existed in hindsight and the issue is whether the defendant had sufficient knowledge of it at an earlier time to trigger an obligation to inform. Existence of hazard remains unproven to this date.
In re Agent Orange, 818 F.2d 187, 193 (2d Cir.1987), cert. denied, 487 U.S. 1234, 108 S.Ct. 2898, 101 L.Ed.2d 932. This case, therefore, falls into the third category where causation must be proved, if at all, by scientific expert testimony as to the reasonable probability of that causal link. Contrary to appellees’ assertions, lay testimony is legally insufficient to establish cause in this case. See Lenger, 455 S.W.2d at 706; American Cyanamid Co., 732 S.W.2d at 657.
Reasonable Probability
To constitute proof, the expert testimony must establish the “reasonable probability” of a causal connection between Mrs. Havner’s ingestion of Bendectin and Kelly’s birth defect. See Schaefer v. Texas Employers’ Ins. Ass’n, 612 S.W.2d 199, 202 (Tex.1980). “Reasonable probability” has been defined as: “testimony predictive of what will happen in the future ... results reasonably to be anticipated.” Insurance Co. of N. Am. v. Myers, 411 S.W.2d 710, 713 (Tex.1966).
[ T]his probability must, in equity and justice, be more than coincidence before there can be deemed sufficient proof for the plaintiff to go to the jury ... In the absence of reasonable probability, the inference of causation amounts to no more than conjecture or speculation.
Schaefer, 612 S.W.2d at 202. “Proof of mere possibilities will not support the submission of an issue to the jury.” Duff v. Yelin, 751 S.W.2d 175, 176 (Tex.1988). While reasonable inferences from the evidence are permissible,
a possible cause only becomes “probable” when in the absence of other reasonable causal explanations it becomes more likely than not that the injury was a result of its action. This is the outer limit of inference upon which an issue can be submitted to the jury.
Parker v. Employers Mutual Liab. Ins. Co. of Wis., 440 S.W.2d 43, 47 (Tex.1969) (emphasis ours) (expert testimony that radiation exposure “could have” caused plaintiffs cancer was held to be “no evidence” of causation). While it is “not absolutely necessary that an expert couch his opinion in terms of a ‘reasonable medical probability,’ we still embrace the principle that a jury issue should not be submitted when it is based merely upon speculation and conjecture.” Duff, 751 S.W.2d at 176 (evaluation of causation in medical malpractice action). Additionally, even when an expert expresses his opinion using the magic words, “reasonable probability,” the entire substance of the expert’s testimony must be examined to determine if the opinion is based on demonstrable facts and does not rely solely on assumptions, possibility, speculation and surmise. See Schaefer, 612 S.W.2d at 204-05 (expert testimony that exposure to contaminated soil, to a “reasonable medical probability,” caused the plaintiffs tuberculosis held to be “no evidence” of causation). See also Myers, 411 S.W.2d at 713 (expert testimony held to be “no evidence” that traumatic injury was producing cause of plaintiffs death from cancer); Ralph v. Mr. Paul’s Shoes, Inc., 572 S.W.2d 812, 814 (Tex.Civ.App.—Corpus Christi 1978, writ ref'd n.r.e.).
Our review is, therefore, confined to a consideration of only the evidence and inferences which tend to support the jury finding of causation, ignoring all contrary; but the supporting evidence must have some probative force. See E-Z Mart Stores, 825 S.W.2d at 458; Schaefer, 612 S.W.2d at 204-05. With expert testimony, in order to possess probative force, the opinion must amount to more than suspicion or speculation. See Schaefer, 612 S.W.2d at 204-05. *542While certainty is not required, an expert scientific opinion must be grounded, at the very least, on some demonstrable underlying scientific data or logical inferences therefrom. Tex.R.Civ.Evid. 702, 703, & 705. When the phrase “reasonable medical probability” is used, it -will amount to some evidence only when it represents the overall substance of the expert’s opinion and is based on more than purely speculative conclusions or personal opinion ungrounded in scientific reality. See Schaefer, 612 S.W.2d at 204-05; see also Tex.R.Civ.Evid. 702, 703, & 705. Reasonable probability cannot be created by the mere utterance of magic words by someone designated an expert. Id.
Application to Facts
Mrs. Havner’s physician testified that he prescribed Bendeetin for Mrs. Havner based on the information provided by Merrell. As proof of that information, the Havners had admitted into evidence the page from the 1981 Physician’s Desk Reference (PDR) which revealed the following warning listed for Bendeetin:2
Precautions: Because of potential drowsiness, Bendeetin should be prescribed with caution for patients who must drive automobiles or operate machinery.
Studies in rats and rabbits have revealed no suggestion of drug-induced fetal abnormalities at doses of Bendeetin up to 90 times the maximum human dose. In addition, several epidemiologic studies in women who received Bendeetin during pregnancy have shown that the incidence of birth defects in their offspring is no higher than in women not taking the drug during pregnancy. Nevertheless, like all drugs considered for use during pregnancy, particularly during the first trimester, Ben-deetin should be used only when clearly needed.
The Havners then offered the testimony of five experts to show that the warning issued by Merrell was “false” and that causation existed.3 The Havners’ experts, however, were unable to offer any demonstrable scientific data available to Merrell in 1981, or even available today, which would indicate that a general causal relationship exists between human birth defects and Bendeetin administered at the normal human therapeutic dose.4 Two opinions were based on scientific data concerning test-tube analysis and chemical composition analogies which, in the absence of animal and human studies, might raise a suspicion that Bendeetin was a human terato-gen.5 The opinions of the other three experts consisted of criticism of the animal studies and the multitude of human epidemi-ologic studies which all indicate that, at normal doses, there is no statistically significant association between Bendeetin and human birth defects.
Without any direct human scientific data to substantiate their theory of causation, the Havners’ experts merely attempted to refute the methodology employed and the conclusions reached in the live animal and human epidemiologic studies relied on by Merrell. They inferred the existence of causation by asserting that proof of the contrary does not exist. Indulging the Havners’ experts to the outer limits of inference in their claim that all the animal studies and all the human epide-miologic studies are wrong, the most we can conclude is that Merrell has no conclusive evidence that Bendeetin never causes birth *543defects in humans and that the warning issued in the 1981 PDR might be inaccurate. However, absence of evidence of a fact, standing alone, does not amount to some evidence of its converse.6 Absence of absolute proof that Bendectin does not cause birth defects in humans, is not some evidence that it does.
In this case, the burden of proof operates in favor of Merrell, the defendant below. The Havners’ defective marketing claim (failure to warn) does not alter this burden. Merrell was not required to prove lack of causation. “A manufacturer who knows or should know of potential harm to a user because of the nature of a product is required to give an adequate warning of such dangers.” American Cyanamid, 732 S.W.2d at 656 (citing to Bristol-Myers Co. v. Gonzales, 561 S.W.2d 801 (Tex.1978)). Implicit in this rule is the presumption that the product causes some harm. Before the adequacy of the warning can even be evaluated, the Havners must produce some evidence that Bendectin is harmful when ingested by humans at normal dosage levels. Merely showing that the warning was inaccurate, as it related to animal studies, is not the same as showing that it inadequately warned of a potential for human harm. Mere inaccuracy, without harm, cannot render a warning tor-tious. Therefore, to prevail on any theory that they advanced, even the failure to warn claim, the Havners were required to offer at least some evidence that Bendectin, more likely than not (to a reasonable probability), caused Kelly’s defect.
This requirement [scientific testimony of causation to a reasonable probability] does in some instances place extraordinary burdens of proof on claimants. But once the theory of causation leaves the realm of lay knowledge for esoteric scientific theories, the scientific theory must be more than a possibility to the scientists who created it. For to the scientific mind, all things are possible. And with all things possible, citizens would have no reasoned protection from the speculations of courts and juries.
Parker, 440 S.W.2d at 49 (citations omitted), Having scrutinized the record, we hold that, despite the use of magic words by the Hav-ners’ experts, there is no evidence of that causal link. The substance of the Havners’ expert testimony is as follows:
1) Dr. Jay Glasser
Dr. Jay Glasser, a Ph.D. in experimental statistics, confined his testimony to a review of the human epidemiologic studies conduct ed with Bendectin by other researchers. Epidemiology measures the statistical relationship, if any, between exposure to an agent, such as Bendectin, and the incidence of a condition, such as birth defects, in a scientifically selected sample population. Depending on the methodology employed, the incidence of the condition in exposed cases is compared with either non-exposed cases exhibiting the same condition, or with exposed cases who do not exhibit the condition. The result is stated as a risk or odds ratio, which is reported with reference to 1.0, or equal risk. This ratio is then converted to a “confidence interval”, a measure which compensates for variables and randomness. Dr. Glasser explained these concepts of relative risk, odds ratio, confidence interval, and statistical significance as they relate to human epidemiology. He stressed that, in investigating the connection between a drug and adverse effects, there is an important difference between “lack of evidence of an association” and “evidence of no association.” He divided the conclusions inferable from epidemiologic studies into three zones, depending on whether the confidence interval includes 1.0, or equal risk. When the entire confidence interval falls below 1.0, the inference is that a drug has a protective effect. When the entire confidence interval falls above 1.0, the inference is that a drug has an adverse effect. When the confidence interval includes 1.0, the study is considered “inconclusive,” “no evidence of anything,” in the “zona obscura.”7
*544Dr. Glasser then testified concerning the more than twenty-one human epidemiologic studies that have been conducted with Ben-dectin. After stating the methodologic faults of all the reported studies, Dr. Glasser notes that all but one are “inconclusive” by virtue of the fact that the confidence interval of each encompasses 1.0. Dr. Glasser admitted that the only study which showed an adverse effect (the Rothman Study) utilized the old three-component Bendeetin, an entirely different drug than that ingested by Mrs. Hav-ner.
On the basis of the very studies he termed “inconclusive,” Dr. Glasser then opined that:
more likely than not, Bendeetin is an associative factor in raising birth defects.... [T]here is an association more likely than not between Bendeetin and limb reductions _I believe that there is an association between — from the sum and substance of the literature that I reviewed and from my point of expertise, that it has an association of Bendeetin with congenital disorders.
His own testimony defines the parameters of the word “association” in this context. Dr. Glasser stated: “Association simply means that one thing is related to the other. But we cannot claim that one thing causes the other.”8 Dr. Glasser revealed the reasoning he employed in finding an association from his review of “inconclusive” studies:
[ B]ecause my background is in public health, as I say, being a faculty member of two schools of public health and being trained in three, that one of the hallmarks of public health is guarding the safety and decision as to whether something is safe or efficacious. You do look at the balance of the data and you always look on the side of safety. And that is again, in trying to reach a judgment, I look beyond what is strictly a statistical interpretation on a one-at-a-time confidence interval or relative risk finding and try to see the direction ...
Because of his public health orientation, Dr. Glasser rejected the very principles, grounded in statistics and epidemiology, that he painstakingly explained to the jury. He “refuted” demonstrable scientific data, not with contradictory scientific data, but with a conclusion based on his personal feeling that it is better to err on the side of safety. Despite his own profession that epidemiologic studies are unable to prove “no association,” proof is precisely what Dr. Glasser would require.
Taken at its best, Dr. Glasser’s testimony might be some evidence that Merrell is unable to prove conclusively that there is no link, no association, no relation between Ben-deetin and human birth defects; but his opinion is not some evidence that a causal link does exist.
2) Dr. Adrian Gross
Dr. Adrian Gross, a veterinarian and former scientific advisor for the Food and Drug Administration (FDA), testified on the basis of animal studies conducted from the early 1960’s through the early 1980’s by both Mer-rell and independent researchers. The studies were conducted on rats, rabbits, monkeys, and baboons, and utilized dosages from two to 2400 times the equivalent human dose. The various researchers concluded generally that at dosages of 50 to 100 times the recommended human dose, Bendeetin caused no effect on developing animal embryos. Dr. Gross reviewed each study and asserted why each was “useless” or “inappropriate” in determining whether Bendeetin was a terato-gen. While all the actual researchers were *545in full concurrence that harmful effects were exhibited only at levels far in excess of the human dose, Dr. Gross disagreed. On the basis of a “data-audit” of each study, he pointed out what he considered discrepancies between the raw data and the summary report, and without any further reason, explanation, or scientific data in support of his position, he concluded that, regardless of dosage, Bendeetin is teratogenic in animals:
Q. Now, Doctor, if I can just read one of those statements, on page two, and go into it a little bit more, then. It says: “In the present case doxylamine succinate looks like ... a clear-cut teratogen in Wistar rats of 100 milligrams per kilogram per day and higher. Fifty milligrams per kilogram are to declare as teratogenic, and of no effect level for rats.”
A. Yes, sir.
Q. Well, first of all, would you agree with the part about being a clear-cut teratogen?
A. Yes, that I certainly agree with.
Q. And, Doctor, from your analysis, then, as you went through it, did you look at the vast numbers of animals that they used?
A. Yes.
Q. And what is your opinion?
A. My view is that — Well, this was highly significant teratogenic in Wistar rats, but it’s the last sentence that I take exception. “50 milligrams per kilogram are to declare as teratogenic no effect level for rats.” What they’re saying, in other words, is that 50 milligrams per kilogram did not seem to have an effect. That’s what I don’t agree with.... I would say that it’s a clear-cut teratogen, period. The higher the dose that one gives, the more noticeable these effects.
Based on the same “data-audit” of the same animal studies, Dr. Gross then extended his opinion to humans as well: “My opinion is that Bendeetin is teratogenic, period, meaning to animals and/or humans.” On the basis of his disagreement with the conclusions of the authors of the various studies, Dr. Gross opined to the jury that the product information listed for Bendeetin in the 1981 PDR was “completely false.”
Dr. Gross agreed that when a drug is available by prescription only, there is risk involved, especially at doses higher than recommended. He admitted that the dosages utilized in the studies that revealed animal birth defects would be equivalent to a human taking 1200 to 4800 Bendeetin tablets at one time. He revealed that:
A What the Food and Drug Administration requires, as a matter of policy, is that tests in animals should be carried out at such doses until they become — the drug is clearly toxic. It kills the animal. So that it becomes sick in some way, and then to back off from that at regular intervals.
Q. Okay.
A. So a test that — in which you see nothing at all, on its face is not a very useful test because the implication is that you haven’t gone high enough.
Despite this FDA requirement, Dr. Gross uses those same ultra-high dosage animal studies to infer human teratogenicity without any qualification as to dosage level. Even under his reanalysis, he offered no evidence of adverse effects at dosage levels at or even near the recommended human dose. In arriving at his opinion, he indicated that human epidemiology studies with Bendeetin were of no concern to him. In substance, his opinion was not grounded in any demonstrable scientific data. It was based merely on criticism of multiple animal studies which all substantiated Merrell’s position. Indulging in every inference possible from his “discrepancy” theory, we are still left with only a possibly inaccurate product precaution regarding the ■ animal studies, but with no underlying data relating either to defects in animals at lower dosages or to causation in humans.
3) Dr. Stuart Newman
Dr. Stuart Newman, a Ph.D. in chemistry, testified on the basis of four in vitro9 studies done with Bendeetin. He initially instructed *546the jury on the general aspects of human pre-natal development, and then focused on the period of limb development. Dr. Newman related the results of experiments done by others on the isolated limb bud cells of chickens and mice. When doxylamine succi-nate was applied directly to the cells, the development of cartilage was inhibited.10 When asked to explain, Dr. Newman stated:
What it means is that, if this substance is present in a developing embryo in such a way that it can get to the cells of the developing embryo, it will cause an impairment or an inhibition of the process that I described before, chondrogenesis11 of the limbs.
Dr. Newman opined that “doxylamine is a teratogen in humans.”12 He stated that the defect exhibited by Kelly was consistent with the type of cartilage impairment reported in the studies and could have been caused by a teratogen ingested during the period of limb development. Dr. Newman then stated that Mrs. Havner’s ingestion of Bendectin coincided with the critical period for Kelly’s limb development. From a review of her medical records he could not identify any other potentially teratogenic substance ingested by Mrs. Havner during this critical period.13
Direct and cross-examination of Dr. Newman revealed that the following assumptions are required in order to infer human terato-genicity from in vitro animal cell studies:
1) that human cells react exactly the same as animal cells;
2) that the human placental barrier is not impervious to doxylamine;
3) that a living human being does not metabolize the chemical (which would render it partially or completely inert); and
4) that the concentrations of the chemical in human fetal tissue are at least equal to that found in the mother.
In an attempt to establish that Bendectin crosses the placenta, Dr. Newman testified that doxylamine succinate was a type of antihistamine, and that studies in the 1950’s revealed that some antihistamines caused abortions or resorptions of pregnancy in rats and mice. However, he admitted that the antihistamines studied differed in chemical structure from doxylamine succinate.
To establish concentration in the fetus, Dr. Newman cited studies in mice showing concentrations of doxylamine to be higher in fetal tissue than in the mother. Indulging every inference in his favor, we still note the total absence of any evidence that human cells react the same as animal cells or that a live human would not metabolize the chemical.
Dr. Newman admitted that all animals do not react identically to the same chemicals, and that animals and humans do not necessarily metabolize chemicals in the same way. He testified that the chromosomal makeup of animals and humans varies widely which would predispose to wide variation in reactions and metabolism. He revealed that although in vitro testing utilizing human cells is available, none has been performed with Bendectin. Neither have any metabolism studies been performed despite the millions of women who have ingested Bendectin dur-*547mg pregnancy. Even assuming the validity of the animal in vitro studies upon which Dr. Newman relies, his own admissions reveal that there is no evidence relating to two of the four underlying assumptions absolutely necessary to his inference. The studies, therefore, raise no more than a possibility that Bendeetin could be a human teratogen. Because Dr. Newman’s opinions rely solely on these mere possibilities, his testimony lacks the probative force necessary to consider it “some evidence” of causation.
tí Dr. Shanna Swan
Dr. Shanna Swan, a Ph.D. in statistics and chief of the Reproductive Epidemiology Program of the State of California, testified as a rebuttal witness to counter data introduced by Merrell from the numerous human epide-miologic studies which all fail to establish a statistically significant link between Bendec-tin ingestion and human birth defects. Her testimony was similar to that of Dr. Glasser, and lacks probative value for the same reasons. She has never conducted a Bendeetin epidemiologic study herself, nor can she even testify to one conducted by someone else that supports her conclusions; but Dr. Swan opines that there is a probable teratogenic risk to a child born after exposure to Bendec-tin during gestation and that “more probably than not [Bendeetin] is associated with limb reduction defects.” In support of her hypothesis, Dr. Swan recounted all the possible faults of the human epidemiologic studies which refute her hypothesis. She concludes that all the studies are flawed in some manner and therefore cannot serve to prove that Bendeetin is free from human teratogenic effect. She recites that the methods of pooling and meta-analysis of data utilized by Merrell to refute causation are “megasilliness,” although she admits to having used the same techniques herself with other products. She does not refute with demonstrable scientific data; she merely injects doubt as to the value and conclusiveness of Merrell’s evidence. Taking Dr. Swan’s testimony to the outer limits of inference, we are left perhaps to wonder how irrefutable Merrell’s proof is, but we are presented with not one scintilla of evidence that a causal link actually exists between Bendeetin and human birth defects.
5) Dr. John Palmer
The Havners’ principal causation expert, Dr. John Palmer, has both a medical degree in internal medicine and a doctorate in pharmacology. His scientific area of concentration is a study of the effectiveness of drugs to combat allergies. Dr. Palmer’s only examination looked at Kelly’s hand the day before trial. He was unable to identify the type or pattern of Kelly’s defect, could make no observations from her x-rays, and admitted that, except for serving as an expert witness in other Bendeetin litigation, he has never studied nor diagnosed the cause of a human birth defect. Dr. Palmer was unaware and unconcerned with the dosage, timing, or number of Bendeetin tablets ingested by Mrs. Havner. However, Dr. Palmer testified as a purported expert not only in pharmacology, but also in toxicology, clinical teratology, chemical structure activity analysis, in vitro studies, animal teratology studies, human fetal development, and genetics. He advanced the critical opinion that Bendeetin not only generally causes human birth defects, but that it specifically caused Kelly’s limb reduction defect. Dr. Palmer based his opinion partially on the same in vitro studies, animal studies and chemical analogy studies utilized by the other experts called by the Havners. As outlined above, a finding of even general causation, in reliance on these existing studies, amounts to no more than pure speculation. Additionally, Dr. Palmer relied on Drug Experience Reports (DERs).14 In 1972, Merrell reviewed 96 DERs that it received concerning birth defects in children of the millions of Bendeetin exposed mothers. DERs report the concurrence of an effect with the use of a drug, and are filed by doctors, lawyers and other individuals. Dr. Palmer admitted that the reports are voluntary and “unsubstantiated,” that they lack scientific methodology and controls, and that they do not suggest causation. However, on this tenuous basis, despite the admitted fact that no scientist who has actually researched the hypothesized relationship between Ben-*548dectin and birth defects has concluded that there is a causal relationship, Dr. Palmer opined specific causation in this case. His opinion is wholly without any foundation or underlying scientific support, and is, in essence, no more than personal lay opinion and surmise disguised as scientific dogma.
Battle of Experts
The Havners argue that this is a classic battle of the experts, and, therefore, our holding in American Cyanamid controls here. See 732 S.W.2d at 648. We note that in American Cyanamid, another drug product liability action, there was underlying, scientifically based, opinion testimony on both sides of the controversy. Id. at 668. In American Cyanamid, experts for both parties relied principally on the research and reports of others, but that underlying research exposed opinions supportive of both causation and lack of causation. Id. The primary researchers had reached differing conclusions. Id. at 656-57. Such is not the case here. All the primary researchers who have studied Bendectin have reached but one conclusion, and it does not support the theory postulated by the Havners’ experts. Casting unsubstantiated aspersions on the work of others cannot be considered a substitute for real scientific data. Our deference to the jury determination in the American Cyanamid case, a true battle of the experts, is inapplicable here.
The Havners have failed to bring forward anything more than suspicion on the essential element of causation. While Dr. Palmer attempted to piece the suspicions together into a mosaic amounting to something more, it remains a basic premise that:
some suspicion linked to other suspicion produces only more suspicion, which is not the same as some evidence- Our system of justice is designed to ensure that our fundamental right of trial by jury does not become some mere game of chance.... We are not empowered to convert some evidence into no evidence. At the same time ... we are not empowered to convert mere suspicion or surmise into some evidence. Where there is real evidence, we must uphold the jury verdict, but in cases such as this where there is only real suspicion, we must overturn it.
Browning-Ferris, Inc., 865 S.W.2d at 928 (reversing a jury verdict and rendering judgment for the defendant on a tortious interference claim). Here there is only real suspicion. We sustain Merrell’s legal insufficiency point.
Having sustained on this ground, it is unnecessary to reach Merrell’s other points of error, or the Havners’ cross-points. Tex. R.App.P. 90(a).
We, therefore, REVERSE the judgment of the trial court, and RENDER that the Hav-ners take nothing by this suit.
Before the court en banc.
OPINION
DORSEY, Justice.This is an appeal from a products liability suit against a drug company, Merrell Dow Pharmaceuticals. Merrell Dow manufactured and marketed Bendectin, a drug to relieve the nausea associated with pregnancy. The Havner family claims that the drug caused their daughter’s birth defect. The plaintiffs obtained a favorable jury verdict and judgment. Merrell Dow appealed. On original submission a panel of this court found no evidence to support causation, and reversed and rendered the judgment.1 We granted rehearing en banc, heard oral argument before the entire court,2 and now AFFIRM the judgment below in part and REVERSE in part. We REVERSE that part of the judgment that awards punitive damages, but otherwise AFFIRM.
I. Procedural History
Marilyn and Ernest Havner sued Merrell Dow Pharmaceuticals and Dr. David Bruce *549on behalf of their daughter, Kelly.3 Kelly was bom with a defect of her right hand; her fingers are not fully formed. During her pregnancy, Marilyn Havner took Bendeetin prescribed by Dr. Bruce. The Havners claim that the Bendeetin caused Kelly’s birth defect.
The case against Merrell Dow was tried for four weeks in September 1991. The jury found that Merrell Dow was negligent, that its negligence proximately caused Kelly’s birth defect, that Bendeetin was defectively designed and marketed, and it awarded $3.75 million in actual damages. The verdict was signed by ten jurors.
The punitive damages portion of the trial was bifurcated. The jury found Merrell Dow to be grossly negligent and awarded $30 million in punitive damages. The verdict was not unanimous and ten different jurors signed the punitive damages charge.
Judgment was entered June 1, 1992, awarding $3.75 million in actual damages, prejudgment interest in the amount of $15 million, and $20,205,821.80 in punitive damages. The punitive damages were reduced from the verdict amount pursuant to section 41.007 of the Civil Practice and Remedies Code. It is from this judgment that Merrell Dow appeals.
Merrell brings four categories of complaints: legal and factual sufficiency of the evidence of scientific causation; that the same ten jurors did not find both liability and actual damages that found gross negligence; a challenge to the award of punitive damages; and several evidentiary challenges.
II. Legal Sufficiency of the Evidence
A. Standard of Review
In reviewing a “no evidence” point of error, we consider the evidence and reasonable inferences therefrom in the light most favorable to the jury’s finding and we disregard all contrary evidence and inferences. Havner v. E-Z Mart Stores, Inc., 825 S.W.2d 456, 458 (Tex.1992); Garza v. Alviar, 395 S.W.2d 821, 823 (Tex.1965). Merrell claims that there is no evidence to support the jury’s finding of causation between Bendeetin and Kelly Havner’s birth defect. Although the Havners put on extensive scientific evidence in support of their claims, Merrell contends that the evidence offered was not reliable, and thus not helpful to the jury. Even if the challenged evidence was admissible, Merrell contends that it is in effect “no evidence” because it is too weak to do more than create a suspicion of causation. Merrell claims that the Havners’ experts did no more than mouth the “magic words” of reasonable medical or scientific probability and that there is no basis in the evidence to support their stated conclusions.
The Havners attempted to do two things with their experts: 1) establish that Bendee-tin caused Kelly’s birth defect and 2) defuse Merrell’s defense that Bendeetin did not cause Kelly’s birth defect because it does not cause birth defects generally. In attempting to prove their first theory, the Havners produced evidence that limb reduction defects are rare, approximately 4 defects for every 10,000 live births.
According to the Havners’ experts, for Bendeetin to cause a limb reduction defect it must be ingested during a twenty day period within the first trimester when the fetal skeleton is developing. This window of vulnerability is between 30 and 50 days after conception. According to her testimony and the medical records, Mrs. Havner was prescribed Bendeetin and took it for her pregnancy induced nausea. Dr. Newman calculated that she took Bendeetin dining that window of vulnerability. The prescribed dosage was one pill daily for twenty days.
Several of the Havners’ experts reanalyzed data from studies performed by other scientists and concluded that animal and epidemiological studies demonstrated that there was a relationship between Bendeetin exposure and limb reduction and other birth defects. These conclusions differed from that of the authors of the studies. In some cases, the reanalysis was performed to focus on a portion of the study that dealt with Bendeetin although much of the other data did not. In other instances, the reanalysis focused on *550claimed discrepancies between the author’s own raw data and his conclusions.
To rebut Merrell’s defense prophylactically, the Havners’ experts extensively criticized the studies that Merrell contends prove that Bendectin is safe. Some criticism was leveled at the objectivity of some of Merrell’s in-house researchers or other researchers who received funding from Merrell or other drug companies. Other criticisms involved sample size, use of inappropriate control groups, failure to properly classify participants as exposed or nonexposed, and other methodological concerns in both the animal and epidemiological studies.
Merrell contends that criticism of existing studies which find no link between Bendectin and birth defects does not equal causation and we agree. But Merrell also claims that reanalysis of data and the resulting opinion of a scientist based on an evaluation of the existing data and that reanalysis is, as a matter of law, insufficient to support a finding of causation in any case. With that conclusion we differ.
B. Admissibility of Scientific Evidence
The trial court admitted the testimony of plaintiffs’ experts at trial after numerous objections by Merrell. Merrell filed numerous pretrial motions seeking to exclude plaintiffs experts and objected to the admission of their testimony at trial. Additionally, Mer-rell filed a motion for a directed verdict on the issue of causation and a motion for judgment nonobstante veredicto on the same issue. From January 1991 through judgment in June 1992, the trial court had numerous opportunities to consider and reconsider the admissibility of plaintiffs’ experts’ testimony.
1. Standards of Admissibility
Expert testimony is admissible if it is helpful to the jury, is scientific, technical or other specialized knowledge, and if the expert is qualified by his knowledge, skill, experience, training, or education. Tex.R.Civ. Evid. 702; E.I. du Pont De Nemours and Co., Inc. v. Robinson, 38 Tex.Sup.Ct.J. 852, 858, - S.W.2d -, - [1995 WL 359024] (June 15, 1995). In interpreting that rule, our supreme court has also determined that the expert’s testimony must be relevant and rehable. DuPont, 38 Tex.Sup.Ct.J. at 858, - S.W.2d at -. The trial court’s responsibility is to make the preliminary determination of whether the testimony meets the standards set out by the court in DuPont. The factors established by the court as part of the reliability inquiry include, but are not limited to:
(1) the extent to which the theory has been or can be tested;
(2) the extent to which the technique relies upon the subjective interpretation of the expert ...;
(3) whether the theory has been subjected to peer review and/or publication;
(4) the technique’s potential rate of error;
(5) whether the underlying theory or technique has been generally accepted as valid by the relevant scientific community; and
(6) the non-judicial uses which have been made of the theory or technique.
Id. Trial courts may consider other factors which are helpful to determining the reliability of the scientific evidence and those factors will differ with each potential case. Id. The DuPont court notes that an expert who testifies based on his own primary research conducted independent of litigation “‘provides important, objective proof that the research comports with the dictates of good science.’ ” Id. at 858 n. 2, at - n. 2 (quoting Daubert v. Merrell Dow Pharmaceuticals, Inc., 43 F.3d 1311, 1317 (9th Cir.1995)).
Once the trial court determines that the evidence is relevant and rehable, the court must then decide whether to exclude the evidence because its probative value is outweighed by the danger of unfair prejudice, confusion of the issues, its potential to mislead the jury, or other issues addressed by Rule 403. Id. at 859, at - (citing Tex.R.Civ.Evid. 403).
2. Application
The trial court considered Merrell’s pretrial motions to exclude the Havners’ experts and held two separate hearings on the issue. We are mindful that in moving for summary judgment and seeking to exclude the Havners’ experts pretrial, the burden *551was on Merrell. This differs from the burden established in DuPont which is on the proponent to demonstrate the admissibility of its proffered evidence. DuPont, 38 Tex. S.Ct.J. at 859, - S.W.2d at -.
The first hearing was in conjunction with MerreE’s motion for summary judgment. In that motion, MerreE contended that there was no proof of a causal connection between Havner’s ingestion of Bendeetin and Kelly’s birth defect. MerreE urged the court to reject the plaintiffs’ experts’ reanalysis of existing epidemiological studies,4 and to reject evidence of animal5 and in vitro studies.6 The court held a hearing on MerreE’s motion and denied it.
Later in the case, MerreE filed motions in limine seeking to exclude the testimony of Shanna Swan, Ph.D., Dr. Gross, and any other evidence of causation based on the Havners’ experts’ conclusions. Again, the court held a hearing during which counsel discussed the testimony to be offered, Mer-reE’s objections thereto, and the Havners’ basis for its admission. There was no evidence taken at that hearing, but the parties discussed in detaE the testimony to be ehcit-ed. During that same hearing, counsel for MerreE referred back to the hearing on its motion for summary judgment which covered some of the same issues.
AdditionaEy, during trial, the court held mini-hearings, outside the jury’s presence, on the issue of the various witnesses’ qualifications and basis for their opinions.
In deciding to admit the testimony of ap-peEees’ experts, the trial court made an initial determination that the experts were qualified for the purposes of Rule 702. Tex. R.Civ.Evid. 702. The rule requires both that expert testimony be helpful to the jury and that an expert be qualified by knowledge, skiE, experience, training or education.
Each of appellees’ experts was either a medical doctor or held a doctorate in a scientific area and had other specialized training or experience. Each met the criteria of specialized education or training. Mer-reE complains that appeEees’ experts aE had the wrong backgrounds to testify about Ben-deetin because they were not primary researchers in the field, had not individuaEy conducted tests on Bendeetin, and did not treat patients with birth defects. This restrictive standard is not encompassed in our rules, nor in the case law interpreting those rules. Each of appeEees’ experts worked in the area of statistical analysis, fetal development, teratology, pharmacology, or epidemiology, and most taught subjects related to the inquiry here at medical schools. Each was qualified to testify under the second part of the test estabEshed by Rule 702.
Rule 702 also requires that expert testimony be helpful to the trier of fact. No one disputes that in cases such as this specialized knowledge is necessary to assist the fact-finder. A jury cannot determine scientific or medical causation in this type of case without the assistance of experts. See Lenger v. Physician’s Gen. Hosp., Inc., 455 S.W.2d 703, 708 (Tex.1970) (medical testimony required to establish causation in medical malpractice case where cause not readüy ascertainable from general experience and common sense); Hernandez v. Texas Employers Ins. Ass’n, 783 S.W.2d 250, 253 (Tex.App.—Corpus Christi 1989, no writ) (reverses? jury finding of causation in absence of medical testimony as to causation of asthma).
To be helpful, scientific evidence must also be relevant and reliable. See DuPont, 38 Tex.Sup.Ct.J. at 858, - S.W.2d at -; see also Daubert v. Merrell Dow Pharmaceuticals, Inc., - U.S. -, -, 113 S.Ct. *5522786, 2797, 125 L.Ed.2d 469 (1993) (interpreting parallel Federal Rules of Evidence).
We review the trial court’s assessment of the relevance and reliability of scientific evidence and its decision to admit or exclude the evidence for an abuse of discretion. DuPont, 38 Tex.Sup.Ct.J. at 860, - S.W.2d at -.
a. Relevance
First, we consider whether the proffered testimony was relevant. The Havners’ experts testified about the process of limb development, the time during gestation when that development can be interrupted, what can cause limb development to be interrupted, and why the experts believed that in Kelly Havner’s case Bendeetin caused her birth defect. All of this testimony is directly related to the determination of the Havners’ cause of action. See Tex.R.Civ.Evid. 401. The evidence satisfies the relevancy test.
b. Reliability
Merrell urges that appellees’ experts could not be helpful to the jury because their opinions were not based on sound scientific principles, but instead were “junk science” and thus unreliable.
The inquiry established in DuPont requires that we now consider whether the trial court could reasonably found the evidence to be reliable under the DuPont standards. DuPont, 38 Tex.Sup.Ct.J. at 860, - S.W.2d at -.
The theory that limb development occurs during a short period of time during conception is well established and accepted by experts for both sides. The theory that limb development can be impaired by exposure to teratogens7 during that period is also well accepted. Merrell denies that Bendeetin is a teratogen in humans, but Dr. Gross testified that Bendeetin appears on the United States Government’s list of human teratogens. Whether Bendeetin is a human teratogen has been the subject of scientific inquiry for several decades. The general consensus from epidemiological studies is that Bendeetin is not a human teratogen and does not cause limb reduction defects. However, even Mer-rell’s experts agreed that biological causation is inherently difficult to assess based on epidemiological studies. They agree that they could not say that thalidomide caused limb reduction defects based solely on epidemiological studies, although the scientific community accepts that it does.
The extent to which an expert’s techniques rely on his subjective interpretation is part of the DuPont inquiry. Id. at 858, at -. The statistical techniques used by the Havners’ experts in examining the epidemiological data does not rely on the subjective interpretation of the expert. Experts for both sides used the same statistical techniques. Evaluation of the effect of Bendeetin and its components in laboratory studies does not depend on the subjective interpretation of experts, although there is a dispute in the scientific community over the proper classification of some observations of the animal fetuses at autopsy.
Publication of the theories espoused by an expert at trial is another factor to consider in determining admissibility of expert opinion. The Havners’ experts generally are not published on the subject of Bendeetin and its role in limb reduction defects. Neither are Merrell’s experts. Dr. Newman, a Havner expert, has published in the general area of limb development. Other Havner experts have published in the areas of fetal development and statistical analysis of epidemiological data. Another Havner expert, Dr. Swan, published an abstract of research involving Bendeetin.
Of Merrell’s experts, Dr. Greenberg has published generally on the subject of birth defects, genetic disorders and teratology; Dr. Friedman has published articles on clinical teratology, and has published a summary of the published studies on Bendeetin concluding that it is not a human teratogen. None of Merrell’s witness have published studies on Bendeetin other than the sum-*553raary Friedman prepared in connection with the database he operates on teratogens.
Another factor is the rate of error of the techniques used by the experts. When considering the potential rate of error of the epidemiological studies, there is a wide variance among the studies. But all the experts use and rely on the same universe of material although their interpretations of the meaning of the data differ. The studies themselves indicate their confidence intervals and whether their results are statistically significant. The animal and in vitro studies likewise indicate the scientific limit of their applicability based on the size of the studies and the statistical significance of the results.
The underlying theory that Bendectin causes birth defects has been investigated for many years by scientists all over the world. Although none of the studies have concluded that it does, the problem of limb reduction defects occurring by themselves and not as part of another cluster of birth abnormalities is very rare. To detect a change in the rate of such defects requires studies on a very large scale and very specific information as to when during the pregnancy Bendectin exposure occurred. None of the studies, even using statistical techniques to group them, has the requisite size or detail as to specific exposure.
The various studies used by both parties to this litigation were generally not prepared for litigation. The reanalysis of the raw data performed by the Havners’ experts was prepared for litigation as was the analysis of existing studies performed by Merrell’s experts.
Another consideration in determining reliability has been whether the expert relies on those things typically relied on by experts of the same type, whether he used methodologies used by other experts of the same type, and whether the testimony is within his area of expertise. See Daubert, - U.S. at ---, 113 S.Ct. at 2796-98. The Hav-ners’ experts relied on data developed by other scientists, as did Merrell’s experts. The reanalysis that several of the Havners’ experts did is a methodology that the Food and Drug Administration and the Environmental Protection Agency use to verify the reports they get from the industries they regulate. They separately analyze the data submitted as part of their regulatory function. The Havners’ experts all testified in areas in which they had expertise. These considerations weigh in favor of admissibility.
It is critical to distinguish the search for truth in law and in science. The purpose of any legal inquiry is to resolve a dispute between the parties. The law is a body of rules that are applied to settle the issue, whether they be rules of substantive or procedural law. The rules may well impede evidence of the truth, such as the statute of frauds prohibiting testimony of an oral agreement between the parties. But these shortcomings and compromises have been accepted when put in balance with other considerations.
We try to achieve justice through the application of the law. In the quest for justice, we say we are engaged in a search for truth, and we accept the jury’s verdict as true except in very limited circumstances. This determination by citizens of the crucial facts of the case is the bedrock of our jurisprudence, however wrong its conclusion is in the eye of God or objective reality. We accept the legitimacy of the verdict if the rules of evidence and procedure have been followed because this is the way our history has accepted as the best way for determining truth in the courtroom, especially when contrasted with the earlier methods of oath, ordeal or combat. What the jury finds may not be really true, but we will accept it as true in order to resolve the dispute.
That one jury finds that a component of an automobile was negligently designed or manufactured is not a universal declaration that all of those same components were negligently designed or manufactured. Other plaintiffs with the same complaint will be put to their proof before different juries which may reach contrary results. Our jurisprudence accepts inconsistent results and restricts the law of the case to that particular case with those plaintiffs and defendants.
Science searches for universal truths in order to understand the workings and composition of our world and ourselves. Its *554methods are continuous testing, probing, examining, and re-examining data; always comparing empirical findings to a given hypothesis to see if the hypothesis is correct. Many theories that were accepted as true 100 years ago have been so changed as to be scarcely recognizable, from our ideas about the culture of the Mayas to particle physics. This search and constant revision will continue into the future as the scientific method is applied by researchers.
Science is not separated into two worlds: true and junk. Reasonable experts may arrive at differing conclusions given the same data. The law has always dealt with this, from the trials of Galileo and Scopes to the ordinary low-back pain collision case. The purpose of the trial is to resolve the dispute, not to find universal truths.
3. Conclusion
We review the trial court’s decision to admit expert testimony for an abuse of discretion. DuPont, 38 Tex.Sup.Ct.J. at 860, - S.W.2d at -. “A reviewing court cannot conclude that a trial court abused its discretion if, in the same circumstances, it would have ruled differently or if the trial court committed a mere error in judgment.” Id. Only when a trial court acts without references to guiding rules or principles, do we conclude that it abused its discretion. Id. Considering all the factors enumerated by the court in DuPont, we cannot say that the trial court here abused its discretion in deciding to admit the testimony of the Havners’ experts.
C. Scientific Evidence to Support Causation
The Havners’ experts testified based on three kinds of studies: animal studies, in vitro studies, and epidemiological studies. All of the studies investigated whether Bendectin is a teratogen. Teratolo-gy is the study of teratogens, which are agents that cause birth defects. A summary of the relevant testimony follows.
1. Adrian Gross
Adrian Gross currently works for the Environmental Protection Agency. He worked for the Federal Drug Administration for fifteen years. He testified that in his work he routinely examined the raw data supplied by drug and pesticide companies, analyzed it, and compared the results of his own analysis to the reports generated by the drug and pesticide companies in order to determine the validity and reliability of the conclusions drawn by the primary researchers. Gross holds a Doctorate of Veterinary Medicine and a Masters of Science in pathology. He has post-graduate training in statistics and microbiology and has had additional training in statistics and biometry. He has testified before Congress regarding the safety of drugs and pesticides and has published in the areas of animal pathology and statistics. The EPA relies solely on animal studies to determine the effects of chemicals; the FDA relies heavily on animal studies in the early stages of drug approval. Animal responses to chemicals (drugs or pesticides) are considered predictive of human response.
Gross testified that in his analysis of the animal studies on Bendectin and its component doxylamine succinate,8 he used the same methodology that he has used throughout his career and that the methodology is generally accepted. He also testified that his opinions were based on a reasonable degree of scientific certainty. Additionally, he noted that Bendectin was on a 1987 governmental list of developmental toxicants. Developmental toxicants are teratogens. Gross’s use of reanalyzed data was not attacked in the testimony of the other experts, nor was it demonstrated that his techniques were unacceptable or disapproved. His conclusions, that Bendectin is an animal and human teratogen that causes limb reduction defects, were challenged by Merrell’s experts, but only Mer-rell’s counsel attacked his methodology. With that background, we cannot say that Gross’s conclusions drawn from animal stud*555ies on Bendectin are without foundation and are no evidence. The trial court did not abuse its discretion in admitting his testimony and his conclusions.
2. Jay Howard Glasser
Jay Howard Glasser holds a doctorate in experimental statistics from the University of North Carolina, a Master of Science in bios-tatistics, and a Bachelor’s degree in zoology. At time of trial, he was an associate professor at the University of Texas School of Public Health at the Texas Medical Center in Houston. He has taught epidemiology, statistics, biometry, and biostatistics. He has published over thirty articles in peer-reviewed journals in the areas of epidemiology, biostatisties, and biometry. At trial, Glasser reviewed the epidemiological studies involving Bendectin exposure. He explained the statistical terminology used by researchers, the kinds of problems inherent in epidemiological studies, and the inability of epidemiological studies to do more than track associations. Epidemiological studies cannot be used to prove or to disprove biologic causation. One of Glasser’s concerns about the majority of the studies and their authors’ conclusions had to do with the timing of Bendectin ingestion as described in the studies. The relative risk of limb reduction defects is significantly greater if exposure occurs in the first trimester rather than merely during the pregnancy as a whole.9 If exposure occurred during the first trimester, the relative risk was 2.18 rather than .97 if the exposure occurred anytime during pregnancy. Glasser’s examination of the study data included a reanalysis of the relative risk when exposure could be broken out of the author’s raw data, even if the author did not do that particular analysis in that fashion. Glasser concluded that to a reasonable degree of epidemiological certainty, Bendectin is a teratogen, is connected with congenital disorders, and that there is an association between Bendectin and limb reduction defects if the exposure occurs within the first trimester of pregnancy.
Dr. Glasser was qualified to testify about statistical techniques and the interpretation of epidemiological studies based on statistical analysis. His testimony constitutes some evidence that Bendectin is associated with limb reduction defects in humans.
3. Stewart Newman
Stewart Newman holds a doctorate in chemistry and is a full professor at New York Medical College where he teaches embryology to medical students. He has also taught anatomy. His field includes cell biology, developmental biology, embryology, and general biology. Newman has done research in the areas of genetics, molecular genetics, and limb development. He has published in the area of limb development and has written over fifty papers in the areas of developmental biology and biologic chemistry. He considers himself an experimental teratologist. His primary research interest is in vitro studies of cell development. Newman initially testified about human fetal development. At twenty-five days after conception, the fetus’s limbs are beginning to form. Between the fourth and seventh week after conception, the limb and skeleton develop on a very small scale. The initial skeleton is not bone, but cartilage. The process of turning cells into cartilage is called chondrogenesis. Chondrogenesis begins at about the fifth week and if disrupted, will effect limb development. Chondrogenesis may be inhibited or stopped altogether by exposure to a tera-togen. When that happens, the succeeding structures do not develop properly.
Newman reviewed Mrs. Havner’s pregnancy medical records. Depending on whether date of conception was based on her last menstrual period or whether it was counted backwards from the date of Kelly’s birth, Newman estimated that Havner took Ben-dectin between 30 and 42 days after conception. Based on fetal developmental timetables, if Bendectin was present between days 42 and 49, he would expect hand development to be disrupted.
*556Newman reviewed studies performed on doxylamine succinate10 and other antihistamines of the same type. Some of the animal studies implicated antihistamines of that type in abortions or resorptions of fetal mice. He also reviewed in vitro studies of mouse and chicken embryos. Even low doses of doxyla-mine succinate inhibited or impaired ehon-drogenesis in the animal embryos.
Newman testified that doxylamine succi-nate crosses the placenta11 in vertebrate animal studies. From that he concludes that it will do so in humans. In vitro studies are done to determine teratogenicity in vertebrate animals and the results are useful in predicting what will occur with humans. He noted that four different laboratories had found that doxylamine succinate is a terato-gen in experimental systems. In contrast, Newman notes that epidemiology is considered to be a “soft” science because the researcher does not control all the variables, compared to the “hard” science in which the researchers control all the variables in laboratory testing.
Newman concluded that Bendectin and doxylamine succinate are teratogens based on a reasonable degree of scientific certainty. He also concluded that Mrs. Havner ingested Bendectin at the time when Kelly’s limbs were developing and that her defect is a loss of a portion of her skeleton which is based on exposure to a teratogen during the late 30 day and early 40 days of her gestational development. Newman is qualified to testify in the areas of cell development and in vitro studies. He is also qualified to draw conclusions in those areas. His testimony is some evidence of the “window of vulnerability” on which the plaintiffs based their claims.
4. John Palmer
John Palmer is a medical doctor and holds a doctorate in pharmacology. He is a full professor of pharmacology at the University of Arizona medical school. As part of his teaching responsibilities, he has taught toxicology and teratology to medical students. He has published book chapters on pharmacology and over thirty papers in major journals.
His current laboratory research work involves antihistamines in animal studies. Doxylamine succinate, a component of Ben-dectin, is an H-l antihistamine. Studies of other H-l antihistamines show teratogenic effects in animal studies. Doxylamine succi-nate crosses the placenta in animal studies and he thinks it likely that it does so in humans. Animal studies are important because similar studies cannot be performed on humans. Animal studies can be used with other information to determine teratogenicity.
Palmer is familiar with in vitro studies and considers them useful to pharmacologists because of the ability to isolate the effects of the agent on tissue without the complications inherent in an entire body system. He has performed in vitro studies during his years as a medical and pharmaceutical researcher. He has reviewed in vitro studies involving Bendectin and doxylamine succinate.
Palmer is also familiar with epidemiological studies. Pharmacology uses epidemiological techniques although he has not personally conducted an epidemiological study. Epidemiological studies are useful although he would not rely solely on epidemiological studies to determine teratology. Epidemiological studies do not prove causation, although they may show association.
In order to determine whether a drug is teratogenic, Palmer considers all of the data to be important including in vitro experiments, animal studies, epidemiological studies, and human studies. After reviewing Havner’s pregnancy medical records, examining Kelly’s hand, and reviewing all the epidemiological, animal, and in vitro data, Palmer concluded, based on a reasonable degree of medical certainty, that Kelly’s limb *557reduction defect was caused by her mother’s ingestion of Bendectin during the critical period of Kelly’s limb development.
Dr. Palmer’s qualifications are sufficient for him to testify about the action of a pharmaceutical product on a developing fetus. He has performed studies involving both animals and in vitro techniques, he works with epidemiological data, he treats patients, and is familiar with medical records and their interpretation. His testimony is some evidence that Bendectin caused Kelly Havner’s limb reduction defect. His testimony adequately supports scientific causation.
D. Conclusion
Appellant cites us to a multitude of cases, including Schaefer v. Texas Employers’ Ins. Ass’n, 612 S.W.2d 199 (Tex.1980), for the proposition that even a qualified expert may not create “some evidence” merely by saying the magic words. See Burroughs Wellcome Co. v. Crye, 907 S.W.2d 497 (Tex.1995). We do not disagree with that proposition.
For an expert opinion to constitute evidence of causation it must rest in reasonable medical probability. Id. at 500. Reasonable probability is determined by the substance and context of the opinion, and does not turn on semantics or on the use of “magic words.” Id.
In Schaefer there was a specific laboratory test that could have conclusively demonstrated the presence of the bacteria which was claimed to cause Schaefer’s particular form of tuberculosis. That test was not performed. Schaefer’s physician believed from other evidence that Schaefer was infected with that particular and rare strain. The supreme court disagreed. In this case there are no such conclusive tests possible.
In Burroughs, the Cryes’ expert hypothesized that the product caused frostbite to Mrs. Crye’s foot and relied for his opinion on several assumptions. One of those assumptions was that Crye did not apply the spray as directed by the packaging and another was that her foot did not become red after she applied the spray. Both assumptions were directly contradicted by the Cryes’ testimony. No other evidence was introduced on those two issues. The court in Burroughs held that the expert’s testimony,
constitutes no evidence that [the product] caused Crye to sustain a frostbite injury. When an expert’s opinion is based on assumed facts that vary materially from the actual, undisputed facts, the opinion is without probative value and cannot support a verdict or judgment.
Id. at 499. That is not the ease here.
The Havners produced evidence that Ben-dectin is a human teratogen and that it caused Kelly’s birth defect.12 We overrule appellant’s challenge to the legal sufficiency of the evidence of causation.
III. Factual Sufficiency of the Evidence
In reviewing the factual sufficiency of the evidence, we consider and weigh all the evidence, but we may only exercise our jurisdiction over factual sufficiency to prevent a manifestly unjust result. We may not reweigh the evidence and set aside a jury verdict “merely because the judges feel that a different result is more reasonable.” Pool v. Ford Motor Co., 715 S.W.2d 629, 634 (Tex.1986); In re King’s Estate, 150 Tex. 662, 244 S.W.2d 660, 661 (1951).
In our discussion of the legal sufficiency of the evidence we outlined only the evidence supporting the jury’s finding of causation. The competing evidence was also substantial.
A. Havners’ Evidence
All of the Havners’ witnesses admitted that none of the authors of the animal or epidemiological studies concluded that Ben-dectin caused limb reduction defects. All of the witnesses at trial agreed that they were not primary researchers into Bendectin and its effect on developing fetuses.
Dr. Newman admitted that birth control pills can cause limb reduction defects and *558that Mrs. Havner took birth control pills up until two months prior to her pregnancy with Kelly. Nevertheless, Newman did not believe that Kelly’s defect was attributable to her mother’s taking oral contraceptives. Newman agreed that doxylamine succinate was on a listing of substances generally considered to be not teratogenic in rats and rabbits. Several of the Havners’ experts agreed that there are many more substances that are teratogenic in animals than in humans.
Dr. Palmer admitted on cross-examination that he was not board certified in internal medicine and that he has faded his boards twice. He agreed that one cannot conclude that Bendeetin is a teratogen from the epidemiological data. He admitted that he did not know how mueh Bendeetin Mrs. Havner took nor for how long.
The Havners testified that no one in their families has a birth defect. They also testified that they had gone for genetic screening after Kelly was born and were told after a lengthy interview that it was unlikely they would have another child with this defect. It was not clear from the testimony whether any laboratory testing had been performed as part of that screening.
B. Merrell Dow’s Evidence
Merrell called five expert witnesses: Stephen Henry Lamm, M.D., James Lee Goddard, M.D., Frank Greenberg, M.D., Jan 'Marshall Friedman, M.D./Ph.D., and Raymond Harbison, Ph.D. All testified that Bendeetin did not cause Kelly’s birth defect.
1. Stephen Lamm
Lamm is a medical epidemiologist in private practice. He also holds a master’s degree in biophysics and is board certified in pediatric medicine. Lamm worked for the Public Health Service at the Center for Disease Control as part of the Epidemic Intelligence Service. His work there involved the use of epidemiologic data. He served as senior epidemiologist to the Child Health Institute. Lamm has intermittently taught epidemiology to medical students. Lamm’s current private practice is not treating patients but as a consultant in the areas of epidemiology and occupational health.
Lamm first reviewed the epidemiological literature on Bendeetin in 1984. His primary focus is on epidemiological data, but he admitted that for others, animal and in vitro studies may be important.
Lamm agreed that fetal limb formation occurs between 26 to 56 days after conception and that the timing of exposure to Ben-deetin is important in looking at the studies. Reviewing an unpublished draft of a study by Jick which reconsidered calculations of Ben-deetin exposed women, some of whom were also exposed to spermicides, Jick found a statistically significant relationship between limb reduction defects and Bendeetin exposure. Lamm testified that he could make calculations from the raw data reported in the studies to answer specific questions not addressed by the researchers and had done so with some but not all.
Lamm conducted a pooling analysis using data from all of the studies and concluded that Bendeetin does not cause birth defects in humans nor does it cause limb reduction defects. He stated that his opinion was based on reasonable medical certainty.
In conducting his pooling analysis, Lamm used the Mantel-Haeszel odds ration calculations which he admits had not been used before in the context of birth defects. He had never seen the technique published in connection with birth defect studies although it has been published for cancer studies. Lamm combined the cohort studies and the case-control studies in his pooling analysis. He did not know whether other scientists combined the two types of studies in pooling. Lamm testified that he had not submitted his analysis for publication although he claims to have presented it to the Society of Epidemio-logic Research. The result of his analysis was an odds ratio of 1.09. He did not consider that odds ratio to be significant because the confidence interval puts the range below 1 at .08, which indicates no relationship. Lamm’s analysis results are inconclusive of a relationship between Bendeetin and limb reduction defects. Lamm admits that over the past 25 years, the only drug that has been conclusively shown to be teratogenic through *559epidemiological studies is Accutane, an acne medication.
2. James Lee Goddard
Goddard is a former Commissioner for the FDA. He no longer practices medicine; he gave up his medical license in 1973. He holds a Master’s degree in Public Health which he obtained in the 1950’s. His sole academic training in epidemiology occurred during his master’s program. Goddard has testified forty times during the thirteen years prior to trial, usually for pharmaceutical companies. He reviewed the epidemiological data, animal, and other published materials on Bendectin and concludes that Ben-dectin does not cause birth defects in humans.
3. Frank Greenberg
Greenberg is a practicing physician and is Director of the Birth Defects Genetics Clinic at Texas Children’s Hospital. He also teaches at the Baylor College of Medicine in the fields of molecular genetics, pediatrics, and obstetrics and gynecology. He saw the Hav-ners and reviewed the medical records. He concludes that Kelly’s defect is genetically caused rather than drug induced. Green-berg found Kelly’s father’s hands to be smaller than average. Mrs. Havner’s hands are normal. According to Greenberg, a defect in her father’s genetic makeup caused Kelly’s birth defect, not exposure to any tera-togen.
4. Jan Marshall Friedman
Friedman is board certified in medical genetics and teaches in that area. He is a clinical teratologist and treats children with birth defects. He developed the Teratogen Information System. His opinion is that Bendectin does not cause birth defects in humans. He agrees that it is extremely difficult to determine teratogenicity based on epidemiological studies and that even thalidomide would not be established as a teratogen based on epidemiological studies.
5.Raymond Harbison
Harbison holds a doctorate in pharmacology and teratology. He specializes in animal teratology and animal testing. He thinks that animal studies are important to determine teratogenicity. He does not think the reliability of in vitro studies has been established to prove human teratogenicity. He admits that Bendectin crosses the placenta, but he does not think that Bendectin is a human teratogen at therapeutic doses.
C. Conclusion
In evaluating the evidence presented, the jury is sole judge of the credibility of the witnesses and the weight to be given to their testimony. Tex.R.Civ.P. 226a(III); Jaffe Aircraft Corp. v. Carr, 867 S.W.2d 27, 28 (Tex.1993); Lopez v. Salazar, 878 S.W.2d 662, 663 (Tex.App.—Corpus Christi 1994, no writ). Such a determination of weight and credibility applies to expert testimony as well. Texas Employers Ins. Ass’n v. Campos, 666 S.W.2d 286, 289 (Tex.App.—Houston [14th Dist.] 1984, no writ). Each of the witnesses was extensively cross-examined; flaws in their credentials, methodological disagreements, and errors of omission in their testimony were explored by opposing counsel. Moreover, each of the experts’ work was criticized by an expert for the opposing party. Our constitutional mandate13 authorizes us to reweigh the evidence presented to the jury in order to determine if its verdict is so contrary to the greater weight of the evidence so as to be manifestly unjust. This we have done and conclude that the verdict as to causation is supported by sufficient evidence. Appellant’s challenge to the factual sufficiency of the evidence is overruled. Point one is overruled in its entirety.
IV. Same Ten Jurors Did Not Agree
By point of error two, appellant complains that the ten jurors who agreed on liability and damages did not agree on the punitive damages verdict. The trial was bifurcated. Only after primary liability was determined were punitive damages submitted to the same jury. Merrell contends that *560the same ten who found liability must have agreed on punitive damages in order for the punitive damages portion of the verdict to be valid. We disagree. We addressed a similar circumstance in Greater Houston Transp. Co. v. Zrubeck, 850 S.W.2d 579 (Tex.App.—Corpus Christi 1993, writ denied). This situation has arisen infrequently and we find no Supreme Court authority. We will follow our previous analysis. Point two is overruled.
V. Punitive Damages
By Point of error three, appellants attack the award of punitive damages. Their attack is three-pronged: 1) that plaintiffs’ had to establish that Merrell acted with conscious indifference; 2) that there is no evidence to support such a finding; and 3) that an award of punitive damages in this ease would violate both the state and federal constitutional guarantees of due process.
A. Constitutional Due Process
In Transportation Ins. Co. v. Moriel, 879 S.W.2d 10, 27-33 (Tex.1994), the supreme court imposed procedural changes to address the very due process concerns raised by Merrell Dow here. Among the changes are bifurcation of the punitive damages portion of the trial, an analysis of the Alamo Nat’l Bank v. Kraus14 factors, articulation of the evidence by the reviewing court that supports the finding of punitive damages, and a refinement of the legal sufficiency standard of review. With these changes, our supreme court considers that the Texas system will pass federal constitutional muster and will satisfy our own constitution. See Moriel, 879 S.W.2d at 27-33. Although Moriel was decided after the trial of this case it applies to it. Ellis County State Bank v. Keever, 888 S.W.2d 790, 799 (Tex.1994).
In this trial, the procedural protection of bifurcation of punitive damage issues was observed. The jury first considered liability and actual damages, then returned to consider the punitive damage issues.
Initially we consider the charge to the jury. The jury was asked:
Question No. 1
Was the negligence of Merrell Dow Pharmaceutical Inc. “gross negligence”?
“Gross negligence” means more than momentary thoughtlessness, inadvertence, or error of judgment. It means such an entire want of care as to establish that the act or omission in question was the result of actual conscious indifference to the rights, welfare, or safety of the person affected by it.
Although this question does not contain the detailed instructions which the Moriel court requires for trials which occur after that opinion, this case was tried in October 1991, long before Moriel was handed down. We do not understand the Moriel decision to require automatic reversal based on a charge that was acceptable at time of trial but is no longer considered adequate.
However, we will review the sufficiency of the evidence to support the finding of gross negligence against the heightened review mandated by Moriel. We will consider that evidence in light of the Kraus factors. The expanded post-judgment review established by Moriel satisfies appellant’s due process complaints. Accordingly, we overrule appellant’s constitutional challenge to the award of punitive damages.
B. Sufficiency of the Evidence to Support Punitive Damages
We consider appellant’s first and second challenges together. The Havners were required to prove actual conscious indifference on Merrell’s part. A gross negligence finding may be upheld on-appeal only if there is legally sufficient evidence that the defendant’s conduct created an extreme risk of harm and that the defendant was aware of the risk. Moriel, 879 S.W.2d at 21. Extreme risk is a function of both the magnitude and the probability of the anticipated *561injury to the plaintiff. Id. at 22. The extreme risk portion of the inquiry is not satisfied by a remote possibility of injury or by a high probability of minor harm. Id. We consider whether the act or omission involves extreme risk by examining all the events and circumstances from the viewpoint of the defendant at the time the events occurred. Id. at 23.
Kelly Havner has a very serious injury. The lack of a functional hand limits one’s ability to perform simple tasks and participate in many of the pleasures of life. Her injury is one of great magnitude.
Whether Merrell knew of the high probability that a child would be born with such a defect if the mother took Bendectin is a more difficult question. To justify punitive damages, Merrell must both have known of the risk and consciously disregarded it. We review the evidence on both questions together. In doing so, we consider what Merrell Dow knew at the time of manufacture. See id. at 23.
The Bendectin taken by Mrs. Havner was manufactured sometime in 1980 or 1981; that formulation had been on the market since 1975. The defendants produced evidence that the rate of limb reduction defects has not varied substantially from periods when Bendectin was heavily prescribed all over the world compared to periods before Bendectin was ever marketed and periods after Bendec-tin was taken off the market.
In support of their claim that Merrell was consciously indifferent to the rights of others, the Havners charge that Merrell did not submit a complete report to the FDA that summarized the Staples study in the 1960’s. In particular, Merrell did not report the bony abnormalities in rabbit fetuses that had been found. The Havners conclude from the omissions that Merrell intended to mislead the FDA about Bendectin’s effect on developing fetuses. But, Dr. Gross, who criticized the omissions, admitted that there was an ongoing debate within the scientific community about what should be classified as a variation versus what should be classified as an anomaly. He also admits that this particular data falls squarely within the debate. The FDA reexamined Merrell’s raw data and found it to support the report as submitted. Dr. Gross disagrees with the audit conclusions. Nevertheless from Merrell’s perspective, the FDA audit cleared it of the charges that it misrepresented the data.
The Havners also contend that the FDA’s heightened scrutiny of Merrell Dow data during that period was the motivation for the discrepancies between the raw data of the Staples study and the report submitted. This heightened scrutiny resulted from data Merrell Dow submitted to the FDA on MER 29. That data was determined to be false. Merrell Dow was indicted by the Justice Department and pled nolo contendere to the charges.15 As a result of Merrell’s false data regarding MER 29, the FDA issued a memo to its employees requiring any data submitted by Merrell be subject to special scrutiny and not be processed under normal procedures but be directed to two specifically designated persons for review. The Havners claim that Merrell’s indictment and plea prove that there was a pattern of falsifying research data, which was continued in the Staples’ report.
Additionally, the Havners produced Dr. Elizabeth Hillman, a Canadian physician, who testified that when she called Merrell Dow in the late 1960’s to ask about reported birth defects in children born to mothers who took Bendectin, she was told there were none. Hillman admitted on cross-examination that she did not specifically recall making the telephone call, that she had no independent recollection of what the man at Mer-rell told her, although she came away with an impression that no other cases of birth defects had been reported. Dr. Hillman was the Poison Control Director at Montreal Children’s Hospital and also an Associate Professor of Pediatrics. She was not the attending physician for the mothers or children but had received reports from the patients’ physician. She recalls that she furnished the physician’s name to Merrell in her contact. Later, Dr. Hillman became aware of reports linking Bendectin and birth defects. In response to a hypothetical question *562assuming that Merrell had reports linking Bendectin to birth defects, Hillman characterizes Merrell’s providing her with misleading and incorrect data as willful and wanton misconduct. The Havners point to the Drug Experience Reports (DERs) which Merrell began to receive in the late 1950s for proof that Merrell received reports before Dr. Hill-man inquired.
Merrell put on evidence that it submitted a collection and analysis of DERs to the FDA, although it was not required to do so. The DERs cover a period from 1957 to 1972. Of the 96 reports of birth defects, 46 involve a variety of limb defects. Merrell examined the reports and concluded in a memo dated September 14, 1972, that some could not have resulted from Bendectin because the mother took the Bendectin after limb formation had occurred, that there was no pattern to the defects, that some were associated with other birth defects, and finally, that the defects resemble a study performed in Birmingham, England, between 1950 and 1959, before Bendectin was marketed. Additionally, over 7 million prescriptions were written for Bendectin through June 1969 and probably more than 9 million by July 1971. Based on the birth defect figures from the Birmingham study, Merrell’s scientists estimate that they should have received over 4000 reports of defects claimed to be associated with Ben-dectin use and they had not.
Dr. Goddard testified during the punitive damages phase that Merrell Dow submitted additional human and animal studies showing the two ingredient Bendectin to be safe and effective, that the company revised its labeling as new information became available, and that the labeling of Bendectin was in conformance with agency requirements. Goddard also recalls that Merrell funded a primate study to determine whether Bendectin was teratogenic and presented the findings to the Maternal Health Advisory Committee. That study was not required by the FDA nor any other governmental agency. He agreed on cross-examination that drug companies have a duty to provide correct, complete, and accurate information to doctors about their products and that it would be irresponsible for a drug company to mislead doctors.
Mark Hoekenga, M.D., a physician who worked for Merrell Dow in the 60’s, testified that during his time at Merrell he was responsible for Merrell’s regulatory contacts for Bendectin and other drugs. During his time there, he testified that he regularly searched the medical and scientific literature for any references to Bendectin that linked it with congenital abnormalities. In August 1964, he compiled a booklet of published studies and DERs on Bendectin and submitted it to the FDA, to the Canadian regulatory agency, the AMA, and generally made the booklet available to interested scientists. Hoekenga had a personal interest in birth defects; he was bom with several, his father died from a congenital heart condition, and one of his children died in infancy from congenital defects; his other child’s defects were surgically corrected. Hoekenga testified that Merrell convened a group of experts from an outside organization called Bio/Basics International to review the Bendectin data in 1975. Bio/Basics grew out of a group of professors at Harvard who reviewed drug situations. Merrell’s purpose was to have an outside group review all the Bendectin information, human studies and animal studies, to determine whether Bendectin was safe and effective to relieve nausea and vomiting during pregnancy. The group concluded that in 1973 there were 2 million therapy starts and the occurrence of congenital malformations does not indicate a change in the incidents of or pattern of congenital malformations between the Bendectin exposed populations and the normal population. The group also concluded that in rat and rabbit testing, Bendec-tin and its component doxylamine were not found to be teratogenic. Overall the group found that the clinical data and animal data fail to reveal a teratogenic tendency in Ben-dectin.
Hoekenga also testified that he submitted the DERs to two outside researchers for review. Neither found the DERs indicative of teratogenic effect by Bendectin for the same reasons that the Merrell researchers did not. Hoekenga remained at Merrell until 1982 when he retired. His conclusion was that Bendectin was not causally linked to birth defects.
*563During cross-examination, Hoekenga admitted that the outside groups and reviewers had been paid for their time in reviewing the materials provided by Merrell.
The evidence that Merrell Dow knew and consciously disregarded the risk of birth defects caused by Bendectin exposure before the 1980 and 1981 consists of the 96 DERs, the testimony of Dr. Hillman that Merrell told her that it did not know of other reports, the alleged falsification of raw data from the Staples study, and Merrell’s plea of no contest to an indictment alleging falsification of data regarding the cholesterol drug MER 29 in the early ’60s. This evidence is too weak to support the conclusion that Merrell Dow was consciously indifferent to an extreme risk that Bendectin could cause limb reduction defects if taken during the critical period of pregnancy. Having found the evidence to be too weak to support the punitive damages, we do not reach the Kraus analysis. We sustain appellant’s challenge to the legal sufficiency of the evidence to support the jury’s finding of gross negligence.
VI. Evidentiary Challenges
By point four, appellant challenges the admission of materials surrounding Merrell’s indictment for falsifying data regarding MER 29, the admission of information regarding a television program on Bendectin in which the drug was linked to birth defects, and admission of the Smithells’ letters to prove that Merrell “purchased” favorable scientific research. We review challenges to the trial court’s admission of evidence under an abuse of discretion standard. Ginsberg v. Fifth Court of Appeals, 686 S.W.2d 105, 107 (Tex.1985); Jones v. Jones, 890 S.W.2d 471, 474 (Tex.App.—Corpus Christi 1994, writ requested). We reverse only when the trial court acts without regard to guiding rules and principles and when “the error complained of amounted to such a denial of rights of the appellant that was reasonably calculated to cause and probably did cause rendition of an improper judgment in the case_” Tex.R.App.P. 81(b)(1); Downer v. Aquamarine Operators, Inc., 701 S.W.2d 238, 242 (Tex.1985).
A. MER 29
The first part of appellant’s complaint involves three separate pieces of evidence related to MER 29. All were introduced during the punitive damages phase. Because of our disposition of point of error three, we need not address this issue.
B. Television Program
Appellant next complains of the admission of evidence of a television program in which Bendectin was linked to birth defects. Counsel asked Mr. Havner when and how he came to suspect that Bendectin caused his daughter’s injury. Appellant objected to hearsay. Havner answered that he first suspected that Bendectin caused the injury in August 1986 while he was watching television. He also testified that he saw on the screen a hand that looked like his daughter’s. Appellant objected again before Havner answered the question. Havner testified that after he saw the hand on television that he asked his wife whether she took Bendectin.
Appellant objected again, requested that the jury be instructed, and moved for a mistrial. The court overruled the objection, and rejected counsel’s request for an instruction and mistrial. But the court also instructed plaintiffs counsel to move away from that area of questioning.
Merrell pled the defense of statute of limitation in its Original Answer and in its answer to the Havners’ First Amended Original Petition. Merrell denies that the statute of limitations was at issue by time of trial. The date on which Havner first suspected that Bendectin caused Kelly’s defect may have been at issue and that information was properly admitted. What Havner heard on the television program is not hearsay if it was not offered to prove that Bendectin causes birth defects. That Havner saw a hand that looked like Kelly’s is not hearsay, and that seeing that hand prompted him to investigate whether his wife took Bendectin while she was pregnant with Kelly is also not hearsay. The trial court did not abuse its discretion in overruling the objection.
We consider whether the admission of the evidence was harmful if its admission was in *564error. The testimony at issue consumes less than a page out of a trial consuming over 2500 pages of statement of facts. It likely was no more than a few minutes out of a trial lasting a month. Although it is not to be doubted that many of us place great faith in what we ourselves hear and see on television; it is less clear what value any of us place on second hand reports of what another saw on television. It is not at all clear, given the extensive scientific testimony, that Havner’s isolated comments provided the nudge that caused rendition of a verdict different than what otherwise would occur. We are not persuaded that admission of the testimony would have irreparably tainted the entire proceeding.
C. Smithells Letters
During the liability and actual damages portion of the case and during its rebuttal, the Havners introduced the Smithells letters. Dr. Smithells solicited funds from Merrell to underwrite the costs of a study on Bendectin. Smithells was a British researcher. Merrell objected to the admission of the letters on several bases: hearsay, lack of authentication, that the letters had little probative value, and that the probative value was substantially outweighed by their unfairly prejudicial effect. The court permitted portions of the letters to be read to the jury and then sustained the objection. Merrell did not request an instruction to disregard or move to strike the testimony from the record. Although we disapprove of the court’s permitting the testimony before ruling on the objection, Merrell received all the relief it requested. Any further complaint was waived. The trial court did not err. Point four is overruled.
VII. Conclusion
We find that the evidence is legally and factually sufficient to support the verdict on the issue of actual damages although it is legally insufficient to support the punitive damages finding. The trial court’s evidentia-ry rulings do not require remand.
The JUDGMENT is AFFIRMED as to actual damages and REVERSED and RENDERED as to punitive damages.
. "This is the traditional use courts have made of expert testimony." Parker v. Employers Mutual Liab. Ins. Co. of Wis., 440 S.W.2d 43, 46 (Tex.1969). "The expert is in effect not telling of facts at all, but of uniform physical rules, natural laws, or general principles, which the jury must apply to the facts.” Id. at n. 5 (quoting Hand, "Historical and Practical Considerations Regarding Expert Testimony”, 15 Harv.L.R. 40, 50 (1901)). In such cases, no dispute exists as to the underlying scientific principle. For example, it may be an undisputed "natural law” that under particular circumstances the ingestion of drug "Z" produces drowsiness and impaired responses. The scientific expert need only explain the underlying principle to the jury and explain under what particular circumstances drowsiness and impaired responses are caused by the ingestion of drug "Z.” The jury may then decide if the facts presented in a particular case render it more likely than not that the ingestion of drug "Z" caused the drowsiness and impaired responses experienced by the plaintiff.
. The PDR is the text generally referred to by physicians for information on available drugs. It provides each drug’s description, action, indications for use, contraindications, warnings, precautions, adverse reactions, dosage and administration information, and treatment for overdose.
. Without rendering an opinion as to the qualifications of the experts or the admissibility of their testimony under an evidentiary analysis utilizing Texas Rules of Civil Evidence 102, 104, 401, 402, 403, 702, 703, 704 and 705, we will initially assume proper qualifications and admissibility and merely review the expert testimony admitted in support of the finding of causation by the jury.
. It was undisputed that every prescription drug, by its very nature, is capable of producing deleterious effects and even death when administered in quantities greater than the determined human therapeutic dose. There was no assertion here, by either party, that Mrs. Havner exceeded the recommended dosage.
. A teratogen is an agent or factor that causes the production of physical defects in the developing embryo.
. For example, absence of proof that a stop light was red, is not some evidence that it was green or yellow, unless some presumption or legal principle, such as the burden of proof, operates in favor of the green-yellow theory.
. While epidemiologic studies are among the most relevant types of information available to establish a human causal link between drugs and adverse reactions, we note that these studies will never produce some evidence of “no effect” or *544"no association”. All "no association” cases would fall within the protective effect zone (drug not only has no adverse effect, but actually decreases the incidence of naturally occurring adverse effects) or within the inconclusive zone (no protective effect and no adverse effect). These studies, by their very nature, are incapable of proving "no effect.” It merely becomes a function of time, repetition, and sheer numbers of individuals exposed before a continued showing of “inconclusive" will equate, in the scientific mind, to "no association” or “no effect.”
. For example, there is a statistically significant, demonstrable "association” between summertime and death by drowning. Summertime is "related" to an increase in drownings. However, summertime does not cause the increase in drownings. Summertime is merely an innocent bystander to the true cause: accidents accompanying an increase in water-related activity.
. The term literally means "in glass”, but here it refers to the scientific process in which cells are removed from the living animal and placed in a controlled environment (a petri dish or test tube) with nutrient fluids to keep them alive. The cells are then exposed to various agents or factors, and the response is noted.
. This is a major oversimplification by this Court of the actual results reported. However, a more scientifically detailed explanation is neither helpful nor relevant.
. Chondrogenesis is the formation or development of cartilage.
. Dr. Newman admitted that the actual researchers concluded only that doxylamine was “potentially capable of inducing genetic damage and should be tested in other systems which detect genetic endpoints.” Additionally, Dr. Newman has continued to selectively ignore subsequent studies, pointed out to him in prior Ben-dectin litigation, which directly refute his theories. Because our review is confined to supporting evidence, we also ignore them.
.On cross-examination, Dr. Newman agreed that recent studies have revealed a link between prolonged use of birth control pills and subsequent limb reduction defects. He explained the causative factor as a deficiency of B vitamins in the mother's system. Use of birth control pills for a period longer than six months can deplete B vitamins. Pregnancy itself depletes these same vitamins. Dr. Newman described the result as a ' 'double-whammy.''
His testimony further revealed that Mrs. Hav-ner had taken birth control pills for eight years. However, it was his personal opinion that, because she had discontinued use of the pills two months prior to conceiving Kelly, her vitamin levels were probably normal.
. The actual DERs were found inadmissible as unreliable hearsay.
. Merrell Dow Pharmaceuticals, Inc. v. Havner, 13-92-540-CV (Tex.App.—Corpus Christi March 17, 1994).
. The Court included Justices Noah Kennedy and Gilberto Hinojosa, who left the court on December 31, 1994. On January 1, 1995, Justices Mel-chor Chavez and Nelda Rodriguez replaced them and participate in this decision.
. The claims against Dr. Bruce were dismissed before trial.
.Epidemiological studies in this context measure the statistical relationship, if any, between exposure to an agent and the incidence of a condition. For example, exposure to Bendeetin and the incidence of birth defects. Such studies compare the incidence of birth defects in populations exposed to bendeetin with the incidence of birth defects in populations who were not exposed to it.
. Animal studies use laboratory animals, often rats, rabbits, or other mammals, to explore the reaction of the experimental animal to the drug or other substance being investigated.
. In vitro studies are conducted in a laboratory also and involve tissue samples, generally animal tissue, to see how cells react to the substance being investigated.
. A teratogen is a chemical agent that causes birth defects. The agent may be ingested or the mother may have an environmental exposure.
. Bendectin had two ingredients: doxylamine succinate (an antihistamine) and pyridoxine hydrochloride (vitamin B-6) at the time it was prescribed for Mrs. Havner. Earlier versions of the drug had three ingredients. Merrell manufactured Bendectin in the United States from 1956 to 1983.
. Relative risk measures the connection between the agent studied and a given result. A relative risk of less than one indicates no connection or a protective effect. If the relative risk is greater than one, there is an association between the agent studied and the result.
. Doxylamine succinate, an antihistamine, is one of two active ingredients in Bendectin as it was formulated at the time it was prescribed to Mrs. Havner.
. The placenta is the structure inside the uterus through which the fetus receives nourishment and excretes waste into the mother’s circulatory system. It has been considered a protective barrier that shields the fetus from harm from agents in the mother’s system.
. The evidence here appears to differ somewhat from the evidence considered in Daubert after remand to the Ninth Circuit. DuPont, 38 Tex. Sup.Ct.J. at 857, - S.W.2d at - (citing Daubert, 43 F.3d at 1322).
. "Provided, that the decision of said courts shall be conclusive on all questions of fact brought before them on appeal or error.” Tex. Const. Art. V § 6.
. 616 S.W.2d 908, 910 (Tex.1981). The factors are: the nature of the wrong; the character of the conduct involved; the degree of culpability of the wrongdoer; the situation and sensibilities of the parties concerned; and the extent to which such conduct offends a public sense of justice and propriety. Id. at 910.
. In its fourth point of error, Merrell challenges the admissibility of this evidence.