concurring.
I write separately because I don’t feel comfortable second-guessing this judge and jury on this record, without clearer guidance from the legislature or the supreme court. While this record is thin, plaintiffs in this case presented sufficient evidence that an alternative design was available and feasible and that GM’s inadequate warning caused Sanchez’s death. Moreover, this jury was in no danger of being misled.
Safer AlteRnative Design
Viewing the supreme court’s most recent guidance, appellate consideration of expert testimony presents two questions. See Maritime Overseas Corp. v. Ellis, 971 S.W.2d 402 (Tex.1998). The threshold question is, of course, whether the expert testimony is scientifically rehable and therefore at least some evidence in support of a judgment. Id. at 409; see also Daubert v. Merrell Dow, Inc., 509 U.S. 579, 595-97, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993); Merrell Dow Pharm., Inc. v. Havner, 953 S.W.2d 706, 711 (Tex. 1997), cert. denied, - U.S. -, 118 S.Ct. 1799, 140 L.Ed.2d 939 (1998); E.I. duPont de Nemours v. Robinson, 923 S.W.2d 549, 558 (Tex.1995). Making this determination requires consideration of the factors listed by the dissent — whether the expert’s theory has been tested, to what extent the technique relies upon the subjective interpretation of the expert, whether the theory has been subjected to peer review and publication, the theory’s potential rate of error, whether the underlying theory has been generally accepted by the relevant scientific community, and any nonjudicial uses of the theory. Havner, 953 S.W.2d at 714. These considerations go to the trial court’s discretion in admitting or excluding the scientific testimony. Ellis, 971 S.W.2d at 409. In Ellis, the supreme court held that error in the admission or exclusion of scientific evidence based on its reliability must be preserved on those grounds before or during trial in order for the appellate court to consider it.1 Id. at 409-10. As the court noted in Ellis, this requirement is grounded in concerns for fairness to appellees:
[Permitting [a party] to challenge on appeal the reliability of [the opposing party’s] scientific evidence under Daubert, in the guise of an insufficiency-of-the-evidence argument, would give [appellant] an unfair advantage. [Appellant] would be ‘free to gamble on an unfavorable judgment before the trial court, knowing that [it could] seek reversal on appeal [despite its] failure to [object at trial].’
Id. (quoting Marbled Murrelet v. Babbitt, 83 F.3d 1060, 1066-67 (9th Cir.1996), cert. denied, - U.S. -, 117 S.Ct. 942, 136 L.Ed.2d 831 (1997)).
GM has attempted to conflate what seems to me to be two distinct questions: whether the evidence of an alternative design, as scientific expert testimony, was reliable, and thus sufficient, under Robinson standards, and whether the evidence, once admitted without objections as to reliability, was otherwise legally sufficient to support the jury’s verdict. I believe the first issue was waived and the fact finder resolved the second.
I cannot fault the trial judge for not sua sponte limiting expert testimony based on the vague guidance now available. GM did not object to the scientific reliability of Tam-ny’s testimony until its motion for a directed *409verdict, made after the close of the plaintiffs’ evidence. That motion claimed that Tamny’s testimony regarding the reasons that the truck moved backward was “based on pure speculation and subjective interpretation.” The objection did not mention the alternative design proposal. If trial courts must be “gatekeepers” charged with the responsibility of preventing unreliable scientific testimony from reaching the jury, then they must be given adequate notice of reliability challenges so that they can make correct determinations based on fully developed evidence. See id. at 411-12 (agreeing with concurrence that “it is impossible for a [trial] court to exercise its gatekeeper function after the evidence has been admitted and the jury discharged”). However, I believe Tamny’s testimony would survive a reliability challenge. While the testimony is pithy under Robinson standards, I find it reliable. The design Tamny proposed had been tested. It was his uncon-troverted testimony that it was being used, in whole or in part, by other auto makers. Specifically, he stated, “Nothing that I am proposing is new.... [N]othing has to be proven as to whether it will work, because people are doing it.” There was testimony of commercial uses of the design, in spite of GM’s attempt to characterize it as merely developed for litigation. Tamny offered testimony on the design’s potential effectiveness and on its limitations — it could not prevent a car or truck from slipping into neutral. That testimony was not based solely on Tamny’s subjective interpretations; it was based on his knowledge that the same or similar designs had been implemented by other auto makers and had been successful in varying degrees. It is true that the proposed design had not been submitted to a publication subject to peer review, but it, or versions of it, is being used in actual auto manufacturing. The Robinson factors should not be used as a win-or-lose litmus test; if judges are to have any value in these eases, those factors must make sense under the facts of each case.
Our next task is to determine whether there was legally sufficient evidence, offered in support of each of the elements of the plaintiffs’ case, supporting the jury’s finding that an alternative design was available and feasible.
To prove that a safer alternative design was available to GM, the plaintiffs were required to show that, in reasonable probability, their design would have prevented or significantly reduced the risk of Sanchez’s injury without significantly impairing the product’s utility and that it was economically and technologically feasible at the time the product left control of the manufacturer, by the application of existing or reasonably achievable scientific knowledge. Tex.Civ.Prac. & Rem.Code Ann. § 82.005(b) (Vernon 1997). The statute does not require the plaintiffs to show that the testifying expert has designed or tested the product in question.2 I can find no Texas case imposing such a requirement on plaintiffs. To the contrary, the supreme court has explicitly stated, “Even if a safer alternative was not being used, evidence that it was available, known about, or capable of being developed is relevant in determining its feasibility.” Boatland of Houston, Inc. v. Bailey, 609 S.W.2d 743, 746 (Tex.1980). Requiring plaintiffs to demonstrate that their experts have personally, actually designed and tested alternative products that are feasible and safe for specific defendants or be subject to losing a no-evidence challenge is to go far beyond the statutory requirements. It is unrealistic and unfair to require plaintiffs to actually place in evidence a transmission that cannot slip into a powered gear.
Simon Tamny’s testimony was limited. There was not an extensive direct or cross. It is not a model of scientific explication. It was, however, nomenclature free and comprehensible to these jurors from a South Texas ranching community. It was not to my mind in the category of “scientific” evidence that Justice Blackmun, the author of Daubert, feared would mislead a jury. I don’t believe a case can be made that this jury lacked the capacity to understand Tam-ny when he explained the hazards of one’s truck slipping into powered reverse during routine ranch work. He testified that his proposed modification was “based on things that have been done before in production.” *410He stated that when he begin thinking about the design, he looked at existing designs to make sure that his proposal was practical. Further, while he admitted he had not designed an automobile transmission, he testified that he had designed transmissions for other kinds of equipment. Tamny testified that his proposal was safer than the transmission in Sanchez’s truck because, while the proposed design might still allow a transmission to slip into neutral, it would prevent the transmission from slipping into reverse, a powered gear. There was testimony at trial that if the truck had slipped into neutral, an unpowered gear, rather than reverse, Sanchez could have pushed the vehicle away. Finally, Tamny offered testimony that the proposal was economically feasible: his suggested modification would cost GM half a cent per truck.
In sum, Tamny offered testimony on each of the requisite statutory elements. The jury was entitled to believe that testimony and to find that a safer, feasible alternative to the design in Sanchez’s GM truck was available. In time, wide disparities in the admission of expert evidence will force a more satisfactory concept. In the absence of clearer guidance from either the supreme court or the legislature, or a record demonstrating that jurors were misled, we should not second-guess the jury’s determination.
FAILURE TO WARN
GM presents a more compelling argument when it complains that there is no evidence to support the jury’s findings that its warnings were inadequate and that this inadequacy was a producing cause of Sanchez’s death. GM produced a clear warning, but in light of Tamny’s testimony regarding the specific defect that caused Sanchez’s injury, I believe it was inadequate.
Tamny testified that there was no way that auto makers could implement a design that would prevent, in every instance, a vehicle from slipping into neutral. Such slippage is always possible. It is this danger of rolling that GM’s owner’s manual warns of:
It can be dangerous to get out of your vehicle if either: your shift lever is not fully in “P” (Park), or your transfer case is in Neutral (on vehicles with 4-wheel drive), or both. Your vehicle can roll. If you have left the engine running, the vehicle can move suddenly.... (If you have to Park on a hill, also turn your front wheels so that the vehicle would roll AWAY from traffic) (emphasis added).
Frankly, I don’t think the jurors needed an expert opinion to find this warning inadequate, in that it never tells the owner that the vehicle will be off and running under its own power if not completely shut off and firmly in park. Tamny’s testimony about the design defect in the GM transmission was that, in addition to the possibility that the vehicles would roll if they slipped into neutral, there was an unreasonably unsafe chance that they would slip into gear. This was the .plaintiffs’ claim — that Sanchez’s truck slipped into gear and that, had it slipped into neutral and merely rolled toward him, he could have pushed the truck off himself. Sanchez had no warning that the truck might slip into a powered gear. Moreover, the warning given by GM is inconsistent. While advising drivers to turn their engines off, it also suggests that leaving the engine running, briefly, is an acceptable practice. Thus, the manual entitled Sanchez to believe that it was safe to leave his truck running for the short amount of time it took to close the gate.
I feel comfortable in speculating that every member of the jury hearing this case has stopped his or her car to open a fence and stopped it again to close that same fence, without placing the car in park, turning the engine off, and, if on an incline, turning the wheels in the proper direction. I suggest this is a practice followed by almost every judge who has reviewed this verdict. The plaintiffs introduced as evidence a study on human behavior conducted by two scientists at Wright Patterson Air Force Base that found that leaving a car running without setting the parking brake is a common practice. According to Tamny’s testimony, GM engineers have testified in other trials that they themselves have done it. If the practice is so common, and the potential danger so horrific, a complete warning is warranted. And yet, according to GM’s manual, the most *411serious danger that could result from such a practice is that the truck would roll or suddenly move. Both these events are brief and considerably less serious than powered motion until an obstruction is met.
In addition, the jury heard testimony about the placement and labeling of the GM warning. GM’s boxed warning regarding using the parking brake is labeled, “Caution.” Tamny testified that published standards for warnings state that when there is a high likelihood of serious injury or death warnings should use the word, “Danger.” Warnings that a product might cause serious injury should use the word, “Warning,” and warnings that the product might cause very minor injury or damage to property should use the word, “Caution.” To be sure, this evidence was controverted by GM, but the jury was entitled to accept the plaintiffs’ version.
Once a plaintiff proves that inadequate warnings rendered a product unreasonably dangerous, there is a presumption that the user would have read and heeded such instructions, Magro v. Ragsdale, Bros., Inc., 721 S.W.2d 832, 834 (Tex.1986). This evidence may be rebutted by showing that the user was “blind, illiterate, intoxicated at the time of the product’s use, irresponsible, lax in judgment, or by some other circumstance tending to show that the improper use would have occurred regardless of the proposed warnings.” Id.
There is some evidence that the failure to provide adequate warnings caused Sanchez’s injury. Sanchez’s father testified that his son would have heeded adequate warnings and that his son had probably read the entire owner’s manual. While this evidence is not overwhelming, it, coupled with the presumption that Sanchez would have heeded adequate warnings, is at least “some” evidence in support of the verdict.
Conclusion
Evidentiary rules are designed to aid the jury, not to undo its work. The expert testimony offered by the plaintiffs regarding an alternative design was not challenged as scientifically unreliable at trial, although, in my view, it would survive such a challenge under the standards set out by the supreme court. Those standards are, to be sure, neither always clear nor fully developed. Nonetheless, I do not believe they mandate reversal of the jury’s verdict in this case.
Once the evidence was admitted without objection to its reliability, the jury was entitled to rely on it in order to reach its very difficult decision, as long as that evidence was legally sufficient under traditional standards. Because I believe the plaintiffs offered at least some evidence on every element of their case against GM, I concur in the majority’s decision to uphold the verdict.
. Ellis went up on appeal on a factual sufficiency point of error, the appellants argued to the supreme court that the court of appeals used an improper standard of review and argued that the trial court should have conducted a Robinson review of the scientific evidence. The plurality opinion is not entirely clear about whether a general no-evidence complaint that does not specifically raise Robinson issues, brought for the first time on appeal, would be sufficient to preserve Robinson error. See Ellis, 971 S.W.2d at 412 (noting that the court has not required preservation of no-evidence claims where, "on the face of the record” the evidence lacked probative value); id. at 414 (Gonzalez, J., concurring) (advocating specific Robinson objection, preferably during trial, in order to allow full development of evidence for appellate record). It seems clear that, in light of the court's emphasis on whether the insufficiency is apparent from the record and the need for a full evidentiary record for appellate review, a specific Robinson objection is preferable, if not mandatory, in order to preserve error.
. Even Robinson and its progeny do not have such a requirement.