Plaintiff Jeffrey K. Reed brought this suit against defendant Chrysler Corporation on a crashworthiness theory. At the close of Reed’s case in chief, following five days of testimony, the trial court sustained Chrysler’s motion for directed verdict. We reverse and remand.

The vehicle involved in this accident was a 1980 Jeep CJ-7, manufactured in 1979, which the driver purchased used in 1983. It came with two removable tops, a canvas top and a fiberglass top. Reed was a passenger.

Traveling east on highway 34, approximately two miles west of Albia, the Jeep’s right side tires went off the road onto the gravel-surfaced shoulder as the Jeep approached a bridge. The Jeep turned sharply back onto the paved surface, missing the bridge wall on the Jeep’s right side. It continued across both lanes of the highway, and at the far corner of the bridge the front left side of the vehicle slammed into the concrete bridge abutment. The force of the impact rapidly propelled the occupants, none of whom were wearing seat belts, out of their seats. Upon impact the driver’s door either was forced open or was torn from the vehicle. The driver was .ejected and struck a wooden support pole headfirst, dying instantly.

After the initial impact the vehicle, now driverless, began to ride up the concrete abutment to the bridge. It then continued[*226] riding along the guardrail attached to the abutment. After leaving the guardrail the Jeep proceeded to roll onto its fiberglass top, breaking it. The overturned vehicle slid on its roll bar more than 300 feet farther down the road. It finally came to rest, still upside down. Reed, the right-rear passenger, was held inside the vehicle by his foot, which became entangled with the right front seat. As the vehicle slid upside down, Reed’s right arm was momentarily pinched between the highway surface and the roll bar. The arm was severely fractured and is said to remain useless. Reed also fractured his right collarbone and dislocated his left hip. The other passengers received only minor injuries.

Accident investigators determined the vehicle had been speeding; one expert projected a speed of 79.47 miles per hour, another a speed in the low sixties. Blood tests showed the driver’s alcoholic content at .185 grams of alcohol and Reed’s at .168 (per 100 milliliters of blood).

I. In Iowa the doctrine of crash-worthiness was first recognized as a theory of design defect by the court of appeals in Wernimont v. International Harvestor Corporation, 309 N.W.2d 137 (Iowa App.1981). The doctrine imposes liability on manufacturers for design defects which only enhance injuries rather than cause them. Id. at 140. The doctrine is applicable when a design defect, not causally connected to the accident, results in injuries greater than those which would have resulted from the accident had there been no design defect. In other words, enhancement of injuries is the gist of crashworthiness cases, not the precipitating cause of the accident.

At the time this case was tried, Werni-mont, a court of appeals decision, was the only Iowa authority to recognize the crash-worthiness doctrine. Since then we also recognized it. Hillrichs v. Avco Corp., 478 N.W.2d 70 (Iowa 1991) (holding crashwor-thiness doctrine applicable where plaintiff was injured after becoming entangled in grain harvesting machinery designed and built without an accessible shut-off switch). We said:

The enhanced injury theory as applied to automobile manufacturers has been widely recognized. See Tafoya v. Sears, Roebuck & Co., 884 F.2d 1330, 1337-38 (10th Cir.1989).

[[Image here]]

We believe that the court of appeals’ recognition of the enhanced injury doctrine in Wernimont was merely a correct application in a particular context of well-established elements of Iowa tort law that permit recovery of damages for injuries caused by the conduct of another that the law identifies as tortious. These principles mandate that an injured party’s right of recovery extend to all injury or degree of injury that would have been prevented through the tortfeasor’s exercise of the proper standard of care.

Hillrichs, 478 N.W.2d at 74-75.

To prevail on a claim of crashwor-thiness a plaintiff at the threshold must establish the existence of a design defect. This showing must include proof that the product was unreasonably dangerous. Wernimont, 309 N.W.2d at 140 (citing Chown v. USM Corp., 297 N.W.2d 218, 220 (Iowa 1980)). To prevail, after making the threshold showing, a plaintiff must then establish the following three elements:

(1) proof of an alternative safer design, practicable under the circumstances; (2) what injuries would have resulted had the alternative safer design been used; and (3) the extent of enhanced injuries attributable to the defective design.

Wernimont, 309 N.W.2d at 140-41.

II. In Hillrichs we did not consider whether expert testimony is required to establish the first element (“proof of an alternative safer design, practicable under the circumstances”). Hillrichs happened to present both expert and nonexpert testimony. Wernimont did discuss the issue, the court concluding that: “[wjhether expert testimony is required ultimately depends on whether it is a fact issue upon which the jury needs assistance to reach an intelligent or correct decision.” Wernimont, 309 N.W.2d at 141. The Wernimont court explained that

[*227] design defect cases sometimes involve technical, scientific issues which cannot be fully understood by the average juror without some expert assistance. In such cases, expert testimony as to the defective nature of defendant’s design will be an indispensable element of plaintiffs case. However, when the issues presented relate to matters which require only common knowledge and experience to understand them, the testimony of experts is not essential.

Id. (Quoting Lynd v. Rockwell Mfg. Co., 276 Or. 341, 349, 554 P.2d 1000, 1005 (1976).)

In Hillrichs we did explore the second (“what injuries would have resulted had the alternative safer design been used”) and third (“the extent of enhanced injuries attributable to the defective design”) elements of the crashworthiness doctrine, and said how the two prongs are to be applied:

[Rjecovery should be denied when it is uncertain or speculative that any damage has actually occurred. Damages may be awarded, however, when the only dispute is the amount of damages and the evidence affords a reasonable basis for estimating the loss. In determining a just and reasonable estimate of damages based on relevant data, a jury may act on probable and inferential, as well as direct and positive proof.

Hillrichs, 478 N.W.2d at 75 (citations omitted). Thus, although the extent of the enhanced injury in Hillrichs was not definitively fixed, we said that “the degree of uncertainty [was not] so great as to preclude the jury from quantifying the enhanced loss with a reasonable margin of error.” Id. at 75.

III. Reed’s crashworthiness claim centers on two components of the Jeep CJ-7, its windshield frame and its removable hardtop. We can assume Reed developed sufficient evidence regarding design defect.^[1] If Reed has tendered sufficient proof on the three additional elements of crashworthiness, the case should have been submitted to the jury and directed verdict was improper. Reed believes it was neither the law nor the evidence which kept the case from the jury, but rather the trial judge’s personal bias.^[2]

The trial court’s stated ground for sustaining the motion for directed verdict was a perceived lack of evidence of the first postthreshold element: “an alternative safer design, practicable under the circumstances.” We of course view the evidence in the light most favorable to Reed. Iowa R.App.P. 14(f)(2). We think a jury question was made out on this element.

Company rollover test reports demonstrated that a removable steel top, sold in three versions by Chrysler as an option in at least one other Jeep model, may have prevented the injury to Reed’s arm. Thus, the safer alternative introduced into evidence was familiar to Chrysler, indeed such an alternative had been designed, tested and sold for twenty years prior to the accident.

Reed’s safety expert indicated that fiberglass tops shattered or came apart in most rollover accidents and thus were not crash-worthy. Another expert said that fiberglass was a brittle material with little structural integrity and “not a material that you would use if you [were] interested in safety.” Available metal tops, alone or in conjunction with struts between the windshield and rollbar, were much safer. This lent credence to Reed’s insistence that safer alternatives were available to the manufacturer.

According to Reed’s evidence, at the time of its manufacture, this Jeep CJ-7 was the only such model sold to the American pub-[*228] lie still utilizing a fiberglass hardtop. The industry standard by 1980 was steel tops. The vehicle in which Reed was injured had the only plastic top sold in North America. This is significant because we have long recognized “custom and practice of the industry,” as well as “state of the art,” as a part of tort law. Chown, 297 N.W.2d at 222.

There is more evidence that Chrysler was aware of industry practice. The Jeep CJ-5, a shorter version of the CJ-7, had an optional steel top while the Jeep CJ-7 had only a fiberglass top. According to the manufacturer’s 1980 accessories catalog, CJ-5 Jeep purchasers had the option of three metal tops. A Chrysler engineer acknowledged that the general public would perceive a metal top to be safer than a plastic top. Although the Jeep CJ-7 was not available to the general public with a metal top, this model, so equipped, was sold to the Alaskan post office.

We think a prima facie showing was made on the first element. We think this is true even in the absence of further expert testimony indicating whether an alternative safer design was practicable and feasible in terms of cost, economy of operation, maintenance requirements, and other safety factors. These questions were not so technologically sophisticated as to require expert testimony. The only difference in design between the fiberglass top and the metal top is the material involved. These matters were not beyond the understanding of an average juror.

IV. Although the trial court based its decision on the first post-threshold element of a crashworthiness claim, the case can be affirmed if Reed failed to produce sufficient evidence for a prima facie showing regarding the other two additional elements. The parties dispute the sufficiency of evidence on both the second and third elements: “(2) what injuries would have resulted had the alternative safer design been used; and (3) the extent of enhanced injuries attributable to the defective design.”

A showing on these elements can be satisfied under the liberal requirements, previously mentioned, we established in Hillrichs. 478 N.W.2d at 75 (holding jury’s just and reasonable estimate of damages may be based on relevant data derived from “probable and inferential, as well as direct and positive proof”).

V. The present case differs from Wemimont where the alternative design would have lessened the injury. It is likely that a metal top would have changed the nature of Reed’s injury, a matter that leads to more speculation. We nevertheless think a jury question was presented on the second element (“what injuries would have resulted had the alternative safer design been used”). Reed contends the “but for— causation” showing (but for the fiberglass top, his arm would not have been injured) is sufficient for this element. In other words, although Reed cannot show the injuries that would have occurred with a metal top, he contends he satisfied this element by showing the injury to his arm would not even have occurred with a metal top. We agree.

Chrysler also contends that Reed’s inability to establish the precise injuries he would have suffered with a metal top is fatal to making a showing of the second element. We are not persuaded. The situation is akin to a general rule of damages. Where some damages appear, recovery should not be denied merely because of difficulty in fixing an exact amount. Bangert v. Osceola County, 456 N.W.2d 188, 190[incorrect citation] (Iowa 1990).

Reed’s expert testified that, had the Jeep been designed with a metal top, and had the top stayed attached, Reed would not have suffered his crushing arm injury. The arm injury is shown to be wholly attributable to the fiberglass top. Reed himself testified that $248,000 of the stipulated $258,000 in total damages is attributable to his arm injury. This evidence, taken together, made out a jury question on both the second and third elements.

VI. Even if our review of this case had led to an affirmance, a directed verdict would have been a very bad idea. This is because of what veteran Iowa judges often refer to as the Uhlenhopp rule. As a dis[*229] trict judge the Honorable Harvey Uhlen-hopp, later a distinguished member of this court, faced a familiar situation. A defendant moved for a directed verdict which Judge Uhlenhopp overruled. A verdict was returned for the plaintiff, after which the defendant filed a motion for judgment notwithstanding the verdict on the same ground urged in the motion for directed verdict. Judge Uhlenhopp sustained the motion, saying:

At the close of the evidence the court was clear that [the defendant was entitled to a judgment as a matter of law]. But to avoid the necessity of another trial in the event of error, although it seemed a nonsuit had to be, the case was submitted to the jury.

Florke v. Peterson, 245 Iowa 1031, 1033, 65 N.W.2d 372, 373 (1954). On appeal we said:

We approve the trial court’s procedure. A verdict for defendant might entirely avert appeal; and the verdict for plaintiff serves to avoid the necessity of retrial in the event of reversal. The procedure conforms to the spirit of rule -243(b) [Iowa R.Civ.P.]_ It is in line with the purpose of the legislation under which the rules were conceived — “promoting the speedy determination of litigation upon its merits.”

Id. (Citations omitted.)

Justice Uhlenhopp’s approach has been encouraged in many of our cases. Recently we said:

[W]e note trial court acted appropriately when it allowed the case to go to the jury rather than [granting the] motion for a directed verdict. By allowing the case to go to the jury, trial court gave the jury an opportunity to consider the evidence, return a verdict, and potentially reach the same conclusion the court tentatively had reached. More importantly, this court, assuming no other prejudicial error is found, may now reinstate that verdict rather than being required to remand [the] claim for new trial.

Johnson v. Junkmann, 395 N.W.2d 862, 866 (Iowa 1986) (citations omitted).

We again emphasize that much is wasted by granting directed verdicts in routine cases, or in eases that are at all close. Considerable time and expense have usually been invested in the trial at that point. Submission to a jury is generally not far off. The so-called Uhlenhopp rule is decidedly in the public interest and should normally be followed.

VII. Because the question is likely to recur on retrial, we address Reed’s claim for punitive damages. Because of the directed verdict, the appropriateness of this claim was not reached, although the trial court did offer its view that a claim for punitive damages was not made out.

On the basis of the showing made in this trial, we agree. Chrysler’s conduct was not shown to be willful or wanton. In order to support such a claim, a plaintiff must show actual or legal malice. We reviewed the controlling principles in Coster v. Crookham, 468 N.W.2d 802, 810-11 (Iowa 1991).

VIII. Another issue apt to recur involved Reed’s failure to use a seat belt. The court overruled Reed’s motion in li-mine to exclude evidence of this failure. The motion should have been sustained. Iowa Code section 321.445(4)(a) (1991) provides:

The nonuse of a safety belt or safety harness by a person is not admissible or material as evidence in a civil action brought for damages in a cause of action arising prior to July 1, 1986.

The accident occurred August 18,1985, and is therefore subject to the quoted language.

Chrysler contends this prohibition is modified by language in Iowa Code section 321.445(4)(b). We think not. The legislature, for reasons it chose, enacted a flat prohibition of seat belt evidence in cases arising prior to the time seat belt use was required in Iowa. We are obliged to follow this clear mandate.

IX. Another issue almost certain to reappear on retrial concerns the admissibility of evidence of the driver’s and Reed’s[*230] intoxication. Reed urges the trial court ruling, admitting this evidence, was error.

In former times comparative negligence was inapplicable to strict liability claims. Speck v. Unit Handling Div., Litton Sys., 366 N.W.2d 543, 546 (Iowa 1985) (reaffirmed in Duggan v. Hallmark Pool Mfg. Co., Inc., 398 N.W.2d 175, 180 (Iowa 1986)). This holding was consistent with the prevailing view. See Restatement (Second) of Torts §§ 402A, 524. Both Speck and Dug-gan were decided before the advent of our comparative fault Act^[3] and were therefore not subject to its provisions.

This view of comparative negligence, as it relates to general strict liability claims, was undone by statute. Under the comparative fault Act, Iowa Code § 668.1(1), fault includes acts or omissions that subject a person to strict liability. A plaintiffs fault, under Iowa Code section 668.3, does not bar recovery; it serves only to diminish it. Iowa Code section 668.1(2) leaves intact existing “legal requirements of cause in fact and proximate cause.”

X. The crashworthiness doctrine is a step removed from ordinary strict liability cases. The crashworthiness doctrine proceeds from the belief that a manufacturer has a duty to minimize the injurious effect of a crash, no matter how the crash is caused. Larsen v. General Motors Corp., 391 F.2d 495, 502 (8th Cir.1968) (holding manufacturer liable for “so-called 'second collision’ of the passenger with the interior part of the automobile” because accidents and enhanced injuries are foreseeable). See Harris, Enhanced Injury Theory: An Analytic Framework, 62 N.C.Law Rev. 643, 673-75 (1984).

The theory, which presupposes the occurrence of accidents precipitated for myriad reasons, focuses alone on the enhancement of resulting injuries. The rule does not pretend that the design defect had anything to do with causing the accident. It is enough if the design defect increased the damages. So any participation by the plaintiff in bringing the accident about is quite beside the point.

In adopting the crashworthiness doctrine in Hillrichs, we indicated a contrary view, stating that the plaintiffs comparative fault could be assessed against him in a claim for enhanced injuries. 478 N.W.2d at 76. On reconsideration, for the reasons just stated, we think a plaintiffs comparative fault should not be so assessed in a crashworthiness case unless it is shown to be a proximate cause of the enhanced injury-

Chrysler points to no way, and the record suggests none, in which the driver’s or Reed’s intoxication could bear on how Reed’s injuries were enhanced by the construction of the Jeep’s roof. Any negligence by Reed’s driver, or even by Reed himself, in connection with the original crash cannot be used by the manufacturer in defending against Reed’s enhancement claim. Evidence of intoxication should have been excluded.

Other arguments relate to issues either not preserved or not likely to recur on retrial.

REVERSED AND REMANDED.