Plaintiff Michael Montag and Plaintiff-Intervenor United of Omaha Life Insurance Company (“Plaintiffs”) appeal from a judgment against them in this products liability action. Plaintiffs claim that Defendants American Honda Motor Company, Honda Motor Co. Ltd., and Honda R & D Co., Ltd. (“Honda”) defectively designed the seat belt in a 1988 Honda Prelude driven by Diane Montag, Mr. Montag’s deceased wife. Mrs. Montag was involved in a collision and was thrown from her car despite the fact that she was wearing a seat belt. The jury rendered a verdict for Honda. Plaintiffs now assert that the trial court made various errors during the course of the trial. We affirm for the reasons that follow.
On her way to work one morning, Diane Montag drove her 1988 Honda Prelude onto a railroad crossing. A freight train was approaching the intersection at the same moment and broadsided the Prelude. Although she was wearing her seat belt, Mrs. Montag was thrown from her vehicle. She received serious brain injuries from which she died twenty-one months later.
Plaintiffs do not claim that Honda is responsible for the initial collision with the train. Rather, Plaintiffs assert that but for the defectively designed seat belt, Mrs. Mon-tag would not have been ejected from her car. The seat belt in question was designed to be automatic: when the door opened, the seat belt automatically disengaged; and when the door closed, the seat belt automatically moved into place, securing both torso and waist. The force of the collision with the train caused Mrs. Montag’s door to open, and, as a result, the seat belt automatically disengaged allowing Mrs. Montag to be ejected from the car. Plaintiffs claim that such a seat belt system is defective under products liability law.
Plaintiffs present several issues on appeal. First, they claim the trial court erred when it concluded that federal motor safety regulations preempted any claim that an airbag was an available alternative safety design. Second, they claim Honda was negligent per se because it failed to comply with certain federal motor safety standards. Thus, they argue the district court erred when it instructed the jury that the seat belt was presumed not to be defective under Colorado law because Honda had complied with all applicable statutes and regulations. Third, they claim the district court erred when it refused to instruct the jury on the consumer expectations test employed in products liability cases. Fourth, Plaintiffs claim the district court erred when it instructed the jury to compare the negligence of Mrs. Montag with the negligence of Honda under Colorado’s comparative fault statute. Finally, they argue the district court made erroneous evidentiary rulings when it allowed Honda to show the jury a film depicting a train-automobile crash, and when the district court restricted the scope of Plaintiffs’ lay witness testimony.
I. Federal Preemption
As part of its defective design case, Plaintiffs sought to introduce evidence of a safer, alternative design. Specifically, Plaintiffs intended to show that Honda could have used an alternative restraint system containing, in part, an air bag. Honda filed a motion in *1417limine seeking to exclude the air bag evidence. It argued that federal motor safety-regulations preempted Plaintiffs’ air bag evidence. The district court agreed and granted Honda’s motion. Plaintiffs claim that this decision was erroneous.
We have previously held that 15 U.S.C. § 1392(d) of the National Traffic and Motor Vehicle Safety Act (the “Safety Act”), 15 U.S.C. § 1381 et seq., and Federal Motor Vehicle Safety Standard 208, 49 C.F.R. § 571.208 (“FMVSS” or “Standard” 208) impliedly preempt air bag claims.1 Kitts v. General Motors Corp., 875 F.2d 787, 789 (10th Cir.1989), cert. denied, 494 U.S. 1065, 110 S.Ct. 1781, 108 L.Ed.2d 783 (1990). Plaintiffs attempt to circumvent this holding by citing to Cipollone v. Liggett Group, Inc., 505 U.S. 504, 112 S.Ct. 2608,120 L.Ed.2d 407 (1992). They argue that, under Cipollone, courts may not consider implied preemption where Congress has provided an express preemption clause. Because the Safety Act contains an express preemption clause in § 1392(d), they reason, the Kitts implied preemption holding does not apply. The Supreme Court, however, recently rejected this interpretation of Cipollone. In Freightliner Corp. v. Myrick, — U.S. -, 115 S.Ct. 1483, 131 L.Ed.2d 385 (1995), the Court held that the existence of an express preemption provision does not preclude an implied preemption analysis. Id. at-, 115 S.Ct. at 1488. At best, an express preemption clause supports the inference that no implied preemption exists. Id.
The question remains, then, whether the express preemption clause in the Safety Act precludes an implied preemption analysis in this case. In Myrick, the Supreme Court engaged in an implied preemption analysis of the Safety Act. Although not explicitly stated, the Court clearly believed that the express preemption clause of the Safety Act did not preclude implied preemption analysis. Thus, our holding in Kitts still applies to this case. Plaintiffs were impliedly preempted from presenting airbag evidence.
II. Negligence Per Se
Colorado Revised Statutes § 13-21-403(l)(b) creates a rebuttable presumption that a product is not defective if the manufacturer complied with applicable federal or state regulations. The district court, holding that Honda had met the applicable federal safety standards, instructed the jury as to this presumption. Plaintiffs argue, however, that Honda did not comply with Federal Motor Vehicle Safety Standard 209, 49 C.F.R. § 571.209 (“FMVSS” or “Standard” 209).2 Thus, they argue not only that the district court’s instruction was erroneous but also that Honda was negligent per se.
FMVSS 209 S4.1(b) provides:
A seat belt assembly shall provide pelvic restraint whether or not upper torso restraint is provided, and the pelvic restraint shall be designed to remain on the pelvis under all conditions, including collision or roll-over of the motor vehicle.
Because the seat belt on Mrs. Montag’s car did not provide pelvic restraint during the collision, Plaintiffs assert that Honda did not comply with FMVSS 209. Honda, on the other hand, argues that FMVSS 208 specifically authorizes the automatic seat belt used in Mrs. Montag’s Prelude. Furthermore, they argue that seat belts in compliance with FMVSS 208 are exempt from the requirements of FMVSS 209. Thus, Honda believes it complied with the applicable federal standards.
FMVSS 208 sets forth the various seat belt options manufacturers may use to meet occu*1418pant crash protection standards. Specifically, FMVSS 208 S4.5.3 provides:
“[A] seat belt assembly that requires no action by vehicle occupants (hereinafter referred to as an “automatic belt”) may be used to meet the crash protection requirements of any option under S4. and in place of any seat belt assembly otherwise required by that option.”
Plaintiffs do not contend that Honda’s seat belt was not an “automatic belt” within the meaning of FMVSS 208 S4.5.3. Thus, the question becomes whether an automatic seat belt which otherwise complies with FMVSS 208 must also meet the requirements of FMVSS 209 S4.1(b). These two standards are in conflict because of the nature of an automatic seat belt. Automatic belts are designed to release when the ear door opens. If the car door opens during a collision, however, the automatic belt still releases. Under these circumstances, the seat belt does not provide pelvic restraint, in direct contravention of FMVSS 209.
Upon further examination of FMVSS 208, we resolve this conflict in favor of Honda. It appears that automatic seat belts complying with FMVSS 208, with certain exceptions, are exempt from the requirements of FMVSS 209. Specifically, FMVSS 208 54.5.3.4 provides:
An automatic belt furnished pursuant to [208] S4.5.3 that is not required to meet the perpendicular frontal crash protection requirements of [208] S5.1 shall conform to the webbing, attachment hardware, and assembly performance requirements of Standard No. 209.
The implication of this regulation is that its inverse also applies: automatic seat belts which must meet the crash protection requirements of Standard 208 S5.1 are not required to conform to Standard 209. This reading is supported by legislative history and by informal interpretive letters of the National Highway Traffic Safety Administration (“NHTSA”). The NHTSA explained the amendment to Standard 208 which added 54.5.3.4 by stating “On reconsideration, the NHTSA has decided that relief from Standard No. 209 should be afforded if a passive belt is capable of meeting the occupant crash protection requirements of S5.1 in a frontal perpendicular impact and amends S4.5.3 accordingly.” 36 F.R. 23725 (Dec. 14, 1971). In an informal interpretive letter, the NHTSA applied this statement when it concluded that automatic seat belts which comply with Standard 208 need not comply with Standard 209. Plaintiffs’ App., vol. I, at 129 (letter from the NHTSA to private individual, 9/10/90). Plaintiffs concede that the Prelude’s seat belt was required to meet the frontal crash protection requirements of Standard 208 S5.1. Appellants’ Br. at 28. Thus, Standard 209 does not apply to the Prelude’s seat belt.
Plaintiffs try to circumvent this result by pointing out that the language in FMVSS 208 S4.5.3 only addresses the webbing, attachment hardware, and assembly performance requirements of FMVSS 209. These requirements are contained in FMVSS 209 S4.2, S4.3, and S4.4, respectively. Plaintiffs argue that because FMVSS 209 S4.1(b) (requiring pelvic restraint) was not specifically addressed in FMVSS 208 S4.5.3, it still applies. This interpretation of the regulations, however, ignores the intent of FMVSS 209. Standard 209 sets forth the technical requirements for the component parts of a seat belt. As expressed in S4.1(b), these component parts should be able to provide pelvic restraint. In effect, S4.1(b) states the general purpose of Standard 209. The webbing, attachment hardware, and assembly performance requirements merely set forth with specificity the conditions under which a seat belt must be able to provide this pelvic restraint. It would be illogical to hold that the specific conditions do not apply but that the general condition does apply. Such a result would make FMVSS 209 internally inconsistent. For this reason, we conclude that FMVSS 208 S4.5.3 includes the pelvic restraint requirement in FMVSS 209 S4.1(b). Thus, the district court correctly concluded that Honda had complied with the applicable federal safety standards.
III. Consumer Expectations Test
The Colorado Supreme Court has held that complex product liability claims involving primarily technical and scientific *1419information require use of a risk-benefit test rather than a consumer expectations test. Camacho v. Honda Motor Co., 741 P.2d 1240, 1246-48 (Colo.1987). The court has listed several factors which should be considered under a risk-benefit analysis but has also noted that this list is not exclusive. Armentrout v. FMC Corp., 842 P.2d 175, 184 (Colo.1992).
Plaintiffs requested that the district court include the consumer expectations test in the jury instructions as one of the factors to be considered in a risk-benefit analysis. The district court refused to give a consumer expectations instruction. Plaintiffs claim this was error. They argue that the consumer expectations test can be used in a risk-benefit analysis because the risk-benefit factors listed by the Colorado Supreme Court are non-exclusive. This novel argument ignores, however, the Colorado Supreme Court’s express statement that the consumer expectations test is inappropriate in cases involving technical and scientific information. Camacho, 741 P.2d at 1246-47. Plaintiffs do not argue that this case does not involve technical and scientific information. Thus, the consumer expectations test should not be used in this case. Engrafting it onto the risk-benefit would render meaningless the express holding of Camacho. The district court did not err when it refused to give the requested instruction.
IV. Comparative Fault
Colorado Revised Statutes § 13-21-406 provides:
In any product liability action, the fault of the person suffering the harm, as well as the fault of all others who are parties to the action for causing the harm, shall be compared by the trier of fact in accordance with this section.
Plaintiffs admitted that Mrs. Montag was negligent in driving her ear into the path of an oncoming train. As a result, the district court instructed the jury that they could compare the fault of Mrs. Montag with the fault of Honda. The district court denied Plaintiffs’ objection to this instruction.
Plaintiffs argue that C.R.S. § 13-21-406 does not apply to Mrs. Montag’s negligence because this is a “crashworthiness” or “second collision” case. Under the crashworthiness doctrine, an automobile manufacturer may be liable “for injuries sustained in a motor vehicle accident where a ... design defect, though not the cause of the accident, caused or enhanced the injuries.” Camacho v. Honda Motor Co., 741 P.2d 1240, 1242-43 (Colo.1987). From this, Plaintiffs reason that the initial accident is irrelevant to their cause of action. They seek damages only for the ' enhanced injuries. Thus, they claim Mrs. Montag’s negligence in causing the initial accident cannot be compared to Honda’s defective design of the seat belt. Rather, Plaintiffs assert that only Mrs. Montag’s misuse of the product, ie., the seat belt, can be compared to Honda’s fault.
We have previously recognized, however, that the term “fault” in § 13-21-406 should be given a broad reading. Huffman v. Caterpillar Tractor Co., 908 F.2d 1470 (10th Cir.1990). In Huffman we found that “[t]he term “fault,” as employed in C.R.S. 13-21-406, is more plausibly construed as a general term encompassing a broad range of culpable behavior including, but not limited to, negligence.” Id. at 1477. Furthermore, fault is not limited to assumption of risk or product misuse. Id. Given this broad interpretation of the word “fault,” no good reason exists not to allow the jury to compare Mrs. Montag’s initial negligence with Honda’s fault in designing the seat belt. In every crashworthiness case, the jury will be required to determine how much of a plaintiffs injuries resulted from the initial collision and how much of the injuries were the result of a second collision. In this case, the jury was required to determine which of Mrs. Mon-tag’s injuries resulted from her initial collision with the train and which of her injuries resulted from the allegedly defective seat belt. Thus, to an extent, the jury is already comparing the plaintiffs and the defendant’s behavior in order to determine causation. Requiring the jury to make a similar determination for the purpose of damages is certainly reasonable and consistent with Colorado’s comparative fault statute.
*1420V. Crash Test Videotape
At trial, Honda introduced a videotape depicting a collision between a train and an automobile. It did so for the limited purpose of demonstrating the physical forces at play in a train-automobile accident. Plaintiffs objected, arguing that the tape was inadmissible because it was substantially dissimilar to the actual accident and because it was inflammatory and misleading under Fed. R.Evid. 403. The district court denied Plaintiffs’ objections.
Tests and experiments may be dissimilar from the actual accident if they are meant to illustrate general physical principles rather than re-create the accident. Robinson v. Missouri Pac. R. Co., 16 F.3d 1083, 1087 (10th Cir.1994); Gilbert v. Cosco Inc., 989 F.2d 399, 402 (10th Cir.1993). If a simulation is offered, however, for the limited purpose of illustrating general principles, the district court should instruct the jury that the evidence is not intended as a re-creation of the accident. Gilbert, 989 F.2d at 402. Here, Honda introduced the videotape for the limited purpose of portraying the forces at work in a train-automobile collision. The district court specifically instructed the jury not to consider the videotape as a re-creation of Mrs. Montag’s accident. Given the district court’s broad discretion in this area, we cannot say that it erred in allowing the jury to view the videotape.
Similarly, a review of the videotape indicates that it was not so inflammatory or misleading as to require us to reverse the district court. Although the tape does repeatedly show the tram’s plow handle striking a dummy’s head, any prejudicial effect was countered by the district court’s limiting instruction and by Plaintiffs’ opportunity for vigorous cross-examination. See Robinson, 16 F.3d at 1088. Additionally, the jury’s verdict indicates that it was not prejudiced by the videotape. The jury specifically found that Mrs. Montag’s injuries did not result from the initial collision with the train. See id. (stating that verdict indicated jury was not prejudiced by accident re-creation). Thus, the district court did not abuse its discretion when it admitted the videotape.
VI. Lay Witness Testimony
Lastly, Plaintiffs argue that the district court improperly restricted the proposed testimony of their lay witness. This witness had been in an accident similar to Mrs. Montag’s accident. While driving a 1987 Honda Accord with a seat belt similar to the seat belt in Mrs. Montag’s Honda Prelude, the witness’s car was struck in the side and she was ejected from the car despite wearing her seat belt. Plaintiffs wanted the witness to testify as to how the accident occurred, how she was ejected, and that the seat belt system was “unsatisfactory” and “poorly conceived.” Appellants’ Br. at 37. Plaintiffs wanted to present this testimony in order to show notice of a defect, to show existence of a defect, and to refute Honda’s expert testimony. Plaintiffs also wanted to use this testimony as evidence of “what a consumer could reasonably expect under the consumer expectation test.” Id. at 38.
The district court decided it would allow the witness to testify with certain restrictions. First, the court restricted the witness from exceeding the scope of lay opinion:
I don’t have any problem with her saying her car was going fast and letting them ask her the question how fast is fast. But if she starts testifying that the restraint system was defective or that the vehicle was poorly designed or unsatisfactory, that goes far beyond a lay opinion as to what she’s testifying about.
Appellants’ App., vol. II, at 432. Second, the district court warned counsel that the witness could not “editorialize”:
I will not allow any editorializing about what happened to her as far as other than she was ejected from the vehicle and she found when she returned to the vehicle that her seat belt was buckled.
... I will absolutely restrict — and you tell her that she’ll be held in contempt if she gets into any commentary or any opinions about what a crummy vehicle this was or her — I forget what her editorializing about Honda’s seat belt system was. But I’ll not tolerate it, period____
She will testify strictly as to the facts of the impact and the fact that she was ejected from the vehicle.
*1421Id., vol. Ill, at 636. Based on these restrictions, Plaintiffs decided not to present the witness’s testimony.
The district court did not abuse its discretion by imposing these restrictions on the witness’s testimony. As we have already held, the consumer expectations test does not apply to this case. As a result, testimony from a lay witness that the seat belt system was “poorly conceived” or “unsatisfactory” would run the risk of presenting improper consumer expectations evidence. Furthermore, testimony from the witness that the seat belt system was “defective” or “poorly conceived” would intrude into the realm of the expert witness. See, e.g., Randolph v. Collectramatic, Inc., 590 F.2d 844, 846-48 (10th Cir.1979) (upholding trial court’s refusal to allow lay witness to testify that pressure cooker was “defectively designed”). In certain circumstances, a lay witness with specialized knowledge may be allowed to testify as to topics typically reserved for expert witnesses. See id. at 848. Plaintiffs, however, have not shown that their lay witness had any type of specialized knowledge. Thus, the district court did not err when it precluded her from testifying about “defects” in the seat belt system. Finally, the district court’s concern with the witness’s “editorializing” appears to be nothing more than an attempt to keep her testimony within the proper confines of a lay witness. Although the district court may have been somewhat restrictive, its action did not rise to the level of an abuse of discretion. See, e.g., Wheeler v. John Deere Co., 862 F.2d 1404, 1408 (10th Cir. 1988) (stating that no abuse of discretion occurred even if district court’s ruling was unduly restrictive).
For the foregoing reasons, we AFFIRM.
. Section 1392(d) provides:
Whenever a Federal motor vehicle safety standard established under this subchapter is in effect, no State or political subdivision of a State shall have any authority either to establish, or to continue in effect, with respect to any motor vehicle or item of motor vehicle equipment any safety standard applicable to the same aspect of performance of such vehicle or item of equipment which is not identical to the Federal standard.
FMVSS 208 outlines the seat belt options available to automobile manufacturers.
. Although Plaintiffs also argue that Honda did not comply with FMVSS 208, this argument appears to be based on Plaintiffs' belief that FMVSS 208 requires compliance with FMVSS 209. Thus, in effect, Plaintiffs only question the applicability of Standard 209.