Shelley A. JODOIN, Individually and as Natural Parent, Next Friend and Guardian of Timothy D. Dorsey, a Minor, and Lawrence H. Jodoin, Plaintiffs, Appellants, v. TOYOTA MOTOR CORPORATION, Toyota Motor Sales U.S.A., Inc., Defendants, Appellees, 284 F.3d 272

TORRUELLA, Circuit Judge. Shelley Jodoin was injured in a car accident and, with her husband and son, sued Toyota Motor Corporation and Toyota Motor Sales U.S.A., Inc. (“Toyota”) alleging a design defect in her vehicle. During trial, the district court excluded all evidence relating to testing done by plaintiffs’ expert. After this ruling, plaintiffs conceded that they would be unable to prove defect, an element of their prima facie case, and the district court granted judgment as a matter of law in favor of Toyota. On appeal, plaintiffs challenge the district court’s exclusion of the testing evidence.

We vacate the judgment in favor of Toyota and remand for a new trial. I. On October 6, 1995, plaintiff-appellant Shelley Jodoin was hit from behind as she drove her 1988 Toyota 4x4 pick-up truck. The impact pushed her vehicle into a counter-clockwise turn. She attempted to correct the course of her truck, but as she turned to the right, her truck flipped, rolling over several times. As a result of the accident, Mrs. Jodoin is permanently paralyzed. Mrs. Jodoin, her husband, and her son brought suit against Toyota alleging a design defect in Mrs. Jodoin’s truck which made it prone to rollover.

At trial, plaintiffs relied on the testimony of their primary liability expert, Robert Loyd Anderson, to prove this defect. They had employed Mr. Anderson to perform an accident reconstruction and test another 1988 Toyota 4x4 truck to determine its rollover propensity. Mr. Anderson was allowed to testify about his accident reconstruction conclusions. However, when plaintiffs attempted to introduce Mr. Anderson’s testimony regarding the testing of the exemplar vehicle, the court refused to allow the testimony for lack of a proper foundation.

The court reasoned as follows: [Yjou’ve got a big problem here that you can’t remedy ... we’re not interested in the date of manufacture what these two vehicles were like. What we’re interested in is what they were like at the time of the ... accident, and whether the exemplar was the same. And we don’t know what condition the exemplar was or what it went through, what its history was. For example, whether it had been in an accident previously, which weakened some structures and other factors. And this witness can’t testify to that. He knows nothing about the history of the vehicle. So you’re wasting your time.

All of this is irrelevant until you establish that the exemplar was virtually identical in all respects with the subject vehicle. And only then can you get into the question of what tests were run.... You can’t get there from here, I can tell you now, not with this witness. The court suggested that plaintiffs could lay a proper foundation by introducing testimony from the people who purchased the car for Mr. Anderson or the people from whom the exemplar vehicle was purchased. Plaintiffs introduced no such testimony. Therefore, Mr.

Anderson was not allowed to testify as to how the design of the exemplar vehicle compared with the design of Mrs. Jodoin’s vehicle. Plaintiffs did try to lay a foundation based on Mr. Anderson’s testimony. First, Mr. Anderson claimed to have checked the vehicle identification tags to verify that the load ratings and tires were the same. Furthermore, the record reflects that the exemplar and Mrs. Jodoin’s vehicle had similar vehicle identification numbers (“VIN”).

Second, he testified to performing a structural examination of the steering components, suspension components, tires, and springs, including crawling under the truck to inspect the undercarriage. Third, he said he had looked at the instrumentation and modifications made for the purposes of testing, which he documented. This examination included “[everything [Mr. Anderson] thought ... would be related to the vehicle dynamics and the issues that [he] was evaluating.” He testified that he detected no evidence of any modifications or any parts that were not Toyota’s original equipment. However, Mr.

Anderson had no personal knowledge of where the exemplar vehicle came from or how it was obtained, and he did not testify to that history. When plaintiffs attempted to question Mr. Anderson regarding the relationship between the design characteristics of Mrs. Jodoin’s vehicle and its rollover stability, the court upheld an objection to the testimony, stating, “obviously [Mr.

Anderson’s testimony is] based on testing; and the results of the testing is not admissible at this point, [sic] it never will be.” When plaintiffs had previously attempted to introduce testimony on the general relationship between a vehicle’s design features and its propensity to rollover, the court excluded that testimony as irrelevant. Therefore, plaintiffs were unable to introduce any testimony regarding design and rollover propensity. The next day, outside the presence of the jury, plaintiffs addressed the court and asked it to reconsider its ruling.

The court again pointed to the lack of information on the exemplar vehicle’s history and repeated that it would exclude any evidence relating to testing of that vehicle without such a history. The court said that plaintiffs needed to “show ... where [the exemplar vehicle] was purchased [and] have some evidence as to whether that’s a legitimate VIN number on [the exemplar vehicle], and somebody who has the expertise to be able to tell us what the VIN number means ... [a]nd whether there’s been any changes in the vehicle since its manufacture.” In response, plaintiffs made an offer of proof regarding what Mr.

Anderson would have testified to regarding the similarities between the exemplar vehicle and Mrs. Jodoin’s vehicle. This included, mainly, Mr. Anderson’s conclusion, based on the VINs of the two vehicles, that the vehicles were “virtually identical.” It also contained his conclusion that the exemplar vehicle was in good condition and “reasonably similar to the kind of condition” he would expect for a vehicle that had not been involved in any accidents, damaged or modified.

Toyota opposed the offer of proof and contended that plaintiffs could never clear the substantial similarity hurdle in regard to the exemplar vehicle. In support, Toyota cited information that the exemplar vehicle had been sent to a dealership for extensive repairs after it had been acquired for Mr. Anderson’s tests; yet Mr. Anderson had no information on those repairs. During the offer of proof, when plaintiffs attempted to introduce evidence relating to the testing, the court stopped them, maintaining the need for a Daubert v.

Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), hearing before the testing or test results could be entered on the record. Plaintiffs declined to hold the hearing because the court iterated that the test results could not be admitted for lack of an adequate foundation, regardless of the outcome of the Daubert hearing. Therefore, the record contains no information on Mr. Anderson’s findings. After the district court reinforced its ruling relating to the inadmissibility of Mr.

Anderson’s testimony regarding the exemplar vehicle, plaintiffs acknowledged they would be unable to prove defect, an element of their prima facie case. The court then asked plaintiffs if they were going to dismiss the case. Plaintiffs responded that they would not do so voluntarily. Then, the court asked, “Are you going to rest your case so I can grant a motion for a directed verdict?” After taking a short recess, plaintiffs accepted the court’s suggestion and rested their case.

Toyota moved for judgment as a matter of law based on plaintiffs’ failure to produce evidence of defect, and the court granted the motion. This appeal followed. II. On appeal, plaintiffs contend that the district court abused its discretion by excluding all testimony regarding the testing of the exemplar vehicle. However, before we can reach that question, we must resolve a threshold matter: whether plaintiffs essentially abandoned their case when they rested before proving any of the elements of their claim.

Because the district court suggested, incorrectly, that plaintiffs needed to rest before the court could consider a Rule 50 motion, we find that, under these facts, plaintiffs did not “abandon” their case.

Under Rhode Island law, plaintiffs must prove five elements, in addition to damages, to prevail in a strict liability claim based on design defect: (1) a defect; (2) the defect existed at the time the product left defendants’ hands; (3) the defect rendered the product unreasonably dangerous; (4) the product was being used as intended at the time of the accident; and (5) the defect was the proximate cause of plaintiffs’ injuries. Raimbeault v. Takeuchi Mfg. (U.S.) Ltd., 772 A.2d 1056, 1063 (R.I.2001). Plaintiffs do not dispute that they failed to produce evidence of causation and damages, not just defect.

Therefore, Toyota contends, it was proper for the district court to grant the Rule 50 motion, regardless of the district court’s evidentiary ruling. Under Toyota’s argument, even if we were to find that the district court abused its discretion by excluding the testing evidence on the exemplar vehicle, plaintiffs still failed to establish the other elements of their case, and, therefore, we cannot reverse the district court’s judgment. The district court entered judgment in favor of Toyota under Rule 50(

a) of the Federal Rules of Civil Procedure. This rule provides that a court may enter judgment as a matter of law after a party has been “fully heard on an issue and there is no legally sufficient evidentiary basis for a reasonable jury to find for that party on that issue.” Fed.R.Civ.P. 50(a). The rule allows the court to entertain a Rule 50 motion at “any time before submission of the case to the jury.” Id. (emphasis added); see also Am. & Foreign Ins. Co. v. Gen. Elec.

Co., 45 F.3d 135, 139 (6th Cir.1995) (holding it proper for a trial court to entertain motions for judgment as a matter of law at any time during trial, not just at the close of a party’s evidence). The advisory committee specifically intended the rule to authorize “the court to consider a motion for judgment as a matter of law as soon as a party has completed a presentation on a fact essential to that party’s case.” Fed.R.Civ.P. 50 advisory committee’s note.

Toyota argues that allowing plaintiffs to appeal the district court’s evidentiary ruling at this point essentially permits an interlocutory appeal, which is at odds with the goal of judicial efficiency. We disagree. The advisory committee specifically contemplated situations like this and stated that “such early action is appropriate when economy and expedition will be served.” Id. Here, the district court ruled that Mr. Anderson would not be able to testify regarding his testing of the exemplar vehicle, and plaintiffs had no other evidence to prove defect.

Therefore, it was entirely appropriate for the district court to consider and grant Toyota’s Rule 50 motion for judgment as a matter of law. However, plaintiffs rested their entire case before the court entered judgment. This, Toyota contends, makes it a more difficult case because the judgment was entered on more than plaintiffs’ failure to prove defect. Plaintiffs also failed to prove other elements of their prima facie case, and those elements, arguably, were not affected by the district court’s exclusion of testimony regarding the exemplar vehicle. We do not find this distinction dis-positive, here.

It is clear from the record that the district court told plaintiffs that they had to rest before it would consider any dispositive motions. Plaintiffs, therefore, faced the option of continuing to present evidence, knowing that the additional evidence would have no bearing on the eventual outcome, or resting their case.

Since Rule 50 allows a judge to issue judgment as a matter of law at any point, once it is clear that a party cannot prevail, we decline to hold that plaintiffs forfeited their entire case when they followed the court’s direction and rested their case, knowing that they could not prevail should they continue. In doing so, plaintiffs did not waive their right to appeal the court’s evidentiary rulings. III. Plaintiffs challenge the district court’s ruling excluding all evidence relating to the testing of the exemplar vehicle. This evidence consists principally of Mr. Anderson’s testimony.

The district court deemed the evidence irrelevant unless plaintiffs could show that the exemplar vehicle was “virtually identical” to Mrs. Jodoin’s truck. Because we find that the district court employed the wrong legal standard, we conclude that the district court abused its discretion by summarily excluding the evidence relating to the testing of the exemplar vehicle and that this error was not harmless. The Federal Rules of Evidence establish a low threshold for relevance, generally.

However, relevant evidence may be excluded if its probative value is “substantially outweighed” by its likelihood to confuse the issue or mislead the jury. Fed.R.Evid. 403. In this regard, courts have treated with skepticism evidence that seeks to recreate accidents. See, e.g., Swajian v. Gen. Motors Corp., 916 F.2d 31, 36 (1st Cir.1990) (upholding exclusion of a videotape test which portrayed the consequences of a car’s axle fracturing). They have not, however, excluded all such evidence. See, e.g., Robbins v.

Whelan, 653 F.2d 47, 49-50 (1st Cir.1981) (overturning district court’s exclusion of report documenting stopping distances for various vehicles). When a party introduces evidence that attempts to reconstruct an accident, that party must show a “substantial similarity in circumstances” between the reconstruction and the original accident. Fusco v. Gen. Motors Corp., 11 F.3d 259, 264 (1st Cir.1993). In contrast, a party may introduce evidence that simply illustrates general scientific principles. See id. Then, we simply inquire whether the test on which the evidence is premised was “properly conducted.” Id.

Differentiating between recreations and illustrations of general scientific principles can be difficult. See McKnight v. Johnson Controls, Inc., 36 F.3d 1396, 1402 (8th Cir.1994). Generally, we look to whether the evidence “is sufficiently close in appearance to the original accident to create the risk of misunderstanding by the jury, for it is that risk that gives rise to the special requirement to show similar conditions.” Fusco, 11 F.3d at 264. Here, Mr. Anderson tested a vehicle of the same make and model year as Mrs. Jodoin’s truck.

He was also prepared to testify about the rollover propensity of the vehicle based on these tests. Because the two trucks are facially similar, we believe that a jury would likely view the testing as a reconstruction of the actual accident, not as simply illustrative of scientific principles. See McKnight, 36 F.3d at 1402-03 (holding that tests performed on a battery of the same type and make which were used to explain what happened when the subject battery exploded “clearly were not limited to a demonstration of scientific principles in the abstract”).

Therefore, the proper test is the substantial similarity standard. See Fusco, 11 F.3d at 264. When reviewing the district court’s application of the substantial similarity test, we accord substantial deference to the trial court, looking only for an abuse of discretion. See Udemba v. Nicoli, 237 F.3d 8, 14 (1st Cir.2001). While this accords the district court considerable latitude, it is not a toothless standard. See Espeaignnette v. Gene Tierney Go., 43 F.3d 1, 5 (1st Cir.1994). An error of law, underlying the evidentiary ruling, constitutes an abuse of discretion. See Koon v.

United States, 518 U.S. 81, 100, 116 S.Ct. 2035, 135 L.Ed.2d 392 (1996) (“A district court by definition abuses its discretion when it makes an error of law.”); see also United States v. Kayser-Roth Corp., 272 F.3d 89, 100 (1st Cir.2001). Here, the district court never specifically applied the substantial similarity standard. Instead, it announced several times that plaintiffs needed to show that the exemplar vehicle was “virtually identical” to Mrs. Jodoin’s vehicle. “Virtually identical” is an incorrect standard. See Robbins v.

Whelan, 653 F.2d 47, 49 (1st Cir.1981) (holding that “perfect identity” is incorrect standard); see also Randall v. Warnaco, Inc., Hirsch-Weis Div., 677 F.2d 1226, 1233-34 (8th Cir.1982) (“Admissibility, however, does not depend on perfect identity between actual and experimental conditions. Ordinarily, dissimilarities affect the weight of the evidence, not its admissibility.”); accord Szeliga v. Gen.

Motors Corp., 728 F.2d 566, 567 (1st Cir.1984) (holding that “[dissimilarities between experimental and actual conditions affect the weight of the evidence, not its admissibility,” but not specifically applying the substantial similarity standard). Therefore, the district court abused its discretion when it required plaintiffs to demonstrate that the exemplar vehicle was “virtually identical” to Mrs. Jodoin’s truck. Nevertheless, we will not reverse the district court judgment if the error was harmless.

See Fed.R.Evid. 103(a) (“Error may not be predicated upon a ruling which admits or excludes evidence, unless a substantial right of the party is affected”); see also United States v. Meserve, 271 F.3d 314, 329 (1st Cir.2001). The error is not harmless if the record indicates that plaintiffs offered sufficient proof that the exemplar vehicle was substantially similar to Mrs. Jodoin’s vehicle.

As this evidence is plaintiffs’ only proof of defect, any improper exclusion adversely affects their substantial rights. “Substantial similarity depends upon the underlying theory of the case.” Four Corners Helicopters, Inc. v. Turbomeca, S.A., 979 F.2d 1434, 1440 (10th Cir.1992). We have looked to the specific variables in various accidents when determining whether the recreation is substantially similar to the original accident.

Compare Swajian, 916 F.2d at 36 (focusing on the fact that the driver during a recreation was a professional driver who knew the axle was going to fracture when driver response to an alleged axle fracture was a key element in the original accident), with Robbins, 653 F.2d at 49-50 (reversing exclusion of test data when the only suggested difference between re-enactment and actual accident was the skill level of the drivers and the trial issue revolved solely around using length of skid marks to estimate the car’s original speed).

When the relevant elements are sufficiently similar, we further emphasize that other differences are for defendants to highlight and the jury to weigh in its deliberations. Robbins, 653 F.2d at 50. Here, plaintiffs alleged a design defect based on the rollover propensity of Mrs. Jodoin’s truck. At this point, only the characteristics of the truck are at issue, not the characteristics of the test. The evidence presented shows that the two vehicles were essentially the same at the time of manufacture.

The question, then, is whether the exemplar vehicle had suffered alterations or damage which could affect its rollover propensity prior to any testing. Mr. Anderson testified that he personally inspected the exemplar vehicle for everything that “would be related to the vehicle dynamics and the issues that [he] was evaluating.” Based on that examination, he found no evidence of any non-original equipment or modifications. Plaintiffs, in their offer of proof, submitted that Mr. Anderson could further testify that the truck showed no evidence of having been in any accidents, otherwise damaged or modified.

The district court, however, required that plaintiffs introduce evidence of the exemplar vehicle’s history. We see no reason such information would need to be presented in order to show substantial similarity. No cases suggest such a requirement. Bogosian v. Mercedes-Benz of N. Am., Inc., 104 F.3d 472, 480 (1st Cir.1997), on which Toyota relies heavily, is inapposite to the current case.

There, the court excluded testing evidence performed on the same car when the parties were unable to show that the car had not been materially changed in the two years since the accident and after that car had been examined by numerous experts in the intervening period. Id. Here, plaintiffs’ expert testified that he had performed a thorough inspection of the exemplar vehicle and detected no evidence of accidents, damage or modification. If the evidence suggested the exemplar had been altered in some material respect, the district court might legitimately require a more complete vehicle history.

However, the district court, sua sponte, instituted this requirement even though nothing in the record suggested any alteration to the exemplar. We are, nonetheless, troubled by Toyota’s allegation that undisclosed modifications may have been made to the exemplar vehicle before testing. Assuming that these allegations can be substantiated with competent evidence, any such modifications may preclude a finding of substantial similarity should they impact the rollover propensity of the exemplar. These, however, are questions for the district court to consider on remand.

Because we find that plaintiffs cleared the “substantial similarity” hurdle and because this evidence is admittedly crucial to their case, the exclusion of the testing evidence was reversible error. IV. For the foregoing reasons, we vacate the judgment below and remand for a new trial. Vacated and remanded. . Toyota also argues that 28 U.S.C. § 1292(

b) suggests that we should not consider plaintiffs’ appeal of the district court’s evidentiary ruling.

Section 1292 gives the federal courts of appeals jurisdiction over appeals from interlocutory orders and decrees entered by federal district courts. Nothing in this grant of jurisdiction can properly be said to conflict with the current case where plaintiffs appeal a properly entered judgment as a matter of law. . Plaintiffs argue that Toyota's motion for judgment as a matter of law was limited to the defect issue, and, therefore, it waived its argument with respect to the remaining elements of the prima facie case.

While it is true that if a party states one ground for granting judgment as a matter of law that party is later precluded from claiming that the motion should have been granted on another ground, see Hammond v. T.J. Litle & Co., 82 F.3d 1166, 1170-72 (1st Cir.1996), we may affirm a district court judgment on any "independently sufficient ground,” including one not raised below. Olsen v. Correiro, 189 F.3d 52, 58 (1st Cir.1999). Therefore, waiver is not applicable to the present case. . We say arguably because it is possible that the exclusion of Mr.

Anderson's testimony may have been fatal to the rest of plaintiffs’ case. Without being able to establish a defect, plaintiffs would have been limited in the evidence they could present on causation and damages, since those elements are related to the defect question. Neither party has fully briefed this issue, however. Therefore, it is unclear whether we could affirm the judgment as a matter of law even if we reverse on the evidentiary ruling. .

Relevant evidence is defined as any evidence having a "tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.” Fed.R.Evid. 401. . Neither party explicitly argues that Mr. Anderson’s testing of the exemplar vehicle was intended merely to illustrate scientific principles. However, plaintiffs repeatedly cite to cases in which evidence of accident reconstructions was admitted solely to illustrate scientific principles.