Kevin W. Ryan, argued, Michie & Hamlett, Charlottesville, VA, for appellants.

Wayne D. Struble, argued, Minneapolis, MN, for appellee.

Before LOKEN, Chief Judge, MELLOY, Circuit Judge, and SCHILTZ,1 District Judge.

SCHILTZ, District Judge.

Richard Olson ("Mr. Olson") spent the evening of September 17, 2002, at the Minot Country Club in Minot, North Dakota. Over the course of the evening, Mr. Olson played golf and drank a number of alcoholic beverages. Mr. Olson left the club at about 11:00 p.m. On the drive home, Mr. Olson lost control of his vehicle, a 1998 Ford Explorer, as he was attempting to navigate a curve. His vehicle slid off the road and crashed into a tree. Mr. Olson suffered extensive injuries and died at the scene.

Appellant Diana Olson ("Ms. Olson"), Mr. Olson's wife, brought a product-liability action against appellee Ford Motor Company ("Ford"), contending that Mr. Olson lost control of his Explorer because its cruise control actuator cable was defectively designed. Ms. Olson contended that this design defect caused a sudden, unexpected acceleration that Mr. Olson was unable to overcome even with hard braking. The case was tried to a jury, which returned a verdict finding that Ford and Mr. Olson were each 50% at fault. Under North Dakota law, this finding precluded Ms. Olson from recovering any damages. See N.D. Cent.Code § 32-03.2-02 (2006).

Ms. Olson brings this appeal, contending that the district court2 erred in ruling on the admissibility of evidence relating to Mr. Olson's ability to use the brakes effectively and his consumption of alcohol. Ms. Olson argues that, but for these errors, the jury would not have found that Mr. Olson's fault equaled Ford's. We affirm.

As noted, Ms. Olson argued at trial that her husband lost control of his Explorer because a defectively designed cruise control actuator cable caused the vehicle to accelerate unexpectedly. Ms. Olson further argued that her husband tried to slow his Explorer by putting as much force as he could on the brake pedal, but that he failed to bring the vehicle under control because of a particular feature of the power-braking system. According to Ms. Olson, when an Explorer is accelerating, the power brakes are much less effective and require much more force to overcome the acceleration, especially if the driver presses the brake pedal more than once (such as by pumping the brakes). The reason for this difficulty is that stepping on the brakes depletes the braking system's vacuum booster. When the vehicle is accelerating, the power braking system does not create another vacuum as efficiently as it does when the vehicle is not accelerating. With less help from the vacuum booster, the driver must rely mainly on manual force to brake the vehicle.

In support of her argument, Ms. Olson relied on two items of physical evidence. First, Ms. Olson pointed to the driver's seat in Mr. Olson's Explorer, which was bent backwards during the accident. Ms. Olson presented evidence that Mr. Olson himself bent the seat backwards while applying tremendous force to the brake pedal in a futile effort to stop his vehicle. Ford countered with evidence that the seat was bent as a result of the impact with the tree.

Second, Ms. Olson contended that the brake-pedal assembly was bent and the rubber brake-pedal pad was distorted. She argued that this was consistent with her theory that Mr. Olson put great force on the brake pedal. Ford argued in response that the brake-pedal assembly was not actually bent. Ford conceded that the rubber brake-pedal assembly was distorted, but introduced evidence that the distortion could not have been caused by Mr. Olson's foot.

In further support of her argument regarding the power-braking system, Ms. Olson presented the testimony of Dr. Rudy Limpert, a mechanical engineer with special expertise in brake design. Dr. Limpert personally tested the braking system of an exemplar Explorer while it was accelerating. Dr. Limpert testified that, at 40 miles per hour and after tapping the brake pedal several times, he was unable to slow the vehicle, no matter how much force he put on the pedal. Trial Tr. 468-69. Dr. Limpert further testified that, even at 32 miles per hour and without first tapping the brake pedal, he had to pull on the steering wheel with both hands in order to exert enough force on the brake pedal to slow the Explorer. Trial Tr. 466-67. Finally, Dr. Limpert testified that the curve on which Mr. Olson was driving when his vehicle left the road could not be navigated at 40 miles per hour if Mr. Olson was also attempting to brake against an open throttle.

In yet further support of her argument regarding the power-braking system, Ms. Olson sought to introduce the testimony of Casey Mulder and Wendy Crowell. Mulder, a former Ford test engineer, would have testified about the difficulty he experienced in braking when an Explorer that he was driving unexpectedly accelerated. Mulder was ultimately able to bring his Explorer under control, but he would have testified that, in general, when an Explorer unexpectedly accelerates, the driver will get a power assist only on the first press of the brake pedal. Any subsequent attempt to apply the brakes, Mulder would have said, will result in substantially less braking power, and the driver will have extreme difficulty in overcoming the acceleration. Mulder would have described the difficulty that he had in bringing his Explorer under control, even though he was a professional test driver.

Crowell was not a professional test driver, but simply an "average" Explorer owner, like Mr. Olson. Crowell would have told the jury that, several years ago, she was driving her 1995 Explorer out of a parking lot when the vehicle accelerated uncontrollably. Crowell stepped on the brake pedal with both feet and applied as much force as she could, but she was unable to slow the vehicle. Fortunately, she was able to steer the Explorer into an empty parking lot, shift the vehicle into neutral, and turn off the ignition.

The district court initially excluded the testimony of both Mulder and Crowell under Federal Rule of Evidence 403. When Ms. Olson again attempted to offer the two witnesses' testimony — this time in rebuttal — the district court again excluded it under Rule 403. (The district court also excluded Mulder as an inappropriate rebuttal witness.) Ms. Olson argues that the district court erred.

Under Rule 403, relevant evidence may be excluded "if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence." Fed.R.Evid. 403. We review a district court's exclusion of evidence under Rule 403 for a clear abuse of discretion. Stafne v. Unicare Homes, 266 F.3d 771, 776 (8th Cir.2001). The reason for this extremely deferential standard of review is obvious: A Rule 403 ruling — as much as any type of determination made by a district court — depends on factors that are uniquely accessible to the trial judge who is present in the courtroom and uniquely inaccessible to an appellate judge who must take the case on a cold record. This case illustrates the point.

No one can doubt that the testimony of Mulder and Crowell had probative value. If the testimony did not have probative value, it would not have been relevant, it would have been excluded under Rule 402, and a Rule 403 balancing would not have been necessary. What the district court had to determine, though, is how much probative value Mulder's and Crowell's testimony had. The amount of probative value of a piece of evidence is not static. Rather, it varies depending on such factors as the importance and complexity of the issue to which the testimony relates, the degree to which that issue is being contested, the other evidence that has already been introduced on the same issue, the attentiveness of the jury when that evidence was being introduced, and myriad other factors that can be fully assessed only by a judge who is present in the courtroom.

After the district court assessed the amount of probative value of the Mulder and Crowell testimony, it had to decide whether that probative value was substantially outweighed by six factors: (1) the danger of creating unfair prejudice; (2) the danger of confusing the issues; (3) the danger of misleading the jury; (4) the need to avoid undue delay; (5) the need to avoid wasting time; and (6) the need to avoid the presentation of cumulative evidence. Even more than assessing the amount of probative value, assessing these factors depends on what a trial judge sees and hears in the courtroom. For example, assessing the risk of confusing or misleading a jury depends to a substantial degree on the intelligence and attentiveness of the particular jurors — factors that are almost invisible to an appellate judge, but that can be appraised by the trial judge who questioned those jurors during voir dire (or who watched them being questioned by counsel) and who observed those jurors over the course of days or weeks of trial.

Here, the district court cited a couple of considerations that, in its view, substantially outweighed the probative value of Mulder's and Crowell's testimony. First, the district court expressed concern about turning the Olson trial "into a whole host of 8, 10, 12, 15 mini-trials on what went on in a 1996 Ford Explorer on the East Coast and the West Coast several years ago." Pretrial Hr'g Tr. 6. In other words, the district court recognized that, if Ms. Olson was permitted to introduce evidence that the Explorers driven by Mulder and Crowell accelerated unexpectedly and could not be brought under control by braking, Ford would have to be permitted to introduce evidence that those Explorers did not accelerate unexpectedly, or that they accelerated for a reason unrelated to the Olson accident, or that they could have been brought under control by braking. No judge wants to see one trial turn into several, especially when the one trial presents complex issues with which the jury may already be struggling.

Second, although Mulder's and Crowell's personal experiences were relevant and could have had an impact on the jury (especially as Mulder was a former Ford employee), the testimony of Mulder and Crowell was unquestionably cumulative. Dr. Limpert had already testified about his difficulty in bringing an accelerating Explorer under control by braking, and he had already explained why the power-braking system was ineffective. Ms. Olson had also introduced physical evidence that Mr. Olson had not been able to slow his Explorer despite putting tremendous force on the brake pedal. If Dr. Limpert's testimony was clear, and if the jury appeared to pay attention to the testimony and understand it, then the probative value of Mulder's and Crowell's testimony would have been substantially reduced. Of course, the judge best able to assess these considerations was the judge who sat a few feet from the jury while Dr. Limpert was testifying.

On this record, we are not persuaded that the district court clearly abused its discretion in finding that the probative value of Crowell's and Mulder's testimony was substantially outweighed by the danger of confusing the issues, the need to avoid undue delay, and the need to avoid the presentation of cumulative evidence. See Hicks v. Six Flags Over Mid-America, 821 F.2d 1311, 1316 (8th Cir.1987) (weighing the dangers of unfairness, confusion, and undue expenditure of time against the probative value of other similar incidents lies within the sound discretion of the district court). Similarly, we do not believe that the district court abused its discretion in excluding Mulder's testimony as improper rebuttal. See Gossett v. Weyerhaeuser Co., 856 F.2d 1154, 1156 (8th Cir.1988) (district court's determination of the admissibility of rebuttal evidence is reviewed for a clear abuse of discretion); Sterkel v. Fruehauf Corp., 975 F.2d 528, 532 n. 3 (8th Cir.1992) (the exclusion of cumulative rebuttal evidence does not prejudice a party's case); Skogen v. Dow Chem. Co., 375 F.2d 692, 706 (8th Cir.1967) (no abuse of discretion to exclude cumulative rebuttal evidence).

The day after the accident, the county coroner attempted to draw a blood sample from Mr. Olson's body but was unable to do so. The coroner instead drew a sample of vitreous humor, which is the clear fluid inside the eyeball. Vitreous humor is almost pure water, but it also contains hyaluronic acid, collagen, and trace amounts of other substances, such as sugar and vitamin C. When a person has been drinking, his vitreous humor will also contain alcohol. In general, a sample of vitreous humor from a person who has imbibed will contain more alcohol than a sample of that person's blood. The ratio will not remain constant, however, as the human body eliminates alcohol more quickly from blood than from vitreous humor.

The coroner drew vitreous humor from both of Mr. Olson's eyes and combined the samples in a tube with a screw-top lid. The tube and the equipment used to collect the sample had been provided by the North Dakota state crime lab, and the collection process was conducted according to standard procedures. The vitreous-humor sample was mailed to the state crime lab on the day that the sample was taken. On September 19, a crime-lab scientist ran two tests on the sample, both of which yielded a result of .22 percent alcohol by weight.

Several months later, the crime lab began retesting the vitreous-humor sample at Ms. Olson's request. Four additional tests were conducted — two in March 2003 and two in April 2003. The crime lab reported the following results for each test (including the two initial tests in September 2002):

  Date of Test      Result

  09/19/02          0.22 (first test)
  09/19/02          0.22 (second test)
  03/14/03          0.128
  03/20/03          0.135
  04/02/03          0.125
  04/16/03          0.095

In sum, although the two September tests yielded identical results, those results were not consistent with the results of the March and April tests, and the results of the March and April tests varied — slightly in the case of the first three tests, and somewhat more significantly in the case of the last test.

The North Dakota state toxicologist testified that the test results may have been inconsistent due to evaporation of alcohol over time and the fact that, because vitreous humor varies in consistency, the small samples that are tested may contain different amounts of alcohol. She agreed, however, that everything else being equal, the first two test results would have been the most accurate.

Ford's expert, Dr. Alan Donelson, applied statistical equations to the results of the first September 2002 test in order to calculate Mr. Olson's blood-alcohol level at the time of his death. Dr. Donelson, who is trained as a pharmacologist, learned of these equations by reviewing scientific literature describing studies of the relationship between vitreous-humor-alcohol levels and blood-alcohol levels. He did not claim to be able to pinpoint Mr. Olson's exact blood-alcohol level on the night of the accident. Instead, Dr. Donelson testified that the equations he used yielded a range in which Mr. Olson's blood-alcohol level most likely fell. Based on his analysis, Dr. Donelson testified that Mr. Olson's blood-alcohol level likely exceeded .10 percent at the time he was killed, and that the impairing effects of alcohol contributed to the crash. Trial Tr. 913-14.

Ms. Olson argues that it was prejudicial error for the district court to permit Dr. Donelson to testify concerning Mr. Olson's blood-alcohol level.3

Federal Rule of Evidence 702, which governs the admissibility of expert testimony, provides: If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise, if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case.

Fed.R.Evid. 702. District courts have wide latitude in determining whether an expert's testimony is reliable. Fireman's Fund Ins. Co. v. Canon U.S.A., Inc., 394 F.3d 1054, 1057 (8th Cir.2005). We review a district court's decision to admit expert testimony under the deferential abuse-of-discretion standard. United States v. Dico, Inc., 266 F.3d 864, 869 (8th Cir. 2001).

Ms. Olson makes several objections to the admissibility of Dr. Donelson's testimony. Before addressing those objections individually, we emphasize a general point that has been overlooked by Ms. Olson: Neither Rule 702 nor Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993) (the seminal case interpreting Rule 702) permits a district court to invade the province of the jury. Rule 702 does not permit a judge to weigh conflicting expert testimony, admit the testimony that he or she personally believes, and exclude the testimony that he or she does not personally believe. Nor does Rule 702 permit a judge to exclude expert testimony just because it seems doubtful or tenuous. The Supreme Court has been clear about how infirmities in expert testimony should be exposed: "Vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are the traditional and appropriate means of attacking shaky but admissible evidence." Daubert, 509 U.S. at 596, 113 S. Ct. 2786.

All of these correctives were brought to bear on Dr. Donelson's testimony. Ms. Olson's counsel vigorously cross-examined Dr. Donelson about his lack of experience in working with vitreous-humor samples and in converting vitreous-humor-alcohol levels to blood-alcohol levels. Ms. Olson also presented testimony about the inherent unreliability of vitreous-humor samples and cast doubt on the reliability of the particular sample drawn from Mr. Olson. Finally, Ms. Olson's experts testified that the level of alcohol in vitreous humor is not a reliable indicator of a person's blood-alcohol level.

At the end of the day, the jury appears to have believed Dr. Donelson's testimony, notwithstanding Ms. Olson's attacks. This does not suggest that something went "wrong" or that the district judge should have excluded the testimony. Rather, it suggests that the adversary system worked exactly as it was supposed to. The jury weighed contradictory evidence and decided which evidence to credit.

Turning now to Ms. Olson's specific complaints about the district court's decision to admit Dr. Donelson's testimony:

Ms. Olson first argues that the district court should have excluded Dr. Donelson's testimony because, before preparing for the trial in this case, he had never previously converted a vitreous-humor-alcohol level to a blood-alcohol level and he had no experience in drawing or handling vitreous humor. Ms. Olson did not make this objection to the district court, and therefore we review it for plain error. See McKnight v. Johnson Controls, Inc., 36 F.3d 1396, 1406-07 (8th Cir.1994) (plain error standard applies where opposing party failed to object that the expert was not qualified); United States v. Martin, 869 F.2d 1118, 1121 (8th Cir.1989) (failure to object to expert's qualifications does not preserve an issue for appeal, despite objections to some of the content of the expert's testimony).

"Plain error review is `narrow and confined to the exceptional case where error has seriously affected the fairness, integrity, or public reputation of the judicial proceedings.'" Chem-Trend, Inc. v. Newport Indus., Inc., 279 F.3d 625, 629 (8th Cir. 2002) (quoting Bd. of Water Works Trs. v. Alvord, Burdick & Howson, 706 F.2d 820, 824 (8th Cir.1983)). Under this standard, a verdict should be reversed only if the error has prejudiced the substantial rights of a party and would result in a miscarriage of justice if left uncorrected. Id. This standard is particularly stringent in the civil context. See Niemiec v. Union Pac. R.R., 449 F.3d 854, 857-58 (8th Cir. 2006).

Ms. Olson has failed to meet this standard. Dr. Donelson is a pharmacologist qualified by education and experience to testify about how alcohol and drugs affect human beings. Although Dr. Donelson had not previously calculated a blood-alcohol level from a vitreous-humor-alcohol level, the actual conversion was a matter of simple arithmetic. Ms. Olson argues that this conversion method is not scientifically reliable — an argument that we address below — but she does not contend that Dr. Donelson applied it incorrectly due to a lack of education or experience. There was no plain error in permitting Dr. Donelson to testify as an expert.

Ms. Olson next argues that Dr. Donelson's testimony should have been excluded because the sample of vitreous humor from which he drew his conclusions was unreliable. We review this issue for abuse of discretion.4

Ms. Olson contends that the state toxicologist testified that the vitreous-humor sample was unreliable. The record does not support this categorical assertion. The toxicologist did testify that the differing test results from the sample caused her some concern, but she also agreed that the first test — the test on which Dr. Donelson relied — was the test most likely to be accurate. She further testified that she was unaware of anything out of the ordinary in how the sample was collected, handled, and stored. Finally, she acknowledged that the first two tests on the sample — the tests conducted in September 2002 — yielded precisely the same result. Only the tests conducted after the sample sat in a tube for six months produced somewhat inconsistent results.

"`As a general rule, the factual basis of an expert opinion goes to the credibility of the testimony, not the admissibility, and it is up to the opposing party to examine the factual basis for the opinion in cross-examination.'" Minn. Supply Co. v. Raymond Corp., 472 F.3d 524, 544 (8th Cir.2006) (quoting Children's Broad. Corp. v. Walt Disney Co., 357 F.3d 860, 865 (8th Cir. 2004)). Ms. Olson fully explored this issue both in her direct examination of the toxicologist and in her cross-examination of Dr. Donelson. At that point, it was up to the jury to decide whether the sample was reliable. We find no abuse of discretion.

Ms. Olson next argues that Dr. Donelson's testimony should have been excluded because there is no reliable method for converting a vitreous-humor-alcohol level to a blood-alcohol level. To support her argument, Ms. Olson cites testimony from the state toxicologist and the county coroner indicating that, in their opinion, it is not appropriate to convert a vitreous-humor-alcohol level to a blood-alcohol level.

The fact that two witnesses did not regard the formulas used by Dr. Donelson as reliable does not mean that testimony based on the formulas was inadmissible under Rule 702. Ms. Olson's argument harkens back to the days of Frye v. United States, 293 F. 1013 (D.C.Cir.1923), when courts would admit expert testimony only when it was based on a scientific technique that was "generally accepted" by members of the relevant scientific community. As Daubert made clear, though, Frye was superseded by the 1975 adoption of the Federal Rules of Evidence. Daubert, 509 U.S. at 587-89, 113 S. Ct. 2786. The Supreme Court could not have been clearer in rejecting the Frye standard: "That austere standard, absent from, and incompatible with, the Federal Rules of Evidence, should not be applied in federal trials." Id. at 589, 113 S. Ct. 2786.

As amended in the wake of Daubert, Rule 702 instructs district courts to ask not whether scientific principles and methods are "generally accepted," but instead whether they are "reliable." In making this reliability determination, courts may consider a wide range of factors, including:

(1) whether the theory or technique can be (and has been) tested;

(2) whether the theory or technique has been subjected to peer review and publication; (3) whether the theory or technique has a known or potential error rate and standards controlling the technique's operation; and (4) whether the theory or technique is generally accepted in the scientific community.

Smith v. Cangieter, 462 F.3d 920, 923 (8th Cir.2006). Because this inquiry is necessarily fact-specific, there is no single standard for reliability. See Unrein v. Timesavers, Inc., 394 F.3d 1008, 1011 (8th Cir. 2005). Instead, these factors are flexible and should be adapted or rejected as the case demands. Id. Here, the district court explicitly considered all four of these factors and further noted that other courts have found that this type of evidence is reliable. See, e.g., Thier v. Lykes Bros., 900 F. Supp. 864, 869 (S.D.Tex.1995).

As noted, Ms. Olson relies on the fact that the state toxicologist and the county coroner did not personally regard the formulas used by Dr. Donelson as reliable. Ms. Olson also points out that the North Dakota legislature has not recognized the use of vitreous humor as a means to determine blood-alcohol levels. See N.D. Cent. Code § 39-20-07 (2006) (stating that evidence of the amount of alcohol in the blood as determined by chemical analysis of blood, breath, saliva, or urine is admissible).5

Ms. Olson's evidence may tend to show that the reliability of calculating a blood-alcohol level based on a vitreous-humor-alcohol level is not generally accepted in the scientific community. But her evidence is not determinative on that point. Ford introduced evidence to the contrary, and the district court was entitled to credit Ford's evidence over Ms. Olson's. More importantly, "general acceptance" is now just one of multiple factors that a district court must consider in deciding whether to admit expert evidence under Rule 702. Ms. Olson could win the battle over general acceptance and still lose the war over admissibility.

The district court did not abuse its discretion in finding that the calculation of Mr. Olson's blood-alcohol level from his vitreous-humor-alcohol level was based on sufficiently reliable principles and methods for purposes of Rule 702.

4. Failure to Take Eyewitness Testimony Into Account

Finally, Ms. Olson argues that Dr. Donelson failed to take eyewitness testimony into account in coming to his conclusion that Mr. Olson's driving was impaired due to his alcohol consumption. But expert testimony does not have to be consistent with all eyewitness testimony in order to be admissible. Because eyewitness testimony is often contradictory, expert testimony often cannot avoid conflicting with something that an eyewitness said. A conflict between the testimony of an eyewitness and the testimony of an expert witness does not mean that the latter was unreliable. After all, it could have been the eyewitness, and not the expert witness, who was wrong.

This case illustrates the point. It is true that no one who saw Mr. Olson on the night that he was killed testified that he appeared to be intoxicated. But Dr. Donelson testified that, in his experience, eyewitness accounts are not very reliable in determining how much a person has had to drink or whether that person is impaired. These observations merely reflect common knowledge; public-service announcements remind us daily that a person can be impaired by alcohol and yet show no outward sign of that impairment.6 Moreover, Ms. Olson had free rein to cross-examine Dr. Donelson about the alleged inconsistencies between his testimony and the testimony of eyewitnesses. Cf. United States v. Gipson, 383 F.3d 689, 696-97 (8th Cir.2004) (challenges to the application of a scientific methodology generally go to the weight, and not the admissibility, of the evidence). We find no abuse of discretion.7

The judgment of the district court is AFFIRMED.