dissenting:
I would affirm the judgment of the district court. I wholeheartedly agree with the majority that Mr. Dixon was not con-tributorily negligent as a matter of North Carolina law, and that the Dixons' state law claims were properly submitted to the jury. In my view, however, the majority’s arguments go further than that. They establish not only that the Dixons’ state law claims were properly submitted to the jury, but also that they were properly resolved.
To illustrate, I can do no better than to quote from the majority opinion. “CSX does not dispute that Dixon could not have seen the approaching train from Highway 27; the train would only have become visible at some point along the fifty-eight foot stretch of Rhyne Road between the highway and the crossing, leaving Dixon a short time at best in which to react.” Majority Op. at 1451. “There was no evidence introduced that Dixon approached the crossing recklessly; in fact, the testimony suggested that he was ‘almost stopped’ as he proceeded slowly southward from Rhyne Road, and that he did look for oncoming trains.” Id. at 1451. “There was, moreover, considerable testimony that Dixon’s view of the tracks from in and around the crossing was obscured by, among other things, the three- to four-foot embankment covered with one- to two-foot tall weeds that ran along the northern edge of the CSX tracks to the east of the crossing.” Id. at 1451. “Finally, there was ample evidence from which a reasonable jury could have concluded that there was no point north of the crossing at which Mr. Dixon could have stopped his long-nosed Buick and looked and listened for an approaching train without putting himself in danger of being struck by a train, particularly given that railroad cars overhang the tracks by as much as two feet.” Id. at 1451.
There is ample reason therefore to credit the jury’s determination that Dixon was not contributorily negligent. Given this documented support in the majority opinion for the jury’s state law verdict, I am at a loss to understand why we should remand this case for another lengthy and exhaustive trial. The justification apparently lies in the prejudicial effect of a pair of trial errors alleged by the majority in its opinion. The prejudicial effect of these errors — if indeed they be errors — is not sufficient to cast any legitimate doubt on the validity of the jury’s state law verdict.
I.
The first of the alleged trial errors is the district court’s refusal to bifurcate the issue of coverage under the Federal Employers’ Liability Act from the issue of substantive liability. The majority appears to offer three rationales for why this refusal to bifurcate might warrant a new trial. Each in my view fails.
The first rationale is that the joint trial of the FELA and state claims allowed the *1455jury to hear “highly prejudicial” and “inflammatory” evidence about Mrs. Dixon’s loss-of-consortium claim — evidence that was only relevant to the Dixons’ state law claim, since FELA does not award damages for loss of consortium. Majority Op. at 1443. This argument cannot support a new trial. Since North Carolina law permits recovery for loss of consortium, any prejudice from the loss-of-consortium evidence as to the Dixons’ common law claims would have been, by definition, proper.
The second rationale is that the joint trial allowed the jury to hear testimony about whether Mr. Dixon was an employee of CSX, which was “wholly irrelevant to the Dixons’ state law claims and undoubtedly prejudicial to CSX.” Majority Op. at 1444. This argument is no more persuasive than the first. The testimony about whether Dixon was an employee of CSX was certainly irrelevant to the state law claims, but it was hardly inflammatory or unduly prejudicial. The majority alludes to the closing arguments in which the Dixons’ counsel accused CSX of being a ruthless employer which had bullied its employees, and implies that these statements somehow subsumed and tainted the testimony about whether Dixon was an employee of CSX. Majority Op. at 1444. This point stands only if the closing argument by itself was reversible error, a ruling the majority appears unwilling to make. Majority Op. at 1444 n. 4. In the context of such a long and drawn-out trial, I have difficulty believing that some closing remarks, to which CSX did not object at trial, could seriously have affected the jury’s verdict.
The final rationale is that the joint trial caused “considerable jury confusion,” because the jury had to resolve two alternative claims. Majority Op. at 1444. I find nothing to back this bald contention except conjecture. Civil plaintiffs assert alternative theories of recovery all the time. Next to the issues posed in a large antitrust or commercial contract dispute, the issues in this case pale in their complexity and scope: How fast did Mr. Dixon approach the crossing? Did he stop first to look for an oncoming train? How clear was his view down the tracks? Did he have enough time to spot the oncoming train and get out of the way? These are just the sorts of determinations for which juries are best suited. In this case, moreover, the district court took particular care to keep the FELA and state law standards distinct, by bifurcating the charges to the jury, and by allowing the litigants to make separate closing arguments about each claim. Unless we have lost faith in the capacity of juries to resolve civil disputes, the assertion that the joint trial somehow confused the jury must fail.*
II.
The second error that the majority alleges is the admission of testimony by the “human factors expert,” Dr. Harry Snyder. At trial, Dr. Snyder testified about how other motorists under his observation had approached the Rhyne Road crossing: 96% did not come to a full stop, and 64% failed to look to their left, the direction from which Mr. Dixon was struck. According to the majority, “Dr. Snyder’s survey results are wholly irrelevant to the question of whether Mr. Dixon fulfilled his burden to act as a reasonably prudent person, a burden that the North Carolina Supreme Court has defined with specificity....” Majority Op. at 1452. “The fact that 64% or 96% or even 100% of the motorists do not comply with the duties imposed upon them by law [does] not excuse them or Mr. Dixon from satisfying these obligations.” Id. In other *1456words, North Carolina defines the motorist’s duty of care in firm rules that have no reference to local custom. Hence evidence like Dr. Snyder’s survey is inadmissible.
This is a startling proclamation, given the long common law tradition to the contrary. “[EJvidence of the usual and customary conduct of others under similar circumstances is normally relevant and admissible, as an indication of what the community regards as proper, and a composite judgment as to the risks of the situation and the precautions required to meet them.” W. Page Keeton et al., Prosser and Keeton on the Law of Torts § 33, at 193 (5th ed. 1984) (Lawyer’s ed.) (footnote omitted). Before departing from this practice here, I should expect to see some North Carolina authority stating that local custom is never a measure of due care at railroad crossings, or that the common law rules of how a motorist should approach a crossing do not take account of local circumstances. The majority provides no such authority; indeed, it ably demonstrates the opposite. “North Carolina law ... does not impose upon the driver of a motor vehicle, on his approach to a public crossing, the duty, under all circumstances, to stop his vehicle before driving on the crossing.” Majority Op. at 1449 (quoting White v. North Carolina R. Co., 216 N.C. 79, 3 S.E.2d 310, 315 (N.C.1939) (internal quotations omitted)). “No inflexible rule can be laid down as to what constitutes contributory negligence as a matter of law, as each case must be decided on its own facts.” Majority Op. at 1449 (quoting Ramey v. Southern Ry., 262 N.C. 230, 136 S.E.2d 638, 643 (N.C.1964)).
Thus it was within the district court’s discretion to conclude that Dr. Snyder’s survey was relevant to whether Dixon was contributorily negligent. Moreover, the district court instructed the jurors, when Dr. Snyder testified, that they were to determine the weight of his testimony. The danger that the jury attached undue significance to the evidence of local custom is therefore illusory.
III.
To conclude, the Dixons’ state law claims were properly resolved by the jury, and any errors committed by the district court were too slight to have “affect[edj the substantial rights of the parties.” Fed.R.Civ.P. 61 (standard for harmless error). To some, the verdict in this case may seem the product of jury sympathy, but I confess I am not certain how jurors can abandon all feeling for a plaintiff who has been rendered a quadriplegic for life. The question on appeal is not whether the jury was moved by sympathy, but whether it rendered a verdict with a solid foundation in the evidence. I am convinced that it did, for the reasons set forth so well in the majority opinion. My vote to affirm is an expression of regret that the majority has failed to follow its own survey of the case: “Certainly, viewing the evidence in the light most favorable to Dixon, a jury could reasonably have concluded that Dixon actually stopped, looked, and listened before entering the crossing.” Majority Op. at 1451.
The majority points to the identical damages awarded on Mr. Dixon’s FELA and state law claims as post hoc evidence of jury confusion. Since the method for calculating damages was different for each claim, the majority reasons that the jury must have been "unable to distinguish between the FELA and the state law claims, at least insofar as the measures of damages were concerned.” Majority Op. at 1444-45 (emphasis added). The inferential leap from confusion about damages to confusion about liability is not so obvious to me. As the majority seems to concede by its own qualifying phrase, the most this argument proves is the need to remand for retrial on damages alone. In no way does it justify the more drastic remedy of a whole new trial.