EDMUNDS, District Judge.

Defendant, Norfolk & Western Railway Co. (“Norfolk”), appeals from the judgment entered against it after a jury verdict in an action for negligence brought pursuant to the Federal Employers’ Liability Act (“FELA”), 45 U.S.C. § 51, et seq. Plaintiff, Claude Smelser, brought suit against his employer, Norfolk, for injuries sustained when the company pick-up truck he was driving to a job site was hit in the rear by another motorist. Smelser alleged that the company truck had a defective safety belt which caused the injuries to his back and aggravated his neck injuries. The jury awarded Smelser $3.5 million compensatory damages, but the trial court remitted the judgment to $1,668,000. Norfolk appeals, arguing it is entitled to judgment as a matter of law in its favor, or in the alternative, a new trial. Smelser cross-appeals asserting that the trial court’s remit-titur decision was in error and should be reversed.

We conclude that the trial court erred when it allowed Smelser’s expert, Ronald Huston, to offer his opinion that a defective shoulder belt in the company pick-up truck, and not the rear-end collision, caused Smel-ser’s back injuries and aggravated his neck injuries. The court did not adequately assess the reliability of the methodology underlying Dr. Huston’s opinions both as to defect and causation and also failed to recognize that Dr. Huston’s opinion as to the cause of Smelser’s specific injuries went beyond his expertise in biomechanics. Accordingly, it failed to adequately perform its gatekeeping functions as recently defined by the Supreme Court in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). The court further erred when it denied Norfolk’s motion for a directed verdict because, without Dr. Huston’s improper testimony, there is not a scintilla of evidence that a defective shoulder belt, and not the rear-end collision, caused or aggravated Smelser’s injuries. Therefore, we REVERSE the judgment against Norfolk. The case is REMANDED for an entry of judgment as a matter of law in favor of Norfolk. Fed.R.Civ.P. 50(a).

Smelser, a 53-year old Norfolk employee, was injured on August 29, 1989 when the company pick-up truck he was driving- to a job site was rear-ended by another motorist traveling at approximately 45 m.p.h. Smel-ser testified that on impact, his head snapped back and hit the rear window of the truck cab, then his upper body moved forward and to the left; he ended up down in between the steering wheel and the door, and his left knee hit the floorboard.

Smelser subsequently sued Norfolk under FELA, alleging that Norfolk negligently failed to maintain the shoulder belt of the company truck he was driving and that this negligence caused his back injuries and aggravated his neck injuries which were initially caused by the rear-end collision. 1 Smelser claimed he was permanently disabled and *302 could no longer perform his job; he sought damages for lost wages and benefits, medical expenses, pain and suffering, and the lost capacity to enjoy the pleasures of life. Smel-ser also filed suit against the driver of the other vehicle involved in the August 1989 accident, seeking damages for the same injuries alleged in this suit. That suit was filed in the Ohio Court of Claims because the driver was an agent of the State of Ohio.

The pick-up truck Smelser was driving on August 29, 1989 was involved in an accident two months earlier in June of 1989. Smel-ser, the driver, testified that in the June accident his safety belt held him in place and left some marks on his neck and shoulder. He reported the accident to his supervisor and showed the marks to another supervisor. Although Smelser complained of soreness to his neck, he did not miss any work and did not file a claim for his injuries. He drove the truck for a while after the June 1989 accident, but the damage to the truck was repaired before the August 29, 1996 accident. The seat belt was not repaired or replaced. The owner’s manual provides that the seat belt should be replaced if damaged or involved in “all but a minor collision.”

Smelser’s FELA action went to trial, and, after the trial court denied Norfolk’s motion in limine challenging the admissibility of Dr. Huston’s expert testimony, and its motions for judgment as a matter of law at the close of Smelser’s case and again at the close of .the entire case, the matter was submitted to the jury. The jury returned a verdict in favor of Smelser in the amount of $8.5 million. Norfolk then filed a motion for judgment as a matter or law and/or a motion for a new trial pursuant to Fed.R.Civ.P. 50 and 59. It also filed a motion for an order of remittitur. The trial court denied Norfolk’s motion for judgment as a matter of law and its motion for a new trial, but granted its motion for an order of remittitur. Judgment was entered on March 31, 1995 reducing the verdict to $1,668,000.

II.

A. Admissibility of Expert Testimony Under Fed.R.Evid. 702

At trial, Smelser called Dr. Ronald Huston, a biomechanical engineer, to testify as an expert, and he rendered an opinion that (1) the shoulder belt, but not the lap belt, in the company truck was defective, (2) these circumstances worked together to cause Smelser’s body to jackknife at the waist, and (3) the defective shoulder belt, not the rear-end collision, caused Smelser’s back injuries and aggravated the neck injuries which were initially caused by the rear-end collision. Norfolk does not object to Dr. Huston’s qualifications to testify as a biomechanical engineer. Rather, Norfolk argues that Huston’s opinion testimony should have been excluded because the methodology underlying his opinions on defect and causation is unreliable, and his opinion on causation goes beyond his expertise in biomechanics. We agree.

“The Federal Rules of Evidence require a judge to undertake ‘a preliminary assessment of whether the reasoning or methodology underlying the testimony is scientifically valid and of whether that reasoning or methodology properly can be applied to the facts in issue.’ ” Frymire-Brinati v. KPMG Peat Marwick, 2 F.3d 183, 186-87 (7th Cir.1993) (quoting Daubert, 509 U.S. at 592-593, 113 S.Ct. at 2796-97). Had the trial court adequately performed its gatekeeping functions as set forth in Daubert, Huston’s expert testimony would not have been admitted into evidence.

1. Standard of Review

In Cook v. American Steamship Co., 53 F.3d 733, 738 (6th Cir.1995), this court set forth the standard of review to be used when examining a trial court’s rulings on the admissibility of expert testimony under Fed.R.Evid. 702. Preliminary factfinding under Fed.R.Evid. 104(a), e.g., a finding that the witness is qualified to testify as an expert on a certain subject, is reviewed for clear error. Id. We review de novo the trial court’s determination “whether the opinion the expert wishes to offer is properly the subject of ‘scientific, technical, or other specialized knowledge’” and this includes its “determination whether ‘the reasoning or methodology underlying the testimony is sci- *303 entificaUy valid.’ ” Id. (quoting Daubert, 509 U.S. at 592-93, 113 S.Ct. at 2796). 2 The party seeking to have the testimony admitted bears the burden of showing “that the expert’s findings are based on sound science, and this will require some objective, independent validation of the expert’s methodology”; “the expert’s bald assurance of validity is not enough.” Daubert v. Merrell Dow Pharmaceuticals, Inc., (on Remand), 43 F.3d 1311, 1316 (9th Cir.1995), cert. denied, — U.S. -, 116 S.Ct. 189, 133 L.Ed.2d 126 (1996).

2. The Trial Court’s “Gatekeeping” Function

The trial court is required under Fed. R.Evid. 104(a) to answer preliminary questions regarding the admissibility of expert testimony under Fed.R.Evid. 702. The Supreme Court in Daubert clarified that a trial court, when performing this gatekeeping function, must use a two-step inquiry which examines the expert’s opinion testimony for reliability and relevance. Cook, 53 F.3d at 737-38. First, the court is to determine “whether the experts’ testimony reflects ‘scientific knowledge,’ whether their findings are ‘derived by the scientific method,’ and whether their work product amounts to ‘good science.’” Daubert (on Remand), 43 F.3d at 1315 (quoting Daubert, 509 U.S. at 590, 593, 113 S.Ct. at 2795, 2797). An expert opinion that is based on scientifically valid principles will satisfy Fed.R.Evid. 702; an expert’s subjective belief or unsupported speculation will not. Daubert (on Remand), 43 F.3d at 1316.

Second, the court “must ensure that the proposed expert testimony is relevant to the task at hand.” Id. at 1315. The Supreme Court referred to this as the “fit” requirement. Id. When making a preliminary finding regarding an expert’s qualifications under Fed.R.Evid. 104(a), the court is to examine “not the qualifications of a witness in the abstract, but whether those qualifications provide a foundation for a witness to answer a specific question.” Berry v. City of Detroit, 25 F.3d 1342, 1351 (6th Cir.1994), cert. denied, — U.S. -, 115 S.Ct. 902, 130 L.Ed.2d 786 (1995). The trial court must determine whether the expert’s training and qualifications relate to the subject matter of his proposed testimony.

When considering reliability, the trial court must focus on the soundness of' the expert’s methodology and not the correctness of his conclusions. Daubert (on Remand), 43 F.3d at 1318. The Supreme Court provided the following non-exclusive list of factors to assist the trial court in its inquiry:

(1) whether a theory or technique can be (and has been) tested, (2) whether the theory or technique has been subjected to peer review and publication, (3) the known or potential rate of error in using a particular scientific technique and the existence and maintenance of standards controlling the technique’s operation, and (4) whether the theory or technique has been generally accepted in the particular scientific field.

Glaser v. Thompson Medical Co., Inc., 32 F.3d 969, 972 (6th Cir.1994) (citations omitted). These factors are to assist the court in determining “whether the analysis under-girding the experts’ testimony falls within the range of accepted standards governing how scientists conduct their research and reach their conclusions.” Daubert (on Remand), 43 F.3d at 1316. The Ninth Circuit has added another factor to assist the court in its inquiry: “whether the experts are proposing to testify about matters growing naturally and directly out of research they have conducted independent of the litigation, or whether they have developed their opinions expressly for purposes of testifying” because the former “provides important, objective proof that the research comports with the dictates of good science.” Id. at 1317.

The trial court considered none of these factors. We conclude that Huston’s opinion testimony cannot withstand Daubert scrutiny, and should not have been admitted into evidence.

a. Huston’s Opinion Testimony on Seat Belt. Defect

Dr. Huston testified that the first time he examined the truck’s safety belt was *304 on March 13, 1992, more than two and one-half, years after the August 29,1989 accident. Initially, he conducted a visual examination of the seat belt webbing, determined that there were stretch marks on both the .lap and shoulder belt, and then measured and photographed the belts and marks. He did not examine the belts under a microscope and did not perform other tests to determine the cause or effect of the marks detected on both, i.e., whether the marks were from normal wear and tear or from being “loaded” and the consequences of that determination. He also did no testing to determine if the marks on the belts were from the June 1989 or August 1989 accidents. Nonetheless, he opined that the marks on the shoulder belt were caused by “loading” in the June 1989 accident. He rendered no opinion about the marks he observed on the lap belt. He also admitted that the presence of “load” marks on a seat belt did not permit the conclusion that its retractor was damaged and would not work. .

On the same day, following this visual inspection, Huston took the pick-up truck on a drive through suburban Batavia, Ohio and tested the seat belt. With his left hand on the steering wheel and his right hand holding the shoulder harness, Huston hit the brakes about three times while driving at 25 to 35 miles an hour and then concluded that the retractor in the shoulder belt did not catch as it should. He testified that it would be difficult to conclude much about the lap belt in that test drive because he was anticipating hitting the brakes, pushing his back against the seat, and resisting the forward motion. Nonetheless, he concluded the lap belt worked properly and the shoulder belt did not. Huston did not document the testing circumstances or results.

Huston did not conduct any further tests that day, but 20 months later, in November of 1993, he tested the retractor assembly of the shoulder belt outside the company truck. To determine whether the locking mechanism was functioning properly, Huston held the assembly in his hand, tilted it in certain directions, pulled out the webbing, and measured the angle at which it locked, if at all, with a protractor. During this test, he assumed, despite his admission that the mounting position varies from vehicle to vehicle, that the mounting plate for the seat belt in the company vehicle was at zero degrees, and then tested it against a Federal Motor Vehicle Safety Standard (“FMVSS”) which requires the mechanism to lock at 35 degrees. Huston admitted that (1) his test results would be inaccurate if the mounting plate was not at zero degrees in the company truck, (2) he never inspected the vehicle or its factory specifications to ascertain the precise degree of the mounting plate, and (3) although he performed this test “about a hundred times,” he took measurements only five or six times. Nonetheless, he concluded that the shoulder belt did not lock at the angle it should.

Based upon his test results, Huston rendered the opinion that the shoulder belt was defective but the lap belt was not. He further opined that, due to these factors, Smel-ser’s body jackknifed or bent around the belt upon impact, causing the injuries to his lower spine and aggravating his neck injuries which, unlike the back injuries, were initially caused by the rear-end collision.

Daubert teaches that expert opinion testimony qualifies as scientific knowledge under Rule 702 only if it is derived by the scientific method and is capable of validation. Huston’s opinion that the shoulder belt, but not the lap belt, failed in the August 29, 1989 accident cannot be based on “good science” when he (1) failed to perform any tests on the lap belt yet concluded it was in proper working condition; (2) conducted no testing to verify his conclusion the shoulder belt was damaged in the June 1989 accident; (3) failed to adequately document testing conditions and the rate of error so the test could be repeated and its results verified and critiqued; and (4) failed to discover, use or at least consider the degree the restraint system was actually mounted at in the subject vehicle and explain whether that information would affect his pendulum test for compliance-with the federal safety standard. Smel-ser failed to establish that any of Huston’s seat belt tests were based on scientifically valid principles, were repeatable,, had been the subject of peer review or publication or *305 were generally accepted methods for testing seat belts in the field of biomechanics. Accordingly, Huston’s opinion testimony that the pick-up truck’s shoulder belt, but not the lap belt, was defective should have been excluded.

b. Huston’s Causation Opinion

Huston’s opinion testimony about causation also should have been excluded. First, it goes beyond his expertise in biomechanics. Second, the opinion lacks reliability.

During preliminary questioning, Huston explained that biomechanics apply the principles in mechanics to the facts of a specific accident and provide information about the forces generated in that accident, explain how the body moves in response to those forces, and thus determine what types of injuries would result from the forces generated. Huston admitted that biomechanics are qualified to determine what injury causation forces are in general and can tell how a hypothetical person’s body will respond to those forces, but are not qualified to render medical opinions regarding the precise cause of a specific injury. He acknowledged that each individual person has his own tolerance level, and therefore, admitted he could testify only in general terms, i.e., that “X” forces would generally lead to “Y” injuries and “Y” injuries are consistent with those the plaintiff claims to have suffered.

Huston also admitted that (1) he was not a medical doctor, had no medical training, and must rely on a medical doctor’s opinion to determine a particular individual’s injuries; (2) each individual has his or her own tolerance level and pre-existing medical conditions could have an effect on what injuries result from an accident; and (3) he had not examined Smelser’s complete medical history. Nonetheless, the trial court permitted Huston to testify that:

the failure of the seat belt, the shoulder belt webbing to lock directly, led to the injuries. The neck injury, in my opinion, was caused by the rear-end collision, with the neck being thrust backward. And then the failure of the shoulder belt allowed the shoulder to go forward, aggravating that injury, causing it to go back further. And at the same time then the lumbar region came from what might be called a jackknifing or a bending around the belt, causing the injury to the lower spine.

This opinion testimony goes beyond Huston’s expertise in biomechanics. As he previously admitted, he was qualified to render an opinion that made úse of his discipline’s general principles, described the forces generated in the August 1989 rear-end collision, and spoke in general about the types of injuries those forces would generate. Huston is not a medical doctor who had reviewed Smelser’s complete medical history, and his expertise in biomechanics did not qualify him to testify about the cause of Smelser’s specific injuries. As this court observed in Berry, “[t]he issue with regard to expert testimony is not the qualifications of a witness in the abstract, but whether those qualifications provide a foundation for a witness to answer a specific question.” Berry, 25 F.3d at 1351.

Huston’s causation opinion has additional flaws. It is based upon the assumption that Smelser had no similar neck or back injuries prior to the August 29,1989 accident. Huston makes this assumption despite his admissions that (1) information about prior back or neck injuries is important and could affect his causation opinion; (2) he had reviewed only four x-ray reports, not Smelser’s complete medical history; (3) he did not consider the fact that Smelser complained of neck soreness after the June 1989 accident; and (4) he had not reviewed Smelser’s deposition or discussed with him his symptoms or his version of how his body moved during the accident. Cross-examination at trial produced uneontroverted evidence that Smelser had been treated and hospitalized for back injuries since as early as 1968 and 1971, and his x-rays taken after the accident showed signs of degenerative conditions that could not have been caused in the August 1989 accident. . Huston’s causation opinion failed to consider this admittedly important information,'and therefore cannot be considered rehable. Accordingly, it should not have been admitted.

*306 B. Directed Verdict

We review de novo the trial court’s decision to deny Norfolk’s motion for judgment as a matter of law brought pursuant to Fed.R.Civ.P. 50(a) at the close of Smelser’s case and apply the same standard the trial court should have applied when considering Norfolk’s motion. 3 Apando v. Norfolk & Western Ry. Co., 84 F.3d 803, 806 (6th Cir.1996). In the usual ease, “this standard requires the district court to decide whether a reasonable jury could conclude that the plaintiff has not proven an element of his or her case.” Id. at 807. The court must make this determination by looking only at the plaintiffs evidence and “[without weighing the evidence,or judging the credibility of witnesses, and drawing all inferences in favor of the nonmoving party.” Id. In FELA cases, however, a different standard applies. Id. The court still considers only the plaintiffs evidence, does not weigh the evidence or judge the credibility of witnesses, and draws all inferences in favor of the non-moving party, but then determines whether the plaintiff has presented more than a scintilla of evidence on each element of his or her FELA ease. 4 Id. at 810.

At issue here is Smelser’s presentation of evidence that Norfolk’s negligence played some part in causing his injuries. The only negligence attributable to Norfolk flows from Smelser’s claim that the shoulder belt in the company truck was defective. Huston is the only witness whose testimony connects Smelser’s injuries to the allegedly defective shoulder belt apart from the rear-end collision. Absent,this improper testimony, we find Smelser failed to meet his burden by presenting more than a scintilla of evidence that Norfolk’s negligence played any part in causing his injuries. Therefore, Norfolk’s motion for a directed verdict should have been granted pursuant to Fed.R.Civ.P. 50(a).

III.

The jury verdict against Norfolk is REVERSED, and the case is REMANDED for an entry of judgment as a matter of law in favor of Norfolk. Fed.R.Civ.P. 50(a).

. Smelser’s complaint names both Norfolk Southern Railway Company and Norfolk & Western Railway Co. as Defendants, characterizes them as “closely interwoven corporations” and alleges he is an employee of both. Both Defendants are collectively referenced here as "Norfolk.”

. In Cook, this Court also observed that the trial court's assessment of whether an expert opinion will assist the jury in understanding the evidence or in determining a fact in issue is a relevancy determination that is reviewed for an abuse of discretion. Cook, 53 F.3d at 738.

. Fed.R.Civ.P. 50(a) provides that:

If during a trial by jury a parly 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, the court may determine the issue against that party and may grant a motion for judgment as a matter of law.

. The elements of a FELA case are:

(1) an injury occurred while the plaintiff was working within the scope of his or her employment with the railroad, (2) the employment was in the furtherance of the railroad's interstate transportation business, (3) the employer railroad was negligent, and (4) the employer’s negligence played some part in causing the injury for which compensation is sought under the Act.

Aparicio, 84 F.3d at 810.