[¶ 1.] In this appeal, we hold that the circuit court erred in allowing an economist to assign the plaintiff a disability rating. We affirm the jury’s verdict on negligence, but reverse for a new trial on damages.
Background
[¶2.] On March 3, 1995, James Garland and Ralph Rossknecht collided at thé intersection of Haines Avenue and Crazy Horse Street in Rapid City, South Dakota. Haines runs north and south and Crazy Horse, east and west. The intersection is controlled by stop signs on Crazy Horse. Vehicles on Haines have the right-of way. Rossknecht testified that he was traveling west on Crazy Horse and approached the intersection intending to turn left onto Haines. He" looked first to the right and then to the left and “felt there was no approaching traffic.” He then proceeded into the intersection and while “double-checking” traffic to the north he first saw Garland’s vehicle “coming from' [the] right side.” Garland testified that he was southbound on Haines, traveling approximately thirty-two or thirty-three miles per hour. The speed limit on Haines is thirty-five. Garland was halfway through the intersection when he saw Rossknecht’s Jeep “coming at [him].” The collision occurred in Garland’s lane of traffic. The right corner of Rossknecht’s front bumper struck Garland’s left rear wheel. According to Garland, the collision caused his car to “scoot” or “fishtail” toward “the direction ... he hit me.”
[¶ 3.] Garland reported no medical problems at the time, but after he went home his neck felt “tingly and, um just weird.” When he awoke the next day, he had a “constant, nagging headache feeling in [his] neck.” He sought medical treatment from a chiropractor, Dr. Garry Gor-such, for problems with “cold feet, neck pain, fatigue, tension, mid-back stiffness and numbness in [his] fingers.” Before the accident, Garland had seen Dr. Gor-such s occasionally for wellness check-ups. After the accident, Garland underwent an extensive course of treatment with Dr. Gorsuch and eventually with Gorsuch’s partner, Dr. Scherr, for pain in his back and neck. The treatments lasted from March 1995 to October 1996.
[¶ 4.] Scherr concluded that Garland suffered from a seven percent whole person permanent impairment. This degree of physical impairment, according to Scherr, would restrict Garland’s ability to sit in one position for long periods and work or lift with his arms overhead. Schérr’s assessment was limited to physical impairment. He would not assign a vocational disability rating, as he felt that a vocational rehábilitationist should make that determination. In Scherr’s view, physical impairment ratings do not translate into disability ratings because “disability is based upbn what you cannot do in regards to a given employment.”
[¶ 5.] Garland brought a negligence action against Rossknecht, seeking damages for medical bills, pain and suffering, lost wages, and dimipished earning capacity. In addition to the expertise of his two treating chiropractors, Garland offered the testimony of Don Frankenfeld, a forensic economist. Frankenfeld assessed Garland’s earning capacity in his chosen field, graphic design.1 His opinions were challenged in a pretrial hearing and later ruled *702admissible over objections from the defense.
[¶ 6.] In the process of calculating reduced earning capacity, Frankenfeld concluded that Garland suffered from a ten percent “disability” or earning capacity reduction. When questioned about how he arrived at this percentage, Frankenfeld stated, “it was a judgment call.” In fact, his opinion was based solely on information conveyed from Garland. Garland reported that when he worked on a four-hour graphic design project, he had to take a half hour break. Frankenfeld extrapolated a half hour break every four hours into one hour of nonproductivity a day. He then took “one and divide[d] it by eight, one hour a day out of an eight-hour day is one-eighth, twelve and a half percent.” Then he discounted the rating to ten percent.
[¶ 7.] Frankenfeld admitted that impairment ratings do not translate directly into disability ratings and that he had no expertise in converting “impairment ratings into physical vocational limitations.” Furthermore, he conceded “a vocational expert has competence that I do not have to take physical limitations and translate them into vocational limitations.” After hearing this testimony, the circuit court nonetheless allowed Frankenfeld to testify without limitation. Frankenfeld offered the jury a range of figures representing the present value of any reduction in Garland’s earning capacity. Included in that range was a calculation based on a ten percent disability and an eight dollar per hour rate of earning, which produced a loss of $90,744. Frankenfeld also calculated damages based on higher and lower disability ratings.2
[¶ 8.] At the end of the two day trial, the jury returned a general verdict of $50,600 for Garland. Rossknecht appeals.
1. Disability Rating
[¶ 9.] Rossknecht urges us to hold that the circuit court erred in allowing an economist to assign a disability rating for Garland, especially as it was based only on Garland’s self-appraisal of his work capacities. Trial courts enjoy broad discretion in ruling on the admissibility of expert opinions. In re Estate of Dokken, 2000 SD 9, ¶ 11, 604 N.W.2d 487, 491 (citing State v. Edelman, 1999 SD 52, ¶ 4, 598 N.W.2d 419, 421) (further citations omitted). We will not reverse a court’s evidentiary rulings absent a clear abuse of discretion. Edelman, 1999 SD 52, ¶ 4, 593 N.W.2d at 421. See also Nickles v. Schild, 2000 SD 131, ¶ 7, 617 N.W.2d 659, 661 (citations omitted).
[¶ 10.] The circuit court found that Rossknecht’s foundational objections went to “credibility, not admissibility,” reasoning that the jury should decide whether the information on which Frankenfeld relied was credible. It is true that challenges to the facts underlying an expert’s opinion present issues for a jury. Stormo v. Strong, 469 N.W.2d 816, 821 (S.D.1991). However, when dealing with expert opinion, the court must fulfill a gatekeeping function, ensuring that the opinion meets the prerequisites of relevance and reliability before admission.3 Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 597, 113 S.Ct. 2786, 2799, 125 L.Ed.2d 469, 485 (1993); State v. Hofer, 512 N.W.2d 482, 484 (S.D.1994) (citations omitted); Rogen v. Monson, 2000 SD 51, ¶¶ 26 27, 609 N.W.2d 456, 462 (Konenkamp, J. concur*703ring specially) (Daubert applies to medical opinions).
[¶ 11.'] Daubert and its progeny qffer general guides for courts to consider in assessing reliability: testing, peer review, error rate, and general acceptance. See Daubert, 509 U.S. at 593-94, 113 S.Ct. at 2796-97, 125 L.Ed.2d at 482-84. These factors cannot be applied in all settings. In some instances, reliability must focus on “knowledge and experience.” Kumho Tire Co., Ltd. v. Carmichael, 526 U.S. 137, 150, 119 S.Ct. 1167, 1175, 143 L.Ed.2d 238, 251 (1999). A fundamental baseline for reliability is that experts are limited to offering opinions within their expertise. See Brain v. Mann, 129 Wis.2d 447, 385 N.W.2d 227, 230 (1986) (citations omitted).
[¶ 12.] Frankenfeld, a forensic economist, specializes in projecting earning loss-over time, not in assessing vocational disabilities. He conceded' as much in the pretrial hearing: “[a] vocational expert has competence that I do not have to take physical limitations and translate those into vocational limitations. That’s not an area that I get into, but it is possible.” According to his chiropractor, Garland suffers a seven percent physical impairment rating; however, a physical impairment does not directly translate into any vocational disability. Marnette v. Morgan, 485 N.W.2d 595, 598 (S.D.1992).4 Frankenfeld’s opinion did not hinge on a physical impairment rating. Instead, he relied solely on Garland’s assertion that he must take a thirty minute break every four hours. Thus, Frankenfeld, who admits that he has no competence as a vocational expert, converted one hour of claimed non-productivity a day into a ten percent disability. Frankenfeld also acknowledged that in most instances where he testifies concerning economic loss, he does so based on the analysis of a vocational expert. We conclude that his opinion cannot be considered reliable.
[¶ 13.] Other jurisdictions have addressed the proper scope of an economist’s testimony in claims for earning impairment. In Lamphere v. Agriew, 94 N.M. 146, 607 P.2d 1164, 1167 (App.1979), the court refused to allow the testimony of an economist for lack of proper foundation. There, the absence of medical testimony on the plaintiffs percentage of disability precluded the economist from translating evidence of specific physical disabilities into a “quantifiable future loss.” Id. Likewise, in Morgan v. Cashion, 638 S.W.2d 387, 391 (Tenn.Ct.App.1982), the court refused to admit an economist’s opinion purporting to convert an anatomical disability (physical impairment) into a work disability-
[¶ 14.] Garland refers to our statement in Mamette that disabilities may be proven with testimony from an expert “other than a doctor.” Marnette, 485 N.W.2d at 598; see also Caldwell v. John Morrell & Co., 489 N.W.2d 353, 363 n. 2 (SD 1992). This argument is unpersuasive for two reasons. First, the statement cannot be viewed as support for allowing experts to testify to matters wholly outside their expertise. Second, the statement merely reiterates the common proposition that in some instances injury will be obvious, thus expert testimony might be advisable but is not required. See Caldwell, 489 N.W.2d at 362. On the other hand, an expert is required when the subject matter falls outside the common experience of lay persons. Id. See also Mc*704Govern v. Murray Taxi Co., 75 S.D. 151, 60 N.W.2d 211, 213-14 (S.D.1953).
[¶ 15.] The circuit court abused its discretion in allowing the economist’s opinion on Garland’s percentage of disability.
2. Denial of Directed Verdict
[¶ 16.] Rossknecht contends that the trial court erred in denying his motion for directed verdict. He claims that Garland made an insufficient showing of negligence. “If any legally sufficient basis exists to support a verdict for the nonmoving party, the motion must be denied.” Jurrens v. Lorenz Mfg. Co., 1998 SD 49, ¶ 5, 578 N.W.2d 151, 153 (citations omitted). We view the evidence in a light most favorable to the nonmoving party. McDonough v. Kahle, 1999 SD 14, ¶ 8, 588 N.W.2d 600, 602 (citing Denke v. Mamola, 437 N.W.2d 205, 207 (S.D.1989)) (further citations omitted). A trial court’s ruling on a directed verdict motion is presumed correct, and we will not seek reasons to reverse it. United States and Hartzell Propellar, Inc. v. South Dakota, 1999 SD 94, ¶ 7, 598 N.W.2d 208, 211 (citations omitted). A directed verdict in a negligence action is ordinarily inadvisable because the “reasonableness” standard must be applied to conflicting testimony. Carpenter v. City of Belle Fourche, 2000 SD 55, ¶ 8, 609 N.W.2d 751, 757.
[¶ 17.] Rossknecht relies on the following language in Carpenter: “the mere occurrence of an accident in a protected intersection will not of itself establish negligence against the unfavored driver.” Id. at ¶ 16, 609 N.W.2d at 759. He insists that the collision itself is the only evidence of negligence; therefore, the circuit court should have directed a verdict in his favor. Rossknecht’s reliance on Carpenter is entirely misplaced. Our statement in Carpenter simply emphasized that questions of negligence or contributory negligence in intersectional accidents are generally factual questions for the jury. 2000 SD 55, ¶ 14, 609 N.W.2d at 759. More specifically, we noted that “[w]hether one looked and failed to see a vehicle within the zone of danger is a jury question, except in those rare instances when the evidence is so definite that reasonable minds could not differ.” Id. at ¶ 16.
[¶ 18.] Rossknecht invokes his own trial testimony as undisputed proof that he acted reasonably and without negligence. Of course, the jury had to listen to both sides and consider the testimony in context with the vehicle damage. The right corner of Rossknecht’s bumper collided with the left rear wheel of Garland’s vehicle. Rossknecht ran into Garland’s vehicle as it proceeded through the intersection. “When witnesses testify that they did in fact look, the question whether they looked with the proper care and saw all that they should have seen is one of fact.” Carpenter, 2000 SD 55, ¶26, 609 N.W.2d at 762. The trial court correctly denied a-directed verdict and submitted the case to the jury under proper instructions.
[¶ 19.] Affirmed in part, reversed in part, and remanded for a new trial on damages.
[¶ 20.] MILLER, Chief Justice, and AMUNDSON, and GILBERTSON, Justices, concur. [¶ 21.] SABERS, J., concurs in part and dissents in part.. Garland was in high school at the time of the accident. In the time between the accident and the trial, he graduated from high school and enrolled in a graphic design program at Brooks College of Design and Merchandising.
. Frankenfeld offered the jury a range of calculations based on varying disability ratings and projected earnings. With a ten percent disability and earnings at minimum wage, for instance, he figured a loss of $76,681. He recalculated damages for earnings at eight dollars an hour, and twelve dollars an hour. He also gave the jury his calculations for disability ratings of five percent, seven percent, and twenty percent.
. Testimony on Garland’s vocational disability was clearly relevant to his prayer for damages for his loss of earning capacity. Consequently our analysis focuses on the reliability of such testimony.
. In Mamette we cited our workers’ compensation case law defining the difference between permanent medical impairment and vocational disability:
Permanent medical impairment is related directly to the health status of the individual, whereas disability can be determined only within the context of the personal, social, or occupational demands, or statutory or regulatory requirements that the individual is unable to meet as a result of the impairment.
Marnette, 485 N.W.2d at 598 (citing Cozine v. Midwest Coast Transport, Inc., 454 N.W.2d 548, 552 (S.D.1990)) (italics omitted).