for the Court:
¶ 1. In this personal-injury, products-liability ease, the jury awarded the plaintiff $7 million, finding that Trellvion Gaines had been brain-damaged from exposure to lead. In an effort to convince this Court to reverse the trial judgment and to render judgment in its favor, the defendant, Sherwin-Williams Company, challenges the reliability of plaintiffs causation experts and fact witnesses. In the alternative, Sherwin-Williams alleges that unreliable, unfairly prejudicial, and untimely disclosed expert testimony and a biased jury pool require a new trial.
STATEMENT OF THE CASE
¶ 2. Trellvion Gaines (“Trellvion”), the plaintiff, was born in 1991 and lived at all relevant times in a wooden house on Highway 553 in Fayette, Mississippi, with his mother Sherneker Pollard (“Pollard”), grandmother Doris Gaines (“Gaines”), and step-grandfather Johnny “Cracker” Crawford (“Crawford”), who bought the house in late 1978 or 1979. The house was built in the early 1900s and was burned in 1994. Expert testimony conclusively established that samples of remains at the house site contained lead-based paint. Two blood tests in September 1993 showed that Trell-vion had elevated blood lead levels of 30 micrograms per deciliter (30 p,g/dL) and 19 p-g/dL five days later. Trellvion now has significant cognitive deficiencies that he claims are a result of lead poisoning.
¶ 3. Fact witnesses for Trellvion testified that they had bought and applied lead-based paint manufactured by the Sherwin-Williams Company (“Sherwin-Williams”) to the interior and exterior of the house at various times. Sherwin-Williams responded with testimony and corporate documents to support its contention that it began to eliminate lead in its products in the 1930s and stopped manufacturing any *43interior lead-based paint in 1954 and all residential lead-based paint in 1972. This lawsuit was filed in 2000. Summary judgment was granted for Sherwin-Williams by the Circuit Court of Jefferson County, affirmed by the Court of Appeals, then reversed by this Court on writ of certiorari in 2007.1 The trial was held in April 2009 in the Circuit Court of Jefferson County before the Honorable Lamar Pickard, and a unanimous jury found for the plaintiff and awarded $7 million in compensatory damages. Because this case has multiple, complex issues, relevant facts are included in the discussion below.
DISCUSSION
I. Did the trial court err in denying Sherwin-Williams’ motion for judgment notwithstanding the verdict?
¶ 4. Whether a party is entitled to a judgment as a matter of law is an issue that we review de novo.2 Evidence should be considered in the light most favorable to the appellee, giving that party the benefit of all favorable inferences which may reasonably be drawn from the evidence.3 We will affirm the denial of a motion for directed verdict or a motion for judgment notwithstanding the verdict if there is substantial evidence to support the verdict, but we will reverse if the evidence, as applied to the elements of a party’s case, is either so indisputable, or so deficient, that the necessity of a trier of fact has been obviated.4 Sherwin-Williams offers three arguments on this issue, however the first argument is dispositive.
Did the plaintiff offer sufficient proof that his injuries were caused by Sherwin-Williams lead-based paint?
¶ 5. The plaintiff concedes that no scientific study has causally connected permanent brain damage to a single, asymptomatic elevation of blood lead to 30 |xg/dL, followed by a rapid decline to baseline. However, he argues, “It was not the opinion of [our] expert witnesses that this case was a ‘single’ exposure. No matter how many times the Defendant says it was a single exposure, does not make it true. Trellvion was exposed to toxic Sherwin-Williams’ lead paint dust for four (4) years.” We must determine whether the trial court erred in determining that the core opinion — lead poisoning caused cognitive impairments — was reliably based on science and the facts presented.
¶ 6. Dr. John Rosen, one of two causation experts, testified by video deposition. He opined that Trellvion was poisoned by lead from infancy through late 1993. He achieved this opinion through “differential diagnosis.” He based his differential diagnosis on “the apparent fact that [Trellvion] was excessively exposed to lead as a young infant.” He noted that Trellvion’s only documented elevated blood levels were 30 |xg/dL on September 15, 1993, and 19 Ixg/dL on September 20,1993.
¶ 7. “Exposure” to lead — or just being around it — is not dangerous. It must be ingested to poison a person. Dr. Paul Mushak, accepted as an expert in the field of toxicology, testified that blood lead levels of 10 |xg/dL or higher are toxic. He further testified that ingestion of fine lead dust is a major pathway for ingestion of lead. His opinion was that Trellvion had *44been “exposed” to a “lead environment” from birth until the house burned, and that lead paint was the source of Trellvion’s elevated blood levels in September 1993.
¶ 8. Dr. Theodore Lidsky, entered as an expert in the field of neuropsychology, performed a differential diagnosis on Trellvion to determine that his brain had been injured by lead poisoning, finding that, without the lead poisoning, Trellvion’s IQ would have been in the “in the bottom of the low average range.” Lidsky explained that, while “lead is eliminated relatively rapidly from the blood ... even when it’s gone from the blood, it’s still in the brain.” To determine whether Trellvion had a brain injury, Lidsky used a method called “deficit measurement ... which is where you look at a person’s baseline level of functioning and look to see if there are scores that deviate ... from that baseline.” The baseline is the level at which one would expect that person to function, but for brain injury. Lidsky testified that “as a matter of course” professionals add points to the patient’s known IQ score to determine the baseline in “lead poisoning” cases because “we know his IQ goes down. It’s one of the most established points in science.... So you have to add a number of points.” Lidsky added ten points to Trellvion’s full-scale IQ (or six to his higher score, a 74 verbal IQ).5
¶ 9. Lidsky explained this “baseline” method further:
When I tested [Trellvion] in 2001, his IQ was about almost 69.... You look at that as his baseline and then you look at his testing. If there’s an impairment at that point, without adjusting his baseline, then you adjust the baseline. If there’s no impairment at that point, then you say, this is what he is, and you can’t adjust it.... Before [I] even looked at his baseline [I] know he’s injured. At that point, he starts at the baseline. Before, like you said, you are assuming the child is injured just by looking at his lead level. You can’t do that.
Lidsky agreed that the neuropsychological tests that he uses are not designed to detect the cause of any particular deficit. He testified, nevertheless, that he still could conclude that Trellvion’s injury was caused by lead. Lidsky testified that no studies had been done linking “acute exposures” to lead with IQ reduction because “based on what the CDC has stated, there are no acute exposures.”
¶ 10. Lead poisoning was, in the opinion of Dr. Rosen, the cause of permanent intellectual impairments uncovered by Dr. Lid-sky. When asked if he would have diagnosed Trellvion with lead poisoning in the absence of the two elevated blood lead results, Dr. Rosen responded,
I would have recognized — I would have diagnosed him as having the disease of — of excessive lead exposure, and that disease is a pervasive disease that involves every organ of a child’s body including their brain, and that is one aspect or one compartment so to speak of ultimately the fact that he was exposed to enough lead in the initial phase of child — of excessive exposure to lead to elevated blood leads.
Finally, when asked how he was able to rule out an idiopathic cause for Trellvion’s mental retardation,6 Dr. Rosen replied, “I *45ruled it out because there is nothing in the medical records to indicate that there is any — any so called etiopathic [sic] or unknown cause. What is unknown is unknown. ...”
¶ 11. The plaintiffs experts seemingly contradict each other and themselves. For example, Dr. Lidsky testified that adding points to determine a baseline IQ is a “matter of course” in “lead poisoning” cases because “we know his IQ goes down ... [ijt’s one of the most established points in science.” However, Dr. Mushak, plaintiffs expert toxicologist, testified that “no one has done the data, gathered the data to show” whether “any time a child has an elevated blood lead level, it’s going to result in the loss of IQ points.” And Dr. Lidsky at one point testified that finding the injury depends on calculating the baseline (including adding points assumed by lead poisoning), but later asserted that it is improper to assume that a child is injured based solely on an increased blood lead level. Further, it is difficult to determine whether Lidsky’s opinion in the case is, proverbially, a chicken or an egg. Dr. Rosen assumed, in reliance upon Dr. Lid-sky’s findings, that Trellvion was, in fact, injured (and not just “slow”) to differentially diagnose exposure to lead as the cause of the injury. However, Lidsky’s admitted method, “deficit measurement,” depended upon the calculation of a baseline, which depended upon assuming that Trellvion had been poisoned by lead. Dr. Lidsky was, essentially, leaning on Dr. Ro-sen’s theory of causation, who was leaning on Dr. Lidsky’s theory of injury, who was leaning on Dr. Rosen’s theory of causation ... ad infinitum. Finally, while Dr. Ro-sen testified that he could have diagnosed Trellvion with the disease of “excessive lead exposure” even without a documented elevated blood level, he subsequently testified that lead poisoning cannot be diagnosed without a blood lead test.
¶ 12. The plaintiff characterizes Drs. Rosen and Lidsky as “some of the world’s leading experts in their respective fields[.]” Sherwin-Williams responds that the trial court failed to exercise its “gatekeeping duty”7 because it allegedly did not “undertake any meaningful evaluation [of the reliability of the opinions] beyond expert credentials.”
¶ 13. Mississippi Rule of Evidence 702 and our familiar Daubert8 standard require trial courts to act as “gatekeepers”9 with regard to expert opinion testimony, because juries tend to place great weight on the testimony of experts and can be misled by unreliable opinions.10 But we are mindful that the exclusion or admission of expert testimony is within the discretion of the trial court,11 and, furthermore, vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are traditional and appropriate means of attacking shaky but admissible evidence.12
¶ 14. The opinions expressed by Drs. Rosen and Lidsky, however, were not reliable. A dose-response ratio is critical to determining the causal connection be*46tween a poison and an injury.13 In this case, the only known dose — or ingestion of lead — is represented by Trellvion’s elevated blood lead levels from September 1993. The experts extrapolated both dose and duration with only circumstantial supporting evidence, engaging in a classic logical fallacy: post hoc ergo propter hoc.14 Dr. Lidsky explained: “[H]e was raised in the house, the same house from the day of birth. What’s the likelihood that [Dr. Timm] just happened to [test Trellvion’s blood] on the day that his blood level was 30?” This opinion was not based on some epidemiological study calculating the likelihood of elevated blood levels based on the levels of paint in the home. Neither was it based on Trellvion’s medical records. The basis for Lidsky’s and Rosen’s causation testimony — that Trellvion was ingesting and being poisoned by lead the entire time he lived in the house — was mere speculation and inadmissible.15
¶ 15. We are urged by amici to instruct trial court judges that consideration of the “relevance and reliability” of an expert’s opinion (and not simply the expert’s qualifications) is a necessary part of the trial court’s duty as a gatekeeper. The trial judge in this case did not consider many of Sherwin-Williams’s arguments concerning the reliability of plaintiffs experts, regarding them as challenges to the weight and not the admissibility of the experts’ testimony. Amici urge that we instruct trial courts to make particularized findings as to the admissibility of expert opinions, when those challenges are raised before trial. In this case, a Daubert challenge was filed before trial, and the court chose to delay ruling until trial, by which point the court did not have time to read the briefs and fully consider the arguments. Our trial judges work exceedingly hard and have discretion in how they discharge their gatekeeping duty,16 but we take this opportunity to reiterate that such duty includes making sure that the opinions themselves are based on sufficient facts or data and are the product of reliable principles and methods.17
¶ 16. Because Drs. Lidsky’s and Ro-sen’s speculation that Trellvion had been ingesting lead throughout his entire residence at the home was unreliable and, therefore, inadmissible, and because the plaintiffs experts did not present any scientific authority that an acute, asymptomatic ingestion of lead could lead to the alleged injuries, the plaintiff did not offer sufficient proof of causation. That admission of the speculation of Drs. Lidsky and Rosen would, alone, trigger a new trial. However, because of the lack of sufficient proof of causation, we find that the court should have granted Sherwin-Williams’ motion for judgment notwithstanding the verdict.
¶ 17. The separate opinion opines that the fact-witness testimony presented by the plaintiffs is insufficient to create an issue for the jury. While the testimony of these fact witnesses is, in various ways and to varying degrees, incredible, self-contradictory, and circumstantial, this Court has held that even contradictory evidence and *47evidence that has been impeached creates an issue for the jury.18
¶ 18. In Rucker v. Hopkins, we said that incredible testimony and the testimony of a plaintiff which is so clearly and manifestly improbable should be disregarded; this also applies to testimony that has been clearly impeached.19 In Elsworth v. Glindmeyer, we reversed and remanded a jury verdict for a defendant in a wrongful death suit because “[n]o court is required to believe, or should be bound by improbable, incredulous, or unreasonable evidence ... which is violative of physical laws, human experience and common sense, and which intrudes into the realm of the supernatural.”20 However, subsequently, in Doe v. Stegall, we found error where a trial judge, relying on Rucker, excluded the testimony of a fact witness who had been “clearly impeached.”21 The logical extension of Stegall is that fact testimony, no matter how thoroughly contradicted or impeached, may create an issue for the jury, so long as it is not supernatural. Indeed, in Rucker, we said, “[ejven where the evidence is such that a judgment for the adverse party would have to be set aside as being contrary to the overwhelming weight of the evidence, it does not necessarily follow that a party is entitled to a directed verdict.”22 The trial court did not err in denying the motions for directed verdict and judgment notwithstanding the verdict based on this argument, because the fact-witness testimony created an issue for the jury.
CONCLUSION
¶ 19. Because Drs. Lidsky’s and Ro-sen’s speculation was inadmissible, and because the plaintiffs experts did not present any scientific authority that an acute, asymptomatic ingestion of lead could lead to the alleged injuries, the plaintiff did not offer sufficient proof of causation. Accordingly, we reverse and render on this issue, and not on whether the fact-witness testimony sufficiently established product identity.
¶ 20. REVERSED AND RENDERED.
WALLER, C.J., CARLSON AND DICKINSON, P.JJ., CONCUR. KITCHENS, J., CONCURS IN RESULT ONLY WITH SEPARATE WRITTEN OPINION JOINED BY CHANDLER AND KING, JJ.; RANDOLPH, J., JOINS IN PART. RANDOLPH, J., CONCURS IN PART AND IN RESULT WITHOUT SEPARATE WRITTEN OPINION. LAMAR, J., NOT PARTICIPATING.. See Pollard v. Sherwin-Williams Company, 955 So.2d 764 (Miss.2007).
. Solanki v. Ervin, 21 So.3d 552, 565 (Miss.2009).
. Spotlite Skating Rink, Inc. v. Barnes, 988 So.2d 364, 368 (Miss.2008).
. U.S. Fid. & Guar. Co. v. Martin, 998 So.2d 956, 964 (Miss.2008).
. Lidsky explained that "full scale” IQ is calculated by averaging scores on "verbal” and "performance” tests. Trellvion’s verbal IQ score was 74, and his score on the performance test was in the "high 60s.”
. The cause of as many as one half of cases of mental retardation is unknown or "idiopathic.” D.K. Daily, H.H. Ardinger, G.E. Holmes, Identification and Evaluation of Mental Retardation, 61(4) Am. Family Physician 1059-67, 1070 (Feb. 2000).
. Miss. Transp. Comm'n v. McLemore, 863 So.2d 31, 40 (Miss.2003).
. Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 590, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993).
. McLemore, 863 So.2d at 40.
. E.g. Edmonds v. State, 955 So.2d 787, 792 (Miss.2007).
. McLemore, 863 So.2d at 34.
. Hyundai Motor America v. Applewhite, 53 So.3d 749, 754 (Miss.2011) (citing Daubert, 509 U.S. at 595, 113 S.Ct. 2786).
. Watts v. Radiator Specialty Co., 990 So.2d 143, 147 n. 9 (Miss.2008).
. "After this, therefore, because of this.”
. Hill v. Mills, 26 So.3d 322, 329 (Miss.2010) (Expert testimony must proceed from "what is known, and the expert must have knowledge that is more than subjective or unsupported speculation.”).
. See Smith ex rel. Smith v. Clement, 983 So.2d 285, 289 (Miss.2008).
. Miss. R. Evid. 702; McLemore, 863 So.2d at 38.
. Rucker v. Hopkins, 499 So.2d 766, 770 (Miss.1986).
. Rucker v. Hopkins, 499 So.2d at 769.
. Elsworth v. Glindmeyer, 234 So.2d 312, 321 (Miss.1970).
. Stegall, 757 So.2d 201, 205 (Miss.2000).
. Rucker, 499 So.2d at 770.