Concept
Beyond Qualification
Qualifications of witnesses
Beyond Qualification
Qualification gets a witness through the courtroom door and no further. Rule 702 allows a witness “qualified as an expert by knowledge, skill, experience, training, or education” to testify in the form of an opinion or otherwise only if the proponent demonstrates to the court that it is more likely than not that four further conditions are met:1
- the expert’s specialized knowledge will help the trier of fact understand the evidence or determine a fact in issue (Rule 702(a));2
- the testimony is based on sufficient facts or data (Rule 702(b));3
- the testimony is the product of reliable principles and methods (Rule 702(c)); and4
- the expert’s opinion reflects a reliable application of those principles and methods to the facts of the case (Rule 702(d)).5
These conditions attach to the testimony and not to the witness, so they bind an expert whose credentials no one disputes. They exist because experts enjoy a latitude that lay witnesses lack. A lay witness must testify from personal knowledge and cannot rely on hearsay. An expert may testify on the basis of hearsay and of general knowledge gleaned from education and experience, so long as other experts in the field “would reasonably rely” on such information.6 Jurors cannot easily check an opinion built on that foundation, because they need the expert precisely when they lack the knowledge to draw the inference alone. The Rule therefore makes the trial judge a gatekeeper who must exclude unreliable expert opinions, and the judge’s inquiry follows three foundation requirements: sufficient facts or data, reliable principles and methods, and reliable application of those methods to the matter at hand.7 The current text of the Rule states those requirements on its face, having been amended in response to the Supreme Court’s decision in Daubert.8
The judge decides each condition. Rule 702 calls on the trial court to make several preliminary determinations under Rule 104(a): whether the witness is qualified, whether the testimony will assist the trier of fact, and whether the facts, methods, and application behind it warrant jury acceptance.9 The proponent carries the burden on each, measured by the more-likely-than-not standard.10 A witness can therefore fail at any step, whether for lack of helpfulness, a thin factual basis, an unreliable method, or a sound method misapplied to the case. Exam hypotheticals usually plant a defect at one of those steps, so the discipline is to run the whole sequence and name the step that fails.
The four conditions follow in the Rule’s own order, beginning with the one that connects the expert’s knowledge to what the jury needs.
Helpfulness to the Jury
The first condition on the testimony itself is that the expert’s “scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue.”11 The trier of fact is the jury or, in a bench trial, the judge. Qualification asks whether the witness knows something worth hearing. Helpfulness asks whether what the witness proposes to say will give the fact finder something it needs on this record. The two questions come apart whenever a well-credentialed witness is offered on a point jurors can resolve alone, or on a topic that bears on no issue in the case.
The condition guards against the costs of expert testimony. The testimony takes time, and jurors may give great deference to someone with significant professional qualifications, so a party may introduce an expert only when the court is persuaded that the jurors, or the court itself in a bench trial, will benefit from help on the topic.12
The test is help, and it replaced a stricter idea. In older cases the courts often required that the subject lie outside the reach, “beyond the ken,” of the trier of fact, which stressed the highly specialized nature of the subject.13 Rule 702 asks for less. The expert need only bring some additional insight beyond that of the average person, and the gap need not be monumental. An expert may aid the jury even on familiar matters when training or experience gives a more thorough understanding than ordinary experience does, because evidence is helpful when it supplies further depth or precision about subjects well within common experience.14 The standard is too vague to call a test, and courts apply it generously. They have admitted testimony on the typical structure of mail fraud schemes, a fire marshal’s opinion in a fire-loss suit that arson started the blaze, and a DEA agent’s description of the purity of heroin commonly sold on the street.15 Some courts still resist testimony on matters of common knowledge. The Texas Supreme Court in K-Mart Corporation v. Honeycutt found error in admitting a human-factors and safety expert’s opinion that a shopping-cart corral lacking a top rail posed an unreasonable risk, because the subject lay within the jurors’ common knowledge.16
Testimony fails the first question when it addresses something other than the case on trial. Expert testimony does not help if it is unrelated to the facts at issue or rests on factual assumptions the evidence does not support, and it also fails when its reasoning is so illogical that it cannot affect the probability of any fact at issue. An economist’s projection of a decedent’s lost future income in a wrongful death case helps no one if it assumes past earnings that have no basis in the facts.18 Knowledge that is sound for one purpose does not become helpful on a different question merely because the witness holds it.
Testimony that relates to the case can still fail the second question, because the jury may not need it. A trier of fact does not need an expert to say that persons wearing gloves do not leave fingerprints, or that compulsive gamblers characteristically mismanage their finances.19 Nor does it need an expert to explain the enormity of a physician’s blunder in amputating the wrong leg of a patient.20 The line can run through a single subject. Jurors understand for themselves that memory weakens with age, so an expert on that general point adds little, but if a party claims that a specific disease impaired her memory, an expert can properly help with the disease.21 These cases fit together once helpfulness is read as a question about what the expert adds. An expert who supplies depth or precision on a familiar subject helps. An expert who restates what every juror already knows does not.
The third question protects the roles of jury and judge. Virtually all courts would exclude an expert’s statement that the expert believes a complainant, because the expert would be vouching for a witness’s credibility.22 Explaining behavior is different from vouching. If an alleged victim of a parking-lot beating told no one because of the effects of earlier armed-forces experience, an expert may explain that silence, which undercuts the inference that no attack occurred, but may not say that the victim is truthful in general.23 Courts also often exclude expert opinions that state legal conclusions without telling the jury anything about the facts, as when an expert would testify that the defendant’s conduct was negligent.24
The fourth question rarely excludes an opinion by itself. Courts generally assume that expert testimony can help even when the opinion is equivocal, because the jury can weigh the expert’s lack of certainty, and an expression of certainty on a complex matter often masks bias or incompetence. Testimony based on rank speculation is another matter, and courts reject it.25
Expert testimony on eyewitness identification shows how helpfulness depends on the topic and the record, because courts divide over it. A court must decide whether jurors can evaluate the risks of mistaken identification from their own knowledge and experience, or whether expert information would help them reach a better conclusion, and trial courts have been affirmed for admitting and for rejecting the testimony.26 The form of the testimony matters. Courts often permit general background about the factors that increase the risk of mistaken identification, but most forbid the expert from opining that a particular eyewitness is mistaken.27 The record matters too. In United States v. Stevens, the defendant offered an expert, Dr. Penrod. The district court admitted his testimony on cross-racial identification, weapon focus, and stress because it was tied to the facts of the case: the victims were white and the assailant black, the assailant brandished a gun, and both victims were under stress. It excluded his testimony on the suggestiveness of the wanted board, the “relation-back” question, and the correlation between confidence and accuracy, because it found no “fit.” The court of appeals reviewed that analysis for abuse of discretion.28
An answer on this topic should be keyed to the record. Eyewitness-identification experts are neither automatically admitted nor automatically excluded. The strong answer asks which topics the expert would address, whether each ties to the facts of this identification, and whether the expert stays with general research or crosses into telling the jury that a particular witness is wrong.
Sufficient Facts or Data
Rule 702(b) requires that the testimony be “based on sufficient facts or data.”30 Helpfulness asks whether the expert’s knowledge matters to the case. This condition asks whether the expert gathered enough information to support the opinion offered. It is the Rule’s quantitative standard, and credentials do not meet it: even when an expert is qualified and the opinion would be helpful, the Rules require a proper foundation, which means enough information to support the opinion.31
The Rules need a separate quantity check because they relax the quality check. An expert may rest an opinion on hearsay and on general knowledge from education and experience, where a lay witness could not.32 That freedom removes the ordinary guarantee about where a witness’s information came from, so the Rules ask instead whether the expert had enough of it.
Rule 703 supplies the qualitative standard and answers a different question: what kinds of information may an expert use. An expert may testify on facts or data of three sorts: facts learned by firsthand observation beforehand, facts learned at the trial or hearing, and outside facts or data gleaned from other sources before trial, provided they are the kinds of data on which experts in the field would “reasonably” rely.33 An expert can pass Rule 703 and fail Rule 702(b), as when every item consulted is the kind a colleague would trust but the items are too few to support the conclusion. An objection that the expert looked at too little invokes sufficiency, and the answer that experts may rely on hearsay does not meet it.
The at-trial route has a standard form. If earlier witnesses have supplied all the evidence on which the expert’s opinion rests, the examiner need not pose a hypothetical question and may instead ask the expert to assume, for purposes of giving an opinion, that the facts to which the preceding witnesses testified are true.34
Two Supreme Court and circuit decisions mark the line between a thin basis and an adequate one. In Pelletier v. Main Street Textiles, LP, the First Circuit affirmed the exclusion of an expert on safety practices who never inspected the facilities and equipment at issue, because he lacked sufficient facts and data on which to base his opinion.35 His general knowledge of safety practices could not fill the gap, since he had no information about the place and machines the jury was being asked to think about. In Whole Woman's Health v. Hellerstedt, the expert’s opinion rested on his participation, with other university researchers, in research that tracked the number of open abortion-care facilities in Texas by requesting information from the state health department, interviewing clinic staff, and reviewing publicly available information. The Court held that the district court acted within its authority in admitting the testimony.36
The two experts differ in the link between what was gathered and what was asserted, and a longer résumé has nothing to do with it. The safety expert never examined the subject of his opinion, while the researcher reached his through a documented effort to collect information about the very facilities in question. No fixed list of required sources exists, because what counts as enough depends on what the opinion claims.
The judge decides sufficiency as a preliminary question under Rule 104(a),37 and the Rule puts the burden on the proponent to show that the basis more likely than not suffices.38 The judge’s final determination keeps sufficiency tied to application: the method must be applied to “facts, data, or opinions sufficiently established to exist,” including those reasonably relied upon under Rule 703.39 Once the court finds the basis sufficient, further criticism of it, such as that the expert did not read every study on the subject, goes to the weight the jury gives the opinion and not to admissibility. A thin record cuts the other way: if the court cannot find that the basis more likely than not suffices, the opinion stays out.
Sufficiency overlaps with the reliability conditions that follow, and one weakness can be framed under more than one of them. In Kumho Tire, the expert asserted that the remaining tire tread had a depth of 3/32 of an inch, while the opposing expert’s apparently undisputed measurements ranged from .5/32 to 4/32 of an inch around the tire. If the conditions are treated separately, one can argue that a theory resting on four signs of abuse lacked an adequate factual basis because the fourth sign was never established, or that a theory using only three signs lacked a reliable method.40 An answer on a thinly supported opinion should name the sufficiency condition first and then check whether the others apply.
Reliability After Daubert
Rule 702(c) requires that “the testimony is the product of reliable principles and methods,” and the proponent must show the court that this condition is more likely than not satisfied.42 Sufficiency counted the information beneath the opinion. This condition tests the technique that turns the information into an opinion, and asks whether that technique has a sound footing in the field it comes from. A witness with sterling credentials and a mountain of data can still fail it, because credentials vouch for the person and data supplies raw material, while the technique is the reasoning that connects the two. The judge’s ultimate question is whether enough assurance of trustworthiness exists to warrant jury acceptance of the expert’s explanation as the product of reliable principles and methods.43
For decades the test for scientific evidence was general acceptance. Beginning in the 1920s, courts applied the Frye test, which required the proponent of testimony based on scientific procedures to show that the procedures were generally accepted in their field.44 The standard took its name from the two-page decision in Frye v. United States, and the concern behind it was that false and unreliable evidence would be offered and juries would be unable to appraise it wisely.45 Under Frye the consensus of scientists in the expert’s field controlled admissibility. The test protected the process against evidence that might later be discredited, but it imposed a lag, because testimony on theories at the cutting edge was rejected until the field caught up.46
Daubert v. Merrell Dow Pharmaceuticals, Inc. ended Frye’s run in the federal courts. In 1993 the Supreme Court discarded Frye in favor of a more flexible approach that was still designed to ensure the reliability of scientific evidence.47 The Court reasoned from the statute. Frye was a creature of purely decisional law, and Rule 402 makes all relevant evidence admissible unless the Constitution, a statute, or another Rule excludes it, so a common-law doctrine could not serve as an extra barrier.48 Dropping Frye did not drop the screen. Because the Rules give experts wide latitude, it would not suffice to say that a qualified expert may testify to anything and that the adversary process is a sufficient safeguard, so the Court made the trial judge responsible for deciding the reliability, or scientific validity, of the testimony and its fit to the issues in the case.49 Kumho Tire followed in 1999 and extended the Daubert standard to all expert testimony presenting technical or specialized material.50
General acceptance survives as one factor. It is no longer enough to poll the scientific community; the judge must decide whether the expert’s methodology is actually reliable, and that decision is sustained on appeal unless the judge abused her discretion. Often the judge will hold a hearing out of the jury’s presence to consider the data and the arguments for and against admissibility.51
Daubert supplied a nonexclusive checklist for examining an expert’s theory or technique:52
- Can the technique be tested, and has the testing been done?
- Has it been described in scientific publications subject to peer review?
- What are its known or potential error rates?
- Are there standards that control its operation, and were they used in developing the testimony?
- Has it gained acceptance in a relevant community?
Each question asks whether anyone besides the proponent has checked the method. The factors speak to the soundness of a method and not to the truth of a result. Peer review and publication, for instance, do not guarantee that the conclusions reached are correct, since much published research meets intense skepticism and is not borne out by later work.53 The checklist is a set of aids and not a test to apply item by item. The Advisory Committee’s Note to the 2000 amendment, which brought Rule 702 into line with Daubert and Kumho Tire, stresses both the judge’s gatekeeping function and the variety of factors relevant to admission, and says no single list of concerns will suffice for all cases.54
A software case shows the list working flexibly. In United States v. Chiaradio, the district court held a Daubert hearing and found a government program, EP2P, reliable because it had no error rate, had never yielded a false positive, and produced results that could be independently verified. The program’s source code had never been peer reviewed, but the court accepted the explanation that the government needed to keep the code secret.55 One missing factor therefore did not defeat reliability when the other indicators were strong and the gap had a reason.
The judge examines the method and not the correctness of the conclusion. On remand in Daubert, the Ninth Circuit put it this way: the test is “not the correctness of the expert’s conclusions but the soundness of his methodology.”56 Two opposing experts can each use a sound method while only one reaches the right answer, and a judge who excluded whichever expert seemed wrong would be deciding the case in place of the jury. Testimony within the range where experts might reasonably differ reaches the jury, which decides among the conflicting views even when the evidence is “shaky.”57 Conclusions still matter at the margin: the Advisory Committee’s Note observes that when a conclusion departs sharply from those of the expert’s peers, there is reason to ask whether the principles and methods have been faithfully applied.58
The same remand shows a method that fails. The plaintiffs claimed that Bendectin caused their injuries, and their experts assured the court that their methods were sound. The Ninth Circuit held that an expert’s bald assurance of validity is not enough: the proponent must show that the findings rest on sound science, which requires some objective, independent validation of the method.59 The experts had none. None had published his Bendectin work in a scientific journal or asked colleagues to review it, the only review the work had received came from judges and juries, and apart from the defendant’s experts no one in the scientific community had thought the studies worth verifying, refuting, or even commenting on.60 One expert offered only the fact that Bendectin is a teratogen and that he had examined the plaintiffs’ medical records, with no tested or testable theory for how, from that limited information, he could eliminate all other potential causes of birth defects.61
The dentist and the psychic both claim expertise, and the result differs according to whether anyone besides the witness has checked the method. The dentist’s method has passed outside checks, and the psychic’s cannot be checked at all. The Bendectin experts fall on the failing side because their proponents could point to no outside confirmation.
The gatekeeping role does not make the judge the referee of every expert dispute. Testimony that falls within the range where experts may reasonably differ goes to the jury, and attacks on shaky but admissible testimony then proceed in the ordinary way, through cross-examination, contrary evidence, and instruction on the burden of proof. A stem in which the opponent attacks the expert’s conclusion, or points to a rival expert who disagrees, raises a weight point for the jury. A stem that describes a technique no one has tested, reviewed, or measured for error, or one built for the lawsuit and backed only by the expert’s say-so, raises the reliability condition and gives the judge a ground to exclude.
Frye has not disappeared from American courtrooms. Almost half the states follow Daubert, but fourteen states, including California and New York, do not, and most of those follow their own versions of Frye.62 A question set in a state court applying general acceptance is testing the older rule, while a federal question is governed by Rule 702 as amended in response to Daubert, which formally requires reliable principles and methods, and by the Daubert inquiry.63
Kumho and Technical Experts
Daubert addressed scientific testimony, which left open whether the judge’s gatekeeping duty also reaches technical and other specialized testimony. The Eleventh Circuit answered narrowly. It held that Daubert applies only where an expert relies on the application of scientific principles and not where the expert relies on skill- or experience-based observation, and it treated the tire expert’s testimony as experience-based and outside Daubert.65
The Supreme Court rejected that limit in Kumho Tire Co. v. Carmichael. Rule 702 imposes on the trial judge a special obligation to ensure that any and all expert testimony is not only relevant but reliable.66 Daubert’s general principles apply to all the expert matters Rule 702 describes: the Rule, in respect to all such matters, “establishes a standard of evidentiary reliability,” and where the factual basis, data, principles, methods, or their application are called sufficiently into question, the judge must decide whether the testimony has “a reliable basis in the knowledge and experience of [the relevant] discipline.”67 The gatekeeping principles therefore apply to engineers and other technical specialists, not just to “scientists.”68
The Court also gave a practical reason that still shapes the doctrine. A gatekeeping duty that turned on whether testimony was “scientific” or “technical” would be difficult, if not impossible, to administer, because no clear line divides them. Engineering rests on scientific knowledge, and pure scientific theory itself may depend on observation and properly engineered machinery.69 Judges would have to litigate the label before reaching reliability.
Extending the duty to all experts did not make the Daubert checklist a mandatory test. The Court rejected the Eleventh Circuit’s view that a judge may ask Daubert-type questions only of scientific experts. Rule 702 creates no schematism that segregates expertise by type while mapping certain kinds of questions to certain kinds of experts, because life and the legal cases it generates are too complex for so definitive a match.70 Some of Daubert’s questions can help even with experience-based testimony: a judge may ask how often an engineering expert’s experience-based method has produced erroneous results or whether it is generally accepted in the relevant engineering community, and may ask even of a perfume tester who can distinguish among 140 odors at a sniff whether his preparation is of a kind others in the field would recognize as acceptable.71 Whether the specific factors are reasonable measures of reliability in a particular case is a matter on which the law grants the trial judge broad latitude. That latitude lets the judge avoid unnecessary reliability proceedings in ordinary cases where an expert’s methods are properly taken for granted, and require appropriate proceedings in the less usual or more complex cases where cause for questioning arises. What stays constant is the aim of making certain that an expert, whether relying on professional studies or personal experience, brings to the courtroom the same intellectual rigor that characterizes the practice of an expert in the field.72
The case began with a minivan whose right rear steel-belted radial tire blew out, killing one passenger and severely injuring others. The Carmichaels sued Kumho Tire, the maker and distributor, in federal court, and relied on deposition testimony by Dennis Carlson, an expert in tire failure analysis.73 The district court did not doubt his qualifications. It excluded his testimony because, despite those qualifications, it found unreliable the methodology he used to analyze the data from his visual inspection of the tire.74
Carlson’s method was a rule of thumb. He testified that in the absence of at least two of four signs of abuse (proportionately greater tread wear on the shoulder, grooves caused by the beads, discolored sidewalls, and marks on the rim flange), he concludes that a defect caused the separation. His analysis also depended on a further implicit proposition: that his visual and tactile inspection could determine that the tire before him had not been abused, even though it showed some of the very signs he looked for and had two punctures.75
The Supreme Court held that the district court’s doubts and its ultimate conclusion were reasonable.76 The reason lies in how narrowly the Court framed the question. The issue was not whether tire experts may reasonably use visual and tactile inspection to decide whether overdeflection caused a tread to separate. It was whether that approach, together with Carlson’s particular method of analyzing the data, could reliably support a conclusion on the particular matter to which his testimony was directly relevant: the likelihood that a defect in this tire caused its tread to separate from its carcass.77 On that question the record showed nothing beyond Carlson’s own assurance that his two-factor test was accurate, with no indication that other experts in the industry use it and no article or paper validating it.78 The district court had not counted failed factors mechanically. It agreed that Daubert should be applied flexibly and that its factors were illustrative, and it still found insufficient indications of reliability in the way Carlson analyzed the data.79
Carlson’s case separates two questions that are easy to run together. Qualification vouches for the person: an engineering background and years of work on tires show that Carlson is the kind of witness who can speak about tire failures. Reliability tests the route from what the expert observed to what the expert concludes, and the route here was a rule of thumb supported by nothing but Carlson’s own word. A strong résumé cannot rescue the opinion, because it answers the first question and is silent on the second.
Experience-based experts are not doomed by this holding. The relevant reliability concerns may focus on personal knowledge or experience, and a court may sensibly ask whether the expert’s preparation is of a kind others in the field would recognize as acceptable.80 For a practical trade, a useful framing asks whether members of the public routinely turn to it for services other than testimony and whether a feedback loop alerts a practitioner when she has erred. An auto mechanic may lack the formal education to qualify as a scientific expert, but her customers tell her whether her repair worked.81 A related first step is to identify the particular methodology the expert will rely on, since it is unacceptable for an expert to rely on a methodology that is literally ineffable.82
Carlson and the builder both rest on years of experience, and neither can point to a published validation study. The difference lies in what stands behind the method. The builder’s probe is the trade’s own check, a wrong diagnosis would be exposed by customer feedback, and the builder applied the method to every joist in this deck. Carlson’s two-factor rule was shown to be used by no one else, and nothing outside his own word showed that it could tell a defect in this tire from abuse. The question is whether the method connecting experience to conclusion has support beyond the witness’s own assertion.
The same attention to method appears in state valuation cases. M.G. Bancorporation, Inc. v. Le Beau extends the point to business valuation, which is other specialized knowledge: the Court of Chancery concluded that the respondents had failed to establish that Reilly’s capital market methodology is generally accepted by the financial community for valuing bank holding companies, as distinguished from other enterprises. It also found that the method included a built-in minority discount, which made it improper in a statutory appraisal proceeding even if otherwise acceptable.83 General acceptance there served as one measure of reliability within the field, which fits the Kumho approach of choosing measures that suit the expertise.
Rule 702 was amended in 2000, largely codifying Kumho’s prescriptions.84 The amended Rule permits expert testimony only if it is grounded on sufficient facts or data and is the result of reliable principles and methods that are themselves reliably applied to the facts of the case.85 Its text states those requirements for every kind of expert, so the engineer and the appraiser meet the same conditions as the scientist.86
A stem that describes the witness as an engineer, a tire technician, or an appraiser and has the proponent argue that reliability review is for scientists offers a rule that Kumho rejected. A stem noting that the expert’s method was never peer reviewed does not by itself require exclusion either, because peer review is one measure that may not fit the field. The answer turns on whether the expert’s way of getting from what was observed to what is concluded is reliable for this question, and credentials answer only the qualification question.
Reliable Application
Rule 702(d) supplies the last condition: the proponent must show that it is more likely than not that “the expert’s opinion reflects a reliable application of the principles and methods to the facts of the case.”88 Subdivision (c) tests the method in the abstract by asking whether the technique has a sound footing in its field. Subdivision (d) follows the method into the case and asks whether the expert used it soundly on these facts and whether the conclusion stays within what the method and the data can support. The two come apart. The bite-mark comparison in the previous subsection cleared subdivision (c) because the technique had been tested, reviewed, and measured for error, but it would still fail subdivision (d) if the dentist skipped the field’s required steps or announced more than the comparison showed.
Two quiet failures survive general validation. The first is execution: a technique validated by testing and review yields dependable results only when its steps are carried out on adequate, accurately gathered inputs. The second is overreach: even a perfectly executed method supports conclusions of only a certain strength, and an expert may state a stronger one. Jurors cannot detect either failure, so the check falls to the judge. Whether the expert reliably applied the method is a preliminary question that affects admissibility under Rule 104(a), and not a matter left to the jury as weight. Decisions before the amendment split on the point, and because the same issues also bear on weight, the jury can take another look at the element if the evidence is admitted.89 An expert who relies primarily on experience must explain how that experience leads to the conclusion.90
Kumho Tire contained an application defect alongside the method problem. Carlson testified that most tires have some bead-groove pattern, that where there was reason to suspect an abnormal bead groove he would ideally “look at a lot of [similar] tires” to learn the groove’s significance, and that he had not looked at many tires similar to the one at issue.91 A method that calls for comparison with many similar tires cannot be reliably applied by an examiner who has looked at few.
The Court in Kumho Tire treated the accuracy of the inspection data, the soundness of the method, and its application together, moving among the Rule’s requirements without separating them and in effect looking at the three as one requirement.92 The four signs of abuse show why. A theory that needs four signs and finds only three can be attacked as resting on an inadequate factual basis, because the fourth sign was never established, or as unreliable, because a three-sign theory was never shown to be sound, or as misapplied, as with the bead-groove testimony.93 The judge’s final determination asks one question: whether the expert’s explanation, as applied to facts sufficiently established, carries enough assurance of trustworthiness to warrant acceptance by the jury.94 Subdivision (d) is the best label when the complaint concerns execution on these facts or the reach of the conclusion, and an answer that names it should note the overlap.
Because subdivision (d) speaks of the expert’s opinion, the judge can test each opinion against the basis and method behind it, and an expert may apply a method reliably for one conclusion and not for the next. The check matters most in forensic comparison work, where examiners match marks by subjective judgment. A ballistics examiner who says that the bullet he tested “matches” the bullet found at the scene, and that his rate of error is “zero,” overstates the case: subjective judgment carries a real rate of error, and it has not been established that each gun makes a unique marking on a bullet, so the expert cannot validly testify to a “match.”95
The dentist and the ballistics examiner both rely on subjective comparison, so the type of method does not separate them. What separates them is the tie between the claim and the work. The dentist’s conclusion rests on a comparison precise enough to support it and on a demonstrated error rate, while the examiner’s claims of a match and of zero error rest on nothing the method can supply.96
The third step is where Carlson’s analysis broke down, because the method he described called for comparisons he had not made, and the fourth is where forensic testimony most often fails.