Every discipline in this series entered a courtroom through the same door. This installment examines the door. It also identifies a Michigan rule that was copied from a federal rule in 2004, and has not moved since, while the federal version was twice amended to address exactly the problem the preceding seven installments documented.
Admissibility asks whether a technique is accepted by its field and, under the modern standard, whether a judge finds it reliable. Neither question requires validation research. Michigan adopted the federal reliability standard effective January 1, 2004 by modeling the version of Federal Rule of Evidence 702 then in force. That federal rule was amended in 2011 and again in 2023, the second time to address overstated forensic certainty. Michigan’s rule still reflects the 2000 text.
If Michigan already applies Daubert, is the wording difference significant?
The gatekeeping duty is the same either way. What the 2023 federal amendment changed is emphasis and specificity, directing attention to whether a stated conclusion exceeds what the method supports. That is the precise failure mode documented throughout this series, and rule text shapes what lawyers brief and what judges are asked to rule on.
Did a stronger admissibility standard keep discredited disciplines out?
Largely no. Bite mark comparison, hair comparison, and arson indicator testimony were admitted in Daubert jurisdictions well after 1993. A reliability standard is only as demanding as the record put in front of the judge, and defense counsel frequently lacked the resources to build that record.
Who is supposed to raise the validity question?
The opposing party, ordinarily by motion before trial. That allocation is the structural problem. It places the burden of detecting a discipline-wide failure on individual defendants, one case at a time, usually without an expert budget.
Is this a Michigan-specific problem?
No. Rule text varies by state and many have not tracked the federal amendments either. Michigan is examined here because it is the jurisdiction this publication covers and because the state has an open process for considering the change.
Why This Series Starts With a Rule
A series about discredited forensic disciplines could reasonably open with a case. I want to explain why this one opens with a rule of evidence instead.
The seven installments that follow examine specific methods: microscopic hair comparison, arson indicators, bite mark analysis, comparative bullet lead, the software now interpreting complex DNA mixtures, and the oversight body that was prevented from assigning responsibility for any of it. Each has its own history and its own failure.
They also have one thing in common that is easy to miss when the cases are read individually. Every one of them was admitted into an American courtroom, repeatedly, by judges applying a standard designed to keep unreliable evidence out. That standard did not fail because judges applied it badly. It failed because of what it asks, and because of who it makes responsible for asking. That is what this installment is about, and it is the reason the rest of the series reads the way it does.
What Admissibility Actually Asks
For most of the twentieth century, American courts evaluated novel scientific evidence by asking whether the technique had achieved general acceptance in the relevant field. Michigan applied a version of this through what its courts called the Davis-Frye standard.
The circularity has been noted for as long as the standard existed, and it is worth stating precisely rather than gesturing at. The relevant field for microscopic hair comparison consisted of microscopic hair comparison examiners. The relevant field for bite mark comparison consisted of forensic odontologists who performed bite mark comparisons. Asking whether those fields accepted those techniques was asking whether a group of practitioners endorsed the practice that defined them.
The Supreme Court’s 1993 decision in Daubert v Merrell Dow Pharmaceuticals replaced that inquiry in federal court with a direct reliability assessment, assigning trial judges a gatekeeping role and offering factors including whether the technique had been tested, whether it had been subject to peer review, whether error rates were known, and whether standards controlled its operation. Later decisions extended the framework beyond strictly scientific testimony.
That is a genuinely better question. It is also, as the preceding installments demonstrate, a question that discredited disciplines passed for decades after it was adopted.
A reliability standard does not generate evidence about reliability. It creates an occasion at which such evidence may be presented, by a party with an interest in presenting it and the resources to obtain it. Where the discipline is long-established, the prosecution’s expert is credentialed, and the defense has no funds for a competing expert, the occasion passes without the question being contested.
The Michigan Rule
Effective January 1, 2004, the Michigan Supreme Court amended MRE 702 to conform to the then-current text of Federal Rule of Evidence 702, which had itself been amended in 2000 to incorporate Daubert. The Michigan Supreme Court construed the amended rule that year in Gilbert v DaimlerChrysler Corp, 470 Mich 749.
Gilbert established the framework Michigan courts apply now. The Daubert factors are the criteria for assessing reliability, the analysis is flexible rather than a checklist, and the trial court carries a gatekeeping obligation to ensure that all expert opinion testimony is reliable regardless of whether it rests on novel science. Michigan courts have since emphasized that the rule reaches each aspect of the testimony, including the data underlying the expert’s theories and the methodology by which conclusions are drawn from that data.
On paper, this is a strong rule. The reach to methodology and underlying data is exactly what an inquiry into a discipline like hair comparison would require.
What Happened to the Federal Rule After 2004
Here is the finding that made this installment worth writing.
Federal Rule of Evidence 702 was amended in 2011. Michigan’s rule was not updated. It was amended again effective in 2023, and this amendment was not stylistic. The federal advisory committee considered various approaches to unreliable forensic testimony and chose to amend the rule to emphasize that each expert opinion must stay within the bounds of what can be concluded from a reliable application of the expert’s basis and methodology. The accompanying committee note states that forensic experts should avoid assertions of absolute or one hundred percent certainty, or of a reasonable degree of scientific certainty, where the methodology is subjective.
Read that against this series. The FBI hair examiners did not fabricate observations. They stated conclusions exceeding what microscopic comparison could support. The arson investigators did not invent burn patterns. They drew inferences their method could not sustain. Bullet lead analysts measured correctly and characterized the result to juries in terms the chemistry did not license.
The 2023 federal amendment is aimed with unusual precision at the failure mode that produced every case in this series. Michigan has not adopted it.
Michigan’s reliability rule reflects a federal text written in 2000. That is before the 2009 National Academy of Sciences report, before the 2015 FBI hair review, before the 2016 PCAST findings, and before every event documented in the seven installments of this series.
The Lab holds free Clutch Justice resources on Michigan court structure, jurisdiction, and record access, built for researchers and advocates working from public records without legal training.
Explore The LabWhy the Standard Was Not Enough
Three structural features explain how a demanding reliability rule coexisted with the record documented in this series.
A reliability challenge must be raised, briefed, and supported by the opposing party. In criminal cases that is a defendant, frequently represented by appointed counsel without an expert budget.
A discipline-wide failure is thereby assigned for detection to the individual least equipped to detect it.
Once a discipline has been admitted repeatedly, later courts rely on that history. The reliability inquiry becomes a question about prior rulings rather than about underlying research.
The first admission decision does the work for every subsequent one, and the first was usually made without validation evidence.
An exclusion in one county resolves one case. There is no mechanism by which a reliability finding propagates to other courts or triggers review of prior admissions.
This is the same absence identified in Parts II through VII from a different direction.
Together these produce the pattern the series documented. The rule is capable of excluding an unvalidated discipline. It just requires the right defendant with the right lawyer and the right expert to arrive in front of the right judge, and it produces no consequence beyond that courtroom when they do.
What Would Change
Harmonize MRE 702 with the 2023 federal amendment. This is the single lowest-cost item in the entire series. It requires no appropriation, no new agency, and no legislation. It is a court rule amendment, the mechanism for considering it already exists, and it would put into Michigan’s rule the language most directly aimed at experts claiming more certainty than a subjective method supports.
Funded expert access for reliability challenges. A gatekeeping standard that depends on the opposing party to trigger it is only as good as that party’s resources. Without a mechanism for appointed counsel to obtain a qualified expert for a validity challenge, the standard functions differently for defendants who can pay and defendants who cannot.
A record of reliability rulings. When a Michigan court excludes or limits testimony from a forensic discipline, that ruling should be recorded somewhere retrievable. At present it exists in one file in one county. This is the admissibility-side version of the register argued for in Part VIII, and it is the same instrument.
Why This Matters
I want to be clear about what this installment is claiming, because it is easy to overstate.
Described in the abstract, the admissibility framework looks like a system for keeping unreliable evidence out of trials. Watch it operate across the cases in this series and it looks like something else: a checkpoint with no staff, operated by whichever traveler happens to be able to afford an inspection, which keeps no log and shares nothing with the next checkpoint down the road.
Every discipline in this series passed through it. Hair comparison passed for decades. Arson indicator testimony passed while the research invalidating it was being published. Bite mark comparison passed after the National Academy of Sciences had identified its foundations as weak. None of that happened because the standard was too lenient in its text. It happened because a standard is a question, and a question only gets asked when somebody with standing and resources is present to ask it.
Michigan’s version of that question has not been updated since 2004. The federal version was revised in 2023 specifically to address forensic witnesses claiming certainty their methods cannot support, which is the precise mechanism by which every person named in this series went to prison.
Amending a court rule is the smallest thing this series asks for. It is also the only thing in it that could be done this year.
Part II takes up the first discipline: microscopic hair comparison. A federal review found erroneous statements in at least ninety percent of the trial transcripts it examined, then drew the boundary of that review at its own walls, excluding the state and local laboratories the FBI itself had trained. A man convicted in Detroit on a single hair examined by one of those laboratories served forty-one years and three months.
Continue Your Investigation
If this reporting raised more questions, use the Clutch Justice ecosystem to keep going.
