Chapter 39 — Key Takeaways

A scannable one-page card for the capstone. For the full assembly and the worked conclusion, see index.md. This chapter does not teach a new method; it teaches how to assemble and weigh everything the book taught — and how to state a conclusion at exactly its true strength.

The core claims

• A complete forensic case is an excluded field plus an independent convergence. You do not "prove" your way to a suspect; you exclude whom the science excludes until the field is small, then ask what the remaining independent threads jointly support.
• Assembly carries each thread forward at the strength it earned — never one degree stronger. A "consistent with" finding does not become "proves" because it now has company. The number of threads creates a feeling of certainty that no single thread supports; resisting that feeling is the whole discipline.
• Exclusion is the surest move, and it overrides even a confession. Three independent findings (DNA, cell-site, the dissected interrogation) excluded the man who confessed — the near-miss wrongful conviction. A confession is not exempt from exclusion.
• Convergence carries weight only to the degree the threads are independent. Independent improbabilities multiply; non-independent ones do not. If one shared cause (usually cognitive bias) bent all the threads the same way, their agreement is a counterfeit (the Beatrice Six) — no matter how overwhelming it feels.
• The honest conclusion is "strongly supports," not "proves." No stacking of limited threads (a DNA mixture, class toolmarks, area-level location, a co-owner's soil) ever sums to metaphysical certainty. "Strongly supports" is not a hedge; it is the strongest verb the evidence jointly earns.
• A conclusion you cannot bound is a belief, not a finding. The mark of a competent assembly is that it can name, out loud, everything its own conclusion does not establish — and can list the further tests that would strengthen or break it. That symmetry is what makes it science.

The Mill Creek result, at its true strength

The residual uncertainty (what the conclusion does NOT establish)

• The DNA is a mixture interpreted as a likelihood ratio, not a single-source match.
• The toolmarks are class, not individual — forced entry, not Keller's pry bar to the exclusion of all others.
• The soil, pollen, and cell-site are class/association evidence: an environment, a vehicle, a phone-area — not a man at a point, and not a guilty reason for a co-owner's soil.
• The scene was compromised in its first hours (treated as an accident); that lost integrity is permanent.
• The original analysis was biased by the accidental-fire frame.

What further testing would do (and why the symmetry matters)

• Resolve the mixture further (Y-STR, repeat probabilistic genotyping) — could strengthen or weaken the DNA.
• Tighten the location (any non-area source) — could move "in the area" toward "at the place," or not.
• Date the trace — could distinguish a guilty deposit from an innocent one, addressing the co-owner problem.
• Blind re-analysis of the contested interpretations — the single most valuable test; neutralizes the bias concession, and if it failed to reproduce a thread, that would be the most important finding of all.

A test you would run only if you were certain of the result is a ritual, not a test. The Mill Creek conclusion is strongly supported and falsifiable — and those two together, not certainty, are what make it science.

Convergence done right vs. the counterfeit

• Right (Golden State Killer, CS 39.1): central claim rests on single-source DNA (top of the validity spectrum); threads genuinely independent; each step at its true verb (genealogy = lead; direct comparison = individualization; rest = corroboration); exclusion did its work first. → a confident identification.
• Counterfeit (Beatrice Six, CS 39.2): a theory came first; multiple "confirmations" (including internalized false confessions) were not independent — all flowed from one contaminating source; forensic evidence read toward the theory. → six innocent people convicted, later excluded by DNA.
• The test: not "how many threads agree?" but "are they independent, and what would exclude the suspect?"

What you can honestly say on the stand

"In my opinion, the totality of the forensic evidence strongly supports that Mr. Keller killed Mr. Diallo and set the fire. I cannot, and do not, state that as a scientific certainty. Here are the specific limits of each finding, and here is the further testing that could strengthen — or could undercut — that opinion."

What you must NOT say: "The evidence proves Mr. Keller did it." "Guilty" is a legal conclusion for the jury (the ultimate-issue problem, Ch.30) — not a scientific one for the expert.

Key terms (one line each)

• Forensic case assembly — combining all evidence, each at its true strength, into the most defensible overall conclusion and its limits.
• Convergence of evidence — the agreement of multiple independent lines, powerful only to the degree they are truly independent.
• Weight of evidence — the overall strength of the assembled whole, stated on the excludes → consistent-with → strongly-supports scale (never proves).
• Exclusion matrix — a persons-×-evidence grid that narrows the field by what the science can rule out before reasoning about who remains.

The cold-case line

A homicide, staged as an accident; the evidence strongly supports that Roy Keller is responsible. It does not prove it — and the surest sign the assembly was done honestly is that it can name everything it does not establish. It is a capital mistake to theorize before one has data. We waited for the data, let it exclude what it could, and said exactly as much as it would let us say — and not one word more.

The four themes, lived (all at once)

• Exclusion over proof — four clean exclusions cleared the field; even the climax stops at "strongly supports."
• The validity spectrum — threads weighed unequally (GC-MS > cell-site; DNA > class toolmarks; the profile got zero weight).
• Cognitive bias — appears twice: it nearly closed Mill Creek on the wrong man, and it manufactured the Beatrice Six counterfeit. The whole method (independence, symmetric scrutiny, blind re-analysis) is built to resist it.
• The CSI effect — refused both the fantasy of certainty and the cynical "no smoking gun, so it's not enough." Calibration is the game.