Chapter 4 — Key Takeaways: The Forensic Laboratory

A one-page card. The lab sets the ceiling on evidence quality — and every safeguard in it is necessary and insufficient.

The core claims

  • The lab is a federation of specialized sections, not one magic room. DNA, drug chemistry, toxicology, trace, latent prints, firearms, documents, digital — each its own discipline, analysts, instruments, and backlog. Small labs refer complex work to a state crime laboratory, splitting a case across institutions and chains of custody.
  • The backlog is the defining real-world fact. Demand vastly exceeds capacity; cases queue for months; the resulting production pressure is the soil in which corner-cutting and fraud grow.
  • Accreditation certifies conformance, not correctness. ISO/IEC 17025 means an outside body verified the lab has and follows a quality system within its declared scope. It does not prove any given result is right, does not make an invalid method valid, and does not make a dishonest analyst honest.
  • QA is the system; QC is the sample. Quality assurance = proactive policies, training, audits, validated SOPs, the two-reviewer rule. Quality control = sample-by-sample checks: positive controls, negative controls, reagent blanks. A failed control invalidates the run — full stop.
  • Proficiency testing measures the analyst — best when blind. Declared tests measure performance under observation; the revealing version is blind (disguised as real casework). Easy samples and known-it's-a-test behavior are its blind spots.
  • Method validation comes before casework. Prove the method on samples of known truth — accuracy, precision, sensitivity, specificity, robustness, error rate. Developmental validation = the method can work; internal validation = this lab can run it. This is PCAST's "foundational validity" at the bench.
  • Contamination is controlled, never eliminated. Routes: scene/collection, carryover between samples, analyst's own cells, contaminated reagents. Defenses: separation + unidirectional workflow, PPE, decontamination, blanks, staff elimination databases. A blank that catches contamination is a success.
  • One unchecked analyst can taint a jurisdiction. Dookhan (dry-labbing), Farak (using the samples), Zain (fabricated, prosecution-tailored serology) corrupted tens of thousands of cases combined — all under systems that, where they existed, failed to catch it.
  • The independence problem hangs over everything. Most U.S. labs sit inside law enforcement. That structural alignment is cognitive bias scaled from the analyst to the institution (Theme 3); the fix is structural independence (Chapter 38), and most jurisdictions have not adopted it.

Method-validity verdict (this chapter's special case)

This chapter is not about one method but about the system that determines whether any method reaches its validity ceiling. The verdict:

Accreditation, QA/QC, proficiency testing, and validation are real and worth defending — but each is NECESSARY AND INSUFFICIENT. They raise the floor on reliability; none guarantees a correct result. A result's trustworthiness depends on the method's own validity (Theme 2), the controls actually run, the analyst's integrity, and the independence of the institution — not on the credential alone.

Critically: a lab can be perfectly accredited and still run a scientifically invalid method (e.g., bite marks). Accreditation does not move a method on the NAS/PCAST validity spectrum.

Themes advanced

  • Theme 2 (not all methods are equally valid): accreditation certifies conformance, not validity; a flawless quality system cannot rescue an unvalidated method.
  • Theme 3 (cognitive bias is the chief threat): context management as a quality control; the independence problem as bias at the institutional scale.
  • (Also Theme 1, in institutional form: a scandal does not "prove innocence" — it excludes tainted evidence as a basis for confidence in guilt.)

Key terms (one line each)

  • Crime laboratory — facility examining physical evidence for the legal system, organized into specialized sections; small labs refer out to state labs.
  • Accreditation — independent recognition that a lab follows a quality system and is competent in its scope; certifies conformance, not correctness or method validity.
  • ISO/IEC 17025 — the international competence standard most crime labs are accredited to.
  • QA/QC — quality assurance (the system) + quality control (the sample-by-sample check).
  • Proficiency testing — testing an analyst on known-answer samples; most revealing when blind.
  • Method validation — proving a method on known-truth samples before casework (developmental + internal).
  • Contamination — foreign material in/between samples; made detectable by blanks and controls, never impossible.

What you can honestly say on the stand

"This result was produced under an accredited quality system with documented controls, which makes it more reliable than work done without one. But accreditation certifies process, not correctness. To trust the conclusion you must also know that the underlying method is validated, that the controls and blanks were actually run and passed, that contamination was guarded against, and how the analyst was insulated from knowing the 'wanted' answer. The credential is the floor, not the proof."

What you may NOT say: "It came from an accredited lab, so it's reliable" — the error the whole chapter exists to refute.