B06

card_id: B06

input: An analytic question, a set of alternative hypotheses, and available evidence/arguments/assumptions bearing on each hypothesis.

operation: An eight-step procedure: (1) brainstorm the full hypothesis set with an analytically diverse group; (2) list evidence/arguments (including assumptions) for and against each hypothesis; (3) build a hypothesis-by-evidence matrix and assess each item's diagnosticity; (4) refine the matrix by removing non-diagnostic items; (5) draw tentative conclusions by trying to disprove rather than prove each hypothesis; (6) test sensitivity of the conclusion to key evidence items; (7) report the relative likelihood of all hypotheses, not just the most likely; (8) identify future milestones/indicators to monitor. DA02 independently operationalizes a related indicators/signposts technique: define observable indicators (and "trigger mechanisms") per hypothesis/scenario and periodically review which have changed.

output: A ranked comparative judgment across all considered hypotheses, an evidence matrix documenting the reasoning, and a forward-looking indicator list to monitor.

scope_conditions: Diagnostic/explanatory analytic problems with multiple plausible alternatives; DA02 states the technique is most effective with a small team working from an already-assembled evidence base, achievable "in a day or less."

failure_conditions: Early rejection of an unproven-but-not-disproved hypothesis biases the rest of the analysis (deception hypotheses are the canonical failure case); more than ~7 hypotheses becomes unmanageable; picking the first satisfactory hypothesis without full comparison ("satisficing") is the failure mode the procedure is designed to counter.

predicted_effect: Helps analysts avoid premature closure on a single favored hypothesis and increases the chance of surfacing disconfirming evidence; stated as design intent by the original author, not a measured effect size.

measurement: No numeric scoring rule verified; diagnosticity assessment and disproof-seeking are qualitative judgment procedures, not calibrated probabilities.

implementation_test: Given hypotheses and evidence, construct the matrix, assess diagnosticity, and check whether the written report ranks all hypotheses. DA02 gives a directly testable analogue: build an indicators matrix and record "is this occurring or not?" per indicator over time.

decision_rights: An analyst or small team constructs and updates the analysis; policy/decision authority over the resulting judgment sits outside the technique in both sources.

uncertainty_handling: Maintains multiple alternatives simultaneously rather than collapsing to one early; seeks disconfirming evidence; no calibrated probability rule verified; DA02 adds indicator-based prospective monitoring to detect divergence from an expected course.

verification: Systematic, team-based re-examination of evidence for diagnosticity and disconfirmation (DA03); periodic indicator review against an explicit matrix (DA02).

provenance: DA03 Chapter 8, pp.95-98; DA02 pp.12-16.

implementation_cost: DA02 states the matrix-construction phase can take "a day or less" once data is already collected, more if data must be reassembled; DA03 frames the full procedure as reserved for "an especially important, difficult, or controversial judgment," implying it is not for routine decisions.

limitations: No controlled-outcome study of ACH's effectiveness was located or read; DA03 is the technique's own inventor's account.

statement_basis: direct: ["8-step ACH procedure box (DA03 pp.96-97)", "footnote 85 attributing ACH's development to the author", "DA02 indicators/trigger-mechanism content and cost statement"] inference: ["predicted_effect summary phrasing is curator paraphrase"]