Annotated Decision Record¶
Documentary artifact — instantiates Cognitive Apprenticeship Modeling
A written record of a real decision that captures not just what was chosen but the cues, evidence, rejected options, and residual doubt behind it — so the reasoning survives the moment.
An Annotated Decision Record is a durable written artifact that captures a decision together with the reasoning that produced it — the cues that were read, the evidence weighed, the options rejected, and the uncertainty that remained. Where most records preserve only the outcome (what was decided), the annotation is the whole point: it preserves the why, in the decision-maker's own words, at the time. Its defining property is persistence — it turns a moment of expert cognition, otherwise gone the instant the meeting ends, into an inspectable object a novice can read next month or next year. It is the memory layer of the apprenticeship: not a live demonstration and not a coached attempt, but the written residue that lets reasoning outlast the moment that produced it.
Example¶
Before taking a position, a portfolio manager writes a short memo: the thesis, the three data points that most moved her, the two disconfirming facts she chose to discount and why, the single scenario that would prove her wrong, and her confidence (~60%). She files it before the outcome is known. A junior analyst reading it six months later — after the position has half-worked — sees something no performance chart shows: which cues she weighted, which she deliberately set aside, and exactly where she flagged her own doubt.
The analyst learns the shape of a real judgment call, including the parts that don't survive in the return number. And because the doubt was recorded in advance, the team can later ask the sharp question — was the reasoning sound and the world unkind, or was the reasoning itself flawed? — instead of rewriting history around whatever happened.
How it works¶
- Record the reasoning, not just the ruling. Capture the cues read, evidence weighed, and criteria applied — the inputs to the choice, not only its output.
- Preserve the rejected branches. Name the options considered and why each was set aside; the discards carry much of the judgment.
- Timestamp the uncertainty. State confidence and the disconfirming facts before the outcome lands, so the record can't be quietly rewritten later.
- Keep it terse and legible. A record no one reads teaches no one; the annotation is a paragraph, not a dossier.
Tuning parameters¶
- Annotation depth — from a one-line rationale to a full reasoning trace. Deeper records teach more but cost time and discourage honest, frequent writing.
- Ex-ante vs. ex-post — whether it's written before or after the outcome. Writing before is far more instructive and far more honest, but harder to enforce.
- Confidence granularity — a rough hunch, a percentage, or a full scenario set. Finer confidence makes the record more testable but risks false precision.
- Audience — a private decision journal versus a shared team artifact. Shared records teach more widely but invite the temptation to write for the audience rather than for the truth.
When it helps, and when it misleads¶
Its strength is persistence: it makes expert reasoning survive the moment and become teachable long after, and it enables honest hindsight by fixing the ex-ante rationale in place. Kept before the outcome, it is essentially a decision journal — the standard corrective to memory's habit of rewriting past reasoning to match present results.[n1]
Its failure modes are documentary. A record written after the outcome quietly launders hindsight into the account, so the reasoning looks better than it was; and records demanded for compliance decay into documentation theater — boxes ticked with boilerplate that no learner ever reads. The classic misuse is the retrospective record assembled to justify a decision already made and already known to have worked. The discipline that keeps it honest is to write the annotation before the result is known, keep it short enough to be read, and treat it as a genuine bet on record rather than a defense exhibit.
How it implements the components¶
visible_reasoning— it externalizes the cues, evidence, and rejected options behind a decision into a written, inspectable form.apprenticeship_record— it is the durable record: the persistent memory that lets a novice study real expert reasoning long after the decision was made.
It preserves reasoning but does not perform it live (narrating cognition as it happens is Think-Aloud Demonstration's role) or coach a learner's attempt (coaching_feedback — Coached Practice Session); and turning the recorded near-misses into shared lessons is the Case Conference's job (error_modeling), not the record's.
Related¶
- Instantiates: Cognitive Apprenticeship Modeling — the record is the artifact that makes expert reasoning persist and become studyable.
- Sibling mechanisms: Cognitive Walkthrough · Case Conference · Worked Example with Reasoning · Think-Aloud Demonstration · Shadowing with Debrief
Editorial Notes¶
Form Classification¶
Form family: Record, Log & Register
Rationale: A written record of a real decision that captures not just what was chosen but the cues, evidence, rejected options, and residual doubt behind it — so the reasoning survives the moment, making its operative form a durable account of actual events, decisions, or transitions whose value depends on history or provenance.
Independent corroboration: The frozen evidence defines Annotated Decision Record as 'A written record of a real decision that captures not just what was chosen but the cues, evidence, rejected options, and residual doubt behind it — so the reasoning survives the moment', so its operative form is Record, Log & Register.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Education & Pedagogy
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Cognitive apprenticeship in education provides the central aim: externalizing expert cues, choices, and uncertainty so novices can inspect otherwise tacit reasoning.
Related originating lineages:
- Computer Science & Software Engineering — Architecture decision records contribute a concise, versioned decision-and-rationale artifact.
- Ethnography & Qualitative Methods — Analytic memoing separates evidence from interpretation.
- Organizational & Management Science — Decision journals and after-action learning provide the durable record practice.
- Psychology — Hindsight-bias research motivates ex-ante recording of confidence and rejected options.
Review resolution: Cognitive-apprenticeship pedagogy is primary. Architecture decision records, analytic memoing, organizational decision logs, and hindsight-bias research materially shape the artifact; exposing expert cues and uncertainty in one record is an Encyclopedia synthesis.
Attribution caveat: The exact artifact deliberately fuses cognitive apprenticeship, decision journaling, and software decision records.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
An Annotated Decision Record is raw material, not instruction on its own — it teaches only when a walkthrough, a conference, or a coach puts it in front of a learner and unpacks it. Filed and unread, it is documentation, not apprenticeship.
[n1] A decision journal records the rationale, expectations, and confidence behind a choice at the moment it is made, so later review can compare reasoning to outcome without the distortion of hindsight. It is the disciplined, ex-ante form of an Annotated Decision Record. ↩