Three-pass pattern spotting: (1) mark recurrences without interpretation, (2) cluster into pattern types, (3) check against counterexamples before naming

Why This Is a Rule

Pattern recognition in personal data (journal entries, review notes, reflection logs) is vulnerable to confirmation bias: you see the patterns you expect to see and miss the ones that don't match your narrative. "I always procrastinate under pressure" feels like a pattern, but when you actually check, you procrastinate under some types of pressure and perform brilliantly under others. The "pattern" was a partial truth mistaken for a universal one.

The three-pass method — borrowed from qualitative research coding — prevents premature pattern naming by separating observation, clustering, and validation into distinct phases. Pass 1 (open coding): mark anything that appears more than once. Don't interpret — just flag recurrences. "Mentioned feeling drained" appears 4 times. "Skipped gym" appears 3 times. Pure observation. Pass 2 (axial coding): group the flagged items into pattern types — emotional (recurring feelings), behavioral (recurring actions), situational (recurring contexts), outcome (recurring results), avoidance (things consistently not done). "Feeling drained" clusters with "worked past 7pm" (situational + emotional pattern). Pass 3 (validation): for each named pattern, actively search for counterexamples. "I feel drained when I work past 7pm" — are there entries where I worked past 7pm and felt fine? The counterexample check prevents over-generalization.

When This Fires

• During monthly or quarterly reviews when analyzing accumulated reflection data
• When you want to identify recurring patterns in your behavior, emotions, or outcomes
• When Read 4-8 weeks of entries in a single sitting for pattern spotting — parallel access to temporally separated experiences reveals what sequential reading misses's batch reading has produced raw material for pattern analysis
• Complements Annual review phase 1: scan calendar month-by-month for 90 minutes, marking + (peak positive) and - (peak negative) before any analysis (annual calendar scan) and Read 4-8 weeks of entries in a single sitting for pattern spotting — parallel access to temporally separated experiences reveals what sequential reading misses (batch reading) with the analytical method for extracting patterns

Common Failure Mode

One-pass pattern naming: reading through entries and immediately declaring "I see a pattern of burnout!" without checking whether burnout is actually a pattern (consistent across many entries) or a salient recent experience (vivid in one entry, projected onto others). Premature naming produces false patterns that misinform future planning.

The Protocol

(1) Pass 1 — Open coding: Read through your entries. Mark (highlight, tag, or note) anything that appears more than once across entries. Do not interpret or name patterns. Just flag recurrences. (2) Pass 2 — Clustering: Group your flagged recurrences into pattern types: Emotional (what feelings recur?), Behavioral (what actions recur?), Situational (what contexts recur?), Outcome (what results recur?), Avoidance (what's consistently absent?). (3) Pass 3 — Counterexample check: For each proposed pattern, search for entries that contradict it. "I always feel drained on meeting-heavy days" — are there meeting-heavy days where I felt energized? If counterexamples exist, refine the pattern: "I feel drained on meeting-heavy days when meetings lack clear agendas." (4) Only name the pattern after it survives counterexample checking. (5) Each named pattern should be specific enough to suggest an intervention and robust enough to survive counterexample scrutiny.