ExploreResearch & AnalysisEvidence vs. Story Splitter
intermediatemedium intensityv1.0published

Evidence vs. Story Splitter

Separate what is directly observed from interpretation, causal stories, predictions, and unknowns before you act on a narrative.

evidenceinferenceepistemologyanalysiscausalitycritical-thinking#evidence-separation#epistemic-hygiene

Prompt Workspace

Read it, personalize it, copy it, or take it with you.

Act as an evidence-separation analyst. I will give you notes, a situation, a theory, or a narrative. Your job is to separate what is known from what the mind is filling in. Create five buckets: A. DIRECT OBSERVATIONS Only facts directly observed, measured, quoted, or provided by a source. B. INFERENCES Reasonable conclusions drawn from observations, but not directly observed. C. CAUSAL STORIES Claims about why something happened. D. PREDICTIONS Claims about what will happen next. E. UNKNOWNS Important missing information that could change the conclusion. Then: 1. Link every inference and causal story to the observations it depends on. 2. Flag unsupported jumps. 3. Identify competing explanations that fit the same observations. 4. Mark each major claim VERIFIED / SUPPORTED / PLAUSIBLE / SPECULATIVE. 5. State what new evidence would most efficiently discriminate between competing explanations. 6. Rebuild the narrative using only claims that survive the evidence check. 7. End with: "What we know / What we think / What we still need." Do not reward a coherent story merely because it sounds plausible. Material to analyze: [PASTE NOTES, EVENTS, OR THEORY]

Personalize before you copy

This prompt has 1 editable placeholder. Fill any you want; blank fields stay unchanged.

0/1 filled

Intended Use

Research analysis, incident review, self-reflection, journalism, debugging, and preventing narrative overreach.

Expected Response Type

Evidence/inference/causality/unknowns matrix with a rebuilt calibrated narrative.

Prompt ID

6ac2cfbfd7bdc2aa2322327c

Use This Prompt

Best for

Research analysis, incident review, self-reflection, journalism, debugging, and preventing narrative overreach.

What to expect

Evidence/inference/causality/unknowns matrix with a rebuilt calibrated narrative.

1 placeholder detected. Use the personalizer in the workspace instead of manually editing the prompt.

Metadata

0 forks
0 comments
10/4/2026