Affect–Evidence Split Prompt¶
Checklist — instantiates Relevance-Substitution Detection and Correction
Separates the emotional response from the evidence and asks what the judgment would be if the affective cue were removed.
Affect–Evidence Split Prompt is a short, self-administered checklist that forces one thing into the open: the feeling a case produces and the evidence it offers are two different signals, and only one of them should move the judgment. The prompt makes you write the affective reaction in one column and the evidence in another, name how the feeling is gripping you, and then run a single counterfactual — if the emotion were gone, would my conclusion change? If the judgment survives the removal of the feeling, the affect was decoration; if it collapses, the affect was doing load-bearing work it may not have earned. The one move that makes this mechanism itself, and not one of its siblings, is that imagined removal: it isolates a single affective cue and tests whether the decision is independent of it, rather than cataloguing every irrelevant move in a debate or masking cues at the source.
Example¶
A product manager at a consumer-software company opens the day to a viral tweet — a furious, widely-shared thread accusing a recently shipped feature of "ruining" the app, screenshots and all. Her stomach drops; her first instinct is to roll the feature back before lunch. Before acting, she runs the Split Prompt. In the affect column she writes: "public embarrassment, a named person angry at me, fear the thread keeps growing." In the evidence column she writes what she actually knows: retention on the feature is flat-to-slightly-up, support tickets are unchanged, and the thread is one loud account plus quote-tweets. Then the counterfactual: if this had come in as an anonymous line in a metrics report instead of a viral thread, would I roll back? The answer is plainly no. The feeling was real and the tweet was worth a reply, but it was not evidence that the feature was failing. She keeps the feature, opens a targeted survey to check the complaint on its merits, and writes one sentence noting that the tweet's force and its informativeness were not the same thing.
How it works¶
- Split before you decide. The reaction is logged in its own column so it is neither suppressed nor smuggled into the evidence. Naming the feeling is what keeps it from operating anonymously.
- Name the grip. State the specific channel the affect uses — vividness, identity threat, disgust, a sympathetic face, fear of public exposure. A cue you can name is a cue you can discount.
- Run the removal counterfactual. Re-imagine the same case stripped of the charged presentation (the anecdote as a statistic, the sympathetic plaintiff as a case number) and ask whether the judgment holds. This is the mechanism's core test.
- Record the verdict, not just the feeling. Note whether the judgment was independent of the affect; if it was not, the case goes back for evidence before any decision is made.
Tuning parameters¶
- Removal strictness — how thoroughly you strip the affective framing in the counterfactual. Strict removal catches more substitution but risks discarding cases where the emotion carries genuine information; loose removal is faster but leaks affect back in.
- Affect-naming granularity — one word ("upset") versus a specific pathway ("fear of being publicly blamed"). Finer naming discounts more precisely but takes longer and can feel clinical.
- Second-person check — self-run versus handed to a colleague who does not share the feeling. Outsourcing the counterfactual is more honest but costs a person's time.
- Trigger threshold — which decisions get the prompt at all. Applying it to every call is exhausting; reserving it for high-stakes or high-heat moments risks missing quiet emotional pulls.
When it helps, and when it misleads¶
Its strength is speed and honesty at the exact moment judgment is hottest: it interrupts the affect heuristic, the well-documented tendency to read a strong feeling about something as information about its facts.[n1] By separating the two channels and testing independence, it lets a decision-maker feel the feeling and still decide on the evidence.
Its failure mode is treating all affect as noise. Sometimes the emotion is the evidence — a clinician's unease, a safety engineer's dread, a grieving family's testimony — and a mechanically applied removal test will strip out a legitimately relevant signal and call it bias. The classic misuse is weaponized detachment: using "let's separate emotion from evidence" to dismiss an inconvenient but relevant emotional truth, especially someone else's. The guarding discipline is to distinguish affect-as-cue (a feeling standing in for evidence it doesn't have) from affect-as-evidence (a feeling that is itself informative), and to run the removal test only against the former.
How it implements the components¶
substitute_signal_register— the affect column is a register of the psychologically active cue that is standing in for evidence.psychological_activation_pathway— "name the grip" records how the cue exerts its force (vividness, identity threat, exposure fear), not just that it does.channel_independence_check— the removal counterfactual is exactly a test of whether the judgment is independent of the affect channel or hostage to it.
It does not fix the claim in a question_at_issue_record or strike diverting moves via cue_quarantine_or_label — that live issue-shift screening belongs to Red Herring Filter Checklist, its checklist twin; nor does it produce the numeric before/after delta of a correction_and_reweighting_rule, which is Reweighted Update Log's job.
Related¶
- Instantiates: Relevance-Substitution Detection and Correction — supplies the affect-isolation front end that flags when a feeling, not the evidence, is driving the call.
- Sibling mechanisms: Red Herring Filter Checklist · Analogy Mapping Table · Question–Evidence Matrix · Decision-Basis Disclosure · Reweighted Update Log · Cue-Validity Audit · Proxy-Relevance Audit · Blinded or Masked Review
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: The mechanism separates affect from evidence, runs a cue-removal counterfactual, and records whether the judgment survives independently, so its operative form is a bounded assessment of evidentiary relevance.
Nearest alternative: Interface, Display & Cue — A prompt presents the questions, but the mechanism's defining result is a finding about whether affect contaminated the judgment rather than the user-facing affordance itself.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Psychology
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Universal
Rationale: Separating an affective response from evidentiary judgment and testing the judgment without the cue directly addresses the affect heuristic and emotion-driven reasoning studied in psychology.
Related originating lineages:
- Behavioral Economics — Affect substitution and bias in risk and value judgments contribute the decision context.
- Philosophy — The distinction between evidential warrant and non-evidential influence supplies the epistemic norm.
- Statistics & Experimental Design — Blinding and cue-ablation logic supply an empirical analogue for asking whether a conclusion survives removal of a potentially biasing channel.
Review resolution: Psychology supplies the affect-heuristic lineage. Behavioral economics, epistemic reasoning, and cue-ablation or blinding logic materially shape this newly synthesized checklist, whose use is domain-general.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] The affect heuristic (Slovic, Finucane, Peters, and MacGregor) is the tendency to let an overall good/bad feeling about something substitute for analysis of its actual risks and benefits — the precise substitution this prompt's removal test is built to catch. ↩