Blind Integrity Questionnaire¶
Test or assessment — instantiates Blinding and Expectancy Bias Reduction
A questionnaire that asks masked roles what condition they believe they encountered and why.
A blind that is never checked may be fictional. Blind Integrity Questionnaire is the instrument that finds out: after the protected judgment is complete, it asks each masked person to guess which condition they were exposed to, how confident they are, and — the part that does the real work — what cue drove the guess. It does not create or maintain any mask; it audits one after the fact by eliciting beliefs. Its defining move is pairing the forced guess with an attribution: a correct guess tells you only that the mask might have leaked, but the stated reason tells you through which channel — a side effect, a packaging tell, a difference in tone — so the finding can be acted on rather than merely alarmed over.
Example¶
A specialty roaster claims a new low-cost bean is indistinguishable in the cup from their flagship single-origin. A tasting panel blind-cups coded samples and scores them. Only afterward does the Blind Integrity Questionnaire go out: "Which cup do you believe was the premium bean? How sure are you? What made you think so?" The scores themselves look close — but eleven of fourteen tasters correctly flag the flagship, and almost all cite the same thing: "brighter, more acidic finish." That converging free-text reason is the payload. It says the two coffees are genuinely distinguishable on acidity, so any downstream "premium" preference is riding a real sensory cue, not a blind that held. The questionnaire didn't rescue the blind; it revealed exactly where and why it never closed.
How it works¶
The questionnaire is administered after the masked decision, never before, so the act of asking cannot itself plant a cue. Each masked role gets a short two-part item: a forced condition guess (with a graded confidence, not just yes/no) and a free-text or checklist attribution of the reason. Responses are aggregated into a blinding index and compared against what chance guessing would produce. The crucial interpretive step separates two very different worlds that both produce correct guesses: one where raters inferred assignment from an avoidable cue (a fixable leak) and one where the treatment effect is simply large and obvious (not a blinding failure at all). The attribution text is what tells them apart.
Tuning parameters¶
- Administration timing — immediately after each judgment versus at end of study. Later is cheaper but risks recall drift; earlier catches the belief while it is fresh.
- Guess granularity — binary guess, graded confidence, or full probability. Finer scales feed a more sensitive index but ask more of the respondent.
- Attribution depth — whether you require the why. Skipping it halves the burden and guts the diagnostic value.
- Index choice — which blinding statistic you compute; some credit "don't know" answers, some penalize confident wrong guesses differently.
- Coverage — which roles you survey (participants, providers, raters), since each can leak through a different channel.
When it helps, and when it misleads¶
Its strength is that it turns "we think the blind held" into evidence, and its attribution field points maintenance effort at the specific leak worth plugging. Its central failure mode is over-reading a correct-guess rate as a broken blind when the treatment effect is genuinely strong — the index rises because the drug works, not because the mask failed.[1] The classic misuse follows directly: treating the index as a pass/fail gate and discarding an otherwise sound study whose "blind failed" only because the effect was real. The guarding discipline is to interpret the index alongside the effect size and the stated reasons, and treat it as a diagnostic that localizes leakage — not a verdict on the study.
How it implements the components¶
blind_integrity_checkpoint— the questionnaire is the checkpoint instrument: the structured elicitation that tests whether masked roles could guess assignment and through what cue.expectancy_pathway_map— the attribution answers populate the map of which cues threaten which roles, turning a generic worry into a named leakage channel.post_blind_debrief_protocol— it is delivered as part of the end-of-exposure debrief, the sanctioned moment for asking masked people what they believed.
It does not build a believable comparator: credible_comparator_experience and sham_or_placebo_similarity_requirement are the province of Sham or Placebo Control. This page only measures whether whatever mask existed actually held.
Related¶
- Instantiates: Blinding and Expectancy Bias Reduction — supplies the integrity evidence the design needs before claiming its comparison was masked.
- Consumes: Double-Blind Trial Protocol and Single-Blind Participant Masking — it audits the blind those protocols create.
- Sibling mechanisms: Blinded Data Analysis Plan · Blinded Outcome Adjudication · Central Randomization and Masking Service · Emergency Unblinding Procedure · Masked Label Codebook · Sham or Placebo Control
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: A questionnaire that asks masked roles what condition they believe they encountered and why, making its operative form a bounded evaluation of existing evidence or work that produces a finding or disposition.
Independent corroboration: The frozen evidence defines Blind Integrity Questionnaire as 'A questionnaire that asks masked roles what condition they believe they encountered and why', so its operative form is Assessment, Review & Assurance.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Single lineage
Present-day reach: Specialized
Rationale: Clinical-trial methodology assesses whether masking held by eliciting condition guesses, confidence, and cue attribution after protected judgments are complete.
Related originating lineages:
- Medicine & Healthcare — Medicine contributes clinical monitoring, trial, adjudication, treatment, or patient-safety practice used here.
- Psychology — Psychology contributes evidence about cognition, bias, belief revision, affect, group judgment, or behavior used here.
Review resolution: Statistics is the agreed primary lineage through formal assessment of whether an experimental blind held. Clinical-trial practice and psychology materially contribute allocation-guessing and cue-attribution checks; the instrument is established rather than Encyclopedia-authored.
Review outcome: Reconciled after independent review; high confidence.
References¶
[1] Bang's Blinding Index (Bang, Ni & Wood, 2004) quantifies masked participants' guessing beyond chance, scaled so that values near zero indicate successful blinding. Its authors stress that a high index in the presence of a large treatment effect reflects the effect, not a failed mask — the reason this page insists on pairing the guess with an attributed cue. withdrawn registry ↩