Software observation · CSV · v1.0 ·
Survey checker: 12 recorded test cases
Exact inputs, observed signal labels, rule codes, interpretations and run date for rules version 1.0. A selected demonstration, not an accuracy benchmark.
Download CSVWorksheets & test files
Take the method into your own work.
Download editable survey questions, planning worksheets, worked examples and our recorded tool checks. Every file explains whether it is a blank template, a constructed example or an observed software output.
Reproduce an example
Follow the scale-scoring walkthrough or repeat the twelve-case checker test. The scoring counts are invented for teaching; the checker alerts are actual software outputs from constructed questions.
Constructed example · CSV · v1.0 ·
Survey scale: worked scoring example
Reproduce 40% favorable among all 20 and 50% among 16 scale answers, with separate cannot-judge coding. All counts are synthetic.
Download CSVPlanning templates and schemas
Choose a blank instrument or adapt the fictional question matrix. Check each field against the actual study before use.
Template · CSV · v1.0 ·
1–5 survey scale template
Three editable agreement, clarity and ease items with complete response labels and coding notes. Not a validated scale or collected responses.
Download CSVTemplate · CSV · v1.0
Aggregate result schema
Empty release structure for condition estimates, uncertainty, suppression, and notes.
Download CSVTemplate · CSV · v1.0
Aggregate-release codebook
Definitions and allowed values for the public result schema.
Download CSVTemplate · CSV · v1.0
Disclosure-test worksheet
Plan exact stimuli, unaided interpretation, recognition, trust, and reporting checks.
Download CSVTemplate · CSV · v1.0
Message-review worksheet
Record independent 0–2 rubric judgments, evidence, blockers, and revisions.
Download CSVTemplate · CSV · v1.1 ·
Question-order audit
Put consent and required safety information first, then audit eligibility, exposure, unaided outcomes, and priming risks.
Download CSVTemplate · CSV · v1.1 ·
Claim register
Map exact and implied claims to evidence scope, owners, decisions, and review triggers.
Download CSVTemplate · CSV · v1.0
Cross-border change log
Separate language, currency, offer, disclosure, and operational differences.
Download CSVTemplate · CSV · v1.0
Sampling inference audit
Document the target, frame, recruitment, error risks, and permitted reporting lane.
Download CSVTemplate · CSV · v1.0
Decision brief
Freeze the decision, actions, contrast, outcomes, thresholds, and stop conditions.
Download CSVTemplate · CSV · v1.0
Reporting checklist
Audit provenance, flow, measurement, estimates, limits, disclosures, and corrections.
Download CSVTemplate · CSV · v1.0 ·
Concept-testing question matrix
Adapt ten fictional example questions with response options, decision uses, interpretation risks, and blank revision fields. Not a validated questionnaire.
Download CSVFictional decision examples
These two filled files accompany the worked decision-brief guide. They demonstrate a method using constructed inputs, not findings from a real study.
Fictional example · CSV · v1.0 ·
Filled fictional decision brief
Follow a constructed message-comparison plan with explicit adoption rules, guardrails, and limits. Approval and collection dates are intentionally blank.
Download CSVFictional example · CSV · v1.0 ·
Annotated fictional decision results
Reproduce planning arithmetic, a +8-point comparison, uncertainty, and a decision not to adopt. Every count and result is synthetic; no fieldwork occurred.
Download CSVTemplate scope and license
The collection includes blank planning structures, a question matrix with fictional wording, and filled fictional examples. These are not validated instruments or industry standards. You may adapt them for internal or commercial work with attribution to Message Test Bench. Do not imply endorsement, certification, or that a schema or teaching example contains empirical evidence.
Release standard
Future public research releases will use aggregate rows only. Names, emails, device identifiers, panel identifiers, verbatim personal narratives, and row-level respondent files do not belong in this catalog. Small cells will be suppressed; additional combinations may be withheld when they could make a person or household reasonably identifiable.
Versioning and provenance
Every future empirical dataset will include a release date, version, study identifier, codebook, methods link, source and transformation notes, license, and corrections history. A material correction produces a new version while the prior change remains documented.