VELLORA RESEARCH EVIDENCE LOG — 2026-10-11
The CSV is blank, not participant data. Duplicate a row for each observation.
Use participant codes; keep the identity key in a separate, restricted system.
source_type: real interview / observed task / survey response / synthetic output / other.
source_reference: stable reference to permissioned material, not an open share link by default.
time_or_location: timestamp, paragraph or record ID that another authorized reviewer can find.
exact_words_or_observed_action: preserve exact text if quoting; identify a paraphrase explicitly.
observation_or_quote: label the record type. Keep interpretations in their own column.
assistance_given: record help and whether the answer followed an explanation or replay.
claim_status: unreviewed / supported within stated scope / needs more evidence / contradicted.
reuse_permission: note actual allowed purpose. Research consent does not imply marketing permission.
Never count synthetic output as another real participant.
Keep disconfirming examples; one observation does not establish prevalence.
Before importing external text into a spreadsheet, set columns to plain text; strings beginning with =, +, -, or @ can be treated as formulas by spreadsheet software. This blank CSV contains no imported user content.
Keep access, retention and deletion aligned with the consent given.
Review the original source before moving a proposed claim into a public product story.
https://vellorastudies.com/templates/research-evidence-log/