You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
PR #618: the 17-Clostridia snippet silently drops (P < 0.05) from mid-quote #620
Adversarial review of my own PR #618, before merge. Minor, same family as #619.
On the primary_fermenter_community -> butyrate edge, the snippet for DOI:10.1016/j.chom.2016.03.004 reads:
inoculation with a community of 17 human Clostridia isolates resulted in a significant
increase in butyrate levels
The source (PMC4832419) reads:
While inoculation with a community of 17 human Clostridia isolates resulted in a
significant (P < 0.05) increase in butyrate levels
(P < 0.05) was dropped from the middle. Truncating a quote at its start is normal
excerpting; removing words from the interior without a marker is an elision, and the
playbook forbids it. Dropping the statistic is also the specific kind of elision that
matters, since "a significant increase" without its test reads as an editorial
characterisation rather than a reported result.
audit_evidence_snippets.py flags ELLIPTICAL_SNIPPET only on a literal ... or …, so
an unmarked interior deletion is invisible to it. Same structural blind spot as #619.
Fix
Use the contiguous text, either from While inoculation or from inoculation, retaining (P < 0.05).
Acceptance test
The snippet appears character-for-character in the source. Blocks merge of #618.
Activity
Correct two bad snippets and one overstated note found reviewing this PR