Chapter 101: The Methods Memo
The most important sentence in the memo was not a finding; it was the rule that no finding stood alone.
On Monday, Samuel placed that sentence at the top of a public draft in KRR's review office.
The page below it described how the firm had tested records without exposing protected people.
My task was to make the explanation readable.
It was not to make the findings sound stronger.
Samuel reminded me that every suggestion would enter the report version log.
I could propose a shorter sentence.
He could reject it, restore the original, or publish neither.
The first section described cross-source corroboration.
Its original table divided support into audio, native documents, independent logs, and named witnesses. Each row also carried a column titled `What This Source Cannot Establish Alone`.
A recording could establish that a meeting occurred and preserve the words spoken there.
It could not, by itself, prove monitoring, payment, retaliation, implementation, or later harm.
A native file could show what its author recorded.
It still required a source, a version history, and another category of support.
A system log could establish delivery, access, timing, or deployment.
It could not read a person's motive.
A named witness could explain a process personally observed.
Memory alone could not replace the underlying record.
I changed `corroborative independence` to `support from a different kind of source`.
Samuel accepted the plain language and kept the technical phrase in a linked note.
I tried to shorten the recording limitation to one line.
The shorter version made the meeting audio sound like the center of the review.
Samuel returned the full boundary.
“No central conclusion depends on one recording, one document, or one witness,” he said.
I placed that sentence in bold.
The memo then explained sampling.
KRR had tested defined departments, periods, workflows, and protected cases.
It had not counted every person ever touched by a Whitmore system.
The supported pattern could not become a claim about an unknown total population.
I replaced a paragraph of statistical language with an example about four tested routes.
Samuel removed my adjective `consistent` because it could imply identical outcomes.
The records supported repeated process steps, not one universal result.
His correction made the sentence less dramatic and more accurate.
I approved the loss.
The witness section described safeguards without names, employers, addresses, or private statements.
I suggested replacing `de-identified participant cohort` with `protected witnesses`. Samuel accepted it only after the protection coordinator confirmed that the simpler phrase revealed no category that had not already been public.
The document section listed categories rather than protected contents.
The log section explained timestamps and access events without pretending either proved intent.
The recording section referred only to its authenticated limits and continuing custody.
Nothing invited a reader to treat an excerpt as the whole case.
I added headings, shortened two definitions, and moved the limitations before the examples.
Samuel controlled every factual sentence and the final publication button.
At 3:26, he compared my readable draft with the prior locked version.
One red line showed where I had moved a limitation upward. Another showed where Samuel had restored `core conclusion` after I had written the broader word `finding`. Minor observations could begin with one source; central responsibility findings could not end there.
The version log showed each change, author, reason, acceptance, and rejection.
I signed only the communications review acknowledgment.
It gave me no authority over findings, sources, redactions, or release timing.
Samuel signed the methods approval.
KRR published the memo with the rule intact: no core conclusion rested on a single recording, file, or witness.
The public could now test the method without receiving anyone's protected evidence.
My best skill had become useful only after someone else could say no to it.
Before we left, Samuel opened the next day's JISC agenda.
It contained two separate resolutions: one for WMG's corporate remediation chair and one for WCT's nonprofit remediation lead.