Historical model
One leak, one visible weather system
- CarrierReporter, column, speech, broadcast
- AudienceBroad and mostly shared
- SignalEditorials, calls, polls, elite response
- Retreat“No decision has been made.”
An interactive documentary
Who launched the story?
A policy appears in the press before anyone owns it. Your source calls it “serious.” Your editor calls it news. The public reaction has already become data.
Chapter 01 / The briefing
A leak can inform. It can also measure.
A trial balloon is a proposal released tentatively—often through the media—to observe the reaction before its sponsor commits.
The technique exploits a newsroom’s legitimate appetite for consequential information. If the idea rises, its sponsor may claim momentum. If it falls, the source can deny that a decision was ever made.
Working definition
Trial balloon: a strategic, reversible disclosure used to test political, institutional, market, or public response. The defining feature is not anonymity alone; it is the feedback loop.
Chapter 02 / Then and now
From tomorrow’s front page to everyone’s private feed.
THE SAME INSTINCT / A DIFFERENT MACHINE
Platform model
Many variants, invisible weather systems
- CarrierAds, posts, influencers, segmented feeds
- AudienceDivided into targetable groups
- SignalClicks, shares, sentiment, conversions
- IterationThe winner scales; the loser disappears
Social testing does not replace the classic trial balloon. It industrializes the same question: what can we say, to whom, with how much deniability, before we decide what we believe?
Chapter 03 / The assignment
The envelope arrives at 9:14 p.m.
Decision 1 of 4
“The government is considering a national press credential.”
A senior adviser messages you from a personal account. They insist the proposal is real, urgent, and “already moving.” They will not be named.
Source claim
“Publish tonight. Reaction by morning will determine whether cabinet sees the draft.”
Chapter 04 / The compact
Anonymity protects sources. It should not conceal method.
The ethical question is not simply “may this person be unnamed?” It is “does the public receive enough information to judge the source’s knowledge, motive, and risk?”
0 of 6 safeguards checked
Establish direct knowledge
Ask exactly how the source knows. Separate firsthand information from inference, hearsay, advocacy, and speculation.
Interrogate motive
Why this information, why now, and why you? A motive does not invalidate evidence, but readers and editors need its relevant contours.
Seek independent corroboration
Use documents, on-record sources, affected parties, and additional independent sources. Do not let confidentiality prohibit further reporting.
Use anonymity as a last resort
Reserve it for vital information unavailable another way, especially when identification risks danger, retribution, or other serious harm.
Give the editor the identity
A senior editor should know who the source is and approve the arrangement. “Anonymous to the public” should not mean “unknown to the newsroom.”
Explain the arrangement
Describe the source as specifically as safety allows, state why anonymity was granted, and disclose relevant motive or vested interest.
The enduring vulnerability
“If the newsroom becomes the instrument that measures reaction, is it still only reporting the experiment?”
Agenda-setting research distinguishes influence over what people think from influence over what they think about. A trial balloon gains leverage the moment editorial attention turns a tentative possibility into a shared public object.
Research desk
Sources & further reading
- McCombs & Shaw, “The Agenda-Setting Function of Mass Media” (1972)
- David E. Pozen, “The Role of News Leaks in Governance and the Law of Journalists’ Confidentiality”
- Associated Press, “Telling the Story” — anonymous-source standards
- Society of Professional Journalists, Code of Ethics
- Canadian Association of Journalists, Protection of Sources
- Tappin et al., political message pretesting and microtargeting
- Tactical Tech, A/B testing in campaign messaging
- Associated Press, “When is it OK to use anonymous sources?”