Harm Tiers and Proportionate Evidence Gates
7 min read
Lesson objectives
- Tier an editorial claim by plausible harm, reach, and reversibility without false numerical precision.
- Convert the tier into a concrete evidence condition, wording, alert choice, and review route.
Please sign in to use all the tools
Once you sign in, the reading tools are yours: larger or smaller text, typeface, line spacing, focus mode, highlighting and bold, and notes saved to your account and gathered in My study.
Published claims do not all carry the same consequences. A wrong museum opening time may inconvenience a visitor; a false warning that neighborhood water is unsafe may change the behavior of thousands. Calling a story “sensitive” leaves the desk guessing what to do. A usable risk tier links plausible harm to an evidence gate, escalation speed, and a correction route. The three-tier scheme in this lesson is a fictional newsroom design aid, not a legal standard or a result of empirical calibration.
Consider three dimensions before counting views. First is plausible harm if the claim is wrong: a small inconvenience, a disrupted service decision, or a safety or health consequence. Second is expected reach: a quiet service page, a widely distributed push alert, or a message likely to be forwarded. Third is reversibility: can an error be corrected before people act, or will screenshots and decisions persist after correction? These are qualitative judgments, not numbers to multiply into false precision. A potentially severe harm may justify a high tier even if initial reach is modest, because the item may travel.
Define three operating tiers for this imagined desk. Low: a limited service fact that can be corrected with little downstream impact; verify the original source and check headline against the source before release. Medium: a claim that could change substantial audience planning or a person's reputation; establish a traceable origin, independently check scope where feasible, and have a named editor examine the headline and alert. High: a health or safety warning, grave allegation, or private detail whose effect may be difficult to reverse; require stronger support, another authorized editorial review where available, a check on harm from publication itself, and a clear hold path. A newsroom must adapt its policy to its work and authority, but should not improvise a weaker tier during one event merely to meet a deadline.
Do not assign one tier to an entire story before examining its assertions. A report about a road closure might contain a relatively low-impact opening time for a car park, a high-impact “water contamination” claim, and a middle-tier estimate of affected passengers. Give each material claim a tier, evidence condition, and allowable wording; then assess the final package because a headline or visual may convey its strongest implication. Removing a high-risk sentence from the article while leaving it in the notification does not reduce publication harm. Tiers help the newsroom release a narrow supported fact while continuing to investigate a more consequential assertion.
Speed still has value. A traveler may need to know about the eastern entrance before arriving, but urgency does not justify dropping “eastern” or “until noon.” Draft a short bounded alert: “Operator: Eastern bridge entrance closed until noon.” If only a screenshot of the announcement exists, a reporter's nearby picture does not automatically authenticate the stated reason or duration. The photo may support congestion, not the operator's schedule. Specify the claim each source actually supports and escalate the remaining gap. A desk can be useful without making its fastest wording its widest wording.
Use a compact decision table:
- claim
- harm if wrong
- expected reach
- reversibility
- tier
- evidence condition
- current release state
- review owner
“High” alone is insufficient. For example: “An anonymous audio warning may drive widespread health behavior, but identity and test result are unverified; no health alert until traceable support.” This explanation enables another editor to review the decision. A later official notice changes the available evidence but does not itself waive text matching. Check its date, area, wording, and whether it reports preliminary findings or recommends an action. A verified source still has a specific scope.
A tier does not predict that harm will occur. It allocates checking effort under uncertainty. Write “readers might change health behavior after a false warning” rather than “harm is certain.” A bridge closure may be medium tier when the error is quickly reversible, yet require higher review if an erroneous direction could divert emergency vehicles. Explain which consequence raised the tier. Keep likelihood of a false claim separate from the severity of its outcome: a high-harm item may have strong evidence, while a low-impact notice may be poorly sourced. The release decision considers both.
Channel choice also changes exposure. A detailed, revisable update page can carry the qualifier that one entrance is closed. A phone alert may arrive without any of the article's context and may circulate as a screenshot. Review the exact text each audience will see. If the short channel cannot preserve “eastern entrance” and “until noon,” do not send a stripped version implying total closure. Use a channel that can carry the bounded meaning or wait for an approved concise form.
Avoid a rigid “two sources for high tier” rule. Two accounts may relay one original notice, while an original operating record and a field observation test different possible errors. Ask what each item can disconfirm. An operator notice supports the operator's stated plan; a reporter observes what happened at one place; a traceable laboratory result may bear on a water-quality claim. Each can still leave gaps in area or time. A useful gate identifies the exact proposition, the plausible alternative that would change the judgment, and the editor authorized to release the warning.
After publication, review the tier decision based on information available at the time, not merely on whether the claim eventually proved true. A newsroom that sent a high-impact warning from weak evidence has not acquired a sound process because luck later vindicated the wording. Conversely, holding a consequential but initially unsubstantiated claim can be responsible even when later evidence confirms it. Keep the timestamped record of what was knowable at each decision point. This distinguishes process quality from hindsight and helps the desk revise a gate that was repeatedly too weak or unnecessarily restrictive. Also test whether the chosen threshold can be performed on a real shift. “Obtain independent confirmation” is ambiguous unless the desk names what independence would look like and who can seek it. For a bridge entrance, a current operator page and a reporter at the entrance may answer different questions; for a school evacuation, a current original notice might establish the action while leaving the cause unknown. Make the gate a task with an owner and a possible outcome, including “not available before deadline.” A high tier can lead to a narrow attributed item or an explicit hold. It must never silently turn uncertainty into a categorical alert.
Lesson summary
The lesson tiers each claim and shows how error consequences and delivery channel shape its evidence gate and alert wording.