Safety management systems depend on people writing things down: hazard reports, risk assessments, investigation notes, corrective actions. The quality of that writing decides whether the system works. Here are the questions safety teams most often raise about peer review in safety management, answered in order, with a focus on why a second reader improves hazard reports and what a practical pairing routine looks like.
Why should a second person read a hazard report at all?
Because the author is too close to it. The person who witnessed an event, or who drafted the assessment, holds a mental picture that fills in whatever the page leaves out. A reader without that picture encounters the report as the organisation will: cold, with only the words to go on. If the words are ambiguous, the second reader finds out in seconds. If an assumption is unstated, they ask about it.
The same logic applies to numbers. A likelihood rating typed in the wrong column or a date that does not match the event log looks fine to the person who entered it, and wrong to anyone comparing it with the log.
What exactly is the second reader checking?
Four things, in rough order of importance. First, completeness: does the report answer who, what, where, when and under what conditions? Second, consistency: do the dates, locations, equipment references and names agree across the report and its attachments? Third, reasoning: does the risk rating follow from what the report describes, or does the conclusion float free of the evidence? Fourth, actions: is each proposed action specific, owned and dated, or is it a general intention?
The reviewer does not have to be a senior safety specialist. What counts is attention and a copy of the source material.
Does a second reader slow the safety process down?
Marginally, and the trade is usually favourable. A short review before filing takes minutes. A report returned from a committee with questions can take days to resolve, and a report that misleads a later reader can cost far more. Teams that adopt pairing commonly report, anecdotally, that follow-up questions drop because the obvious ones were asked in advance. Treat that as a plausible benefit, not a promise; each organisation should observe its own experience.
Who makes a good partner?
Someone who did not witness the event and did not draft the assessment, and who is comfortable asking plain questions. Rotating partners helps: a colleague from a different department often asks the questions insiders no longer think to ask. Avoid always pairing a junior with a senior if you can, since juniors may feel unable to challenge, while seniors may skim.
What should the author hand over?
The draft and the evidence: photographs, logs, relevant procedure references, any data the conclusion rests on. A reviewer who sees only the narrative can proofread it. A reviewer who also sees the evidence can test it. That is the difference between a second pair of eyes and a second opinion on style.
How should feedback be given?
As questions and observations about the page, not about the person. “The report says the fuel truck was on the left, the photo shows the right, which is correct?” is useful. “You should be more careful” is not. Safety culture relies on people being willing to report; a review style that makes authors defensive undermines the very reporting the system needs.
Authors, for their part, should treat a query as information. If one reader stumbled on a sentence, later readers will too.
Should the reviewing be reciprocal?
Yes, wherever possible. Reciprocity turns review from an audit into a shared craft. People who check each other’s reports learn what makes a report easy to follow, and their own writing improves. It also spreads the load, so review does not become one person’s burden.
What if both readers share the same blind spot?
It happens. Two colleagues from the same team, trained identically, may both accept an assumption that is wrong. That is why rotating pairs and occasionally including someone from outside the immediate area is helpful. A second reader reduces the chance of an unnoticed slip; it does not eliminate the chance. Treat pairing as one layer, not as proof that a report is correct.
Does this apply outside the aviation workplace?
Very much so, and in everyday administration. Anyone making travel arrangements or submitting paperwork benefits from having a second person read it. Families researching care abroad, for instance, may use a medical tourism site like KneeHospitals, a site that describes itself as a medical tourism resource with a section per country and enquiry forms for quotes and consultations. Before submitting such a form, ask someone you trust to check the typed details against your documents. This is a clerical check; decisions about care are for qualified professionals.
What does a weak review look like?
It looks like agreement without evidence. The reviewer returns the report in two minutes with “fine by me” and cannot say which line they compared with which attachment. Another weak pattern is the review that rewrites the author’s prose for style while leaving a wrong date untouched. A third is the review that happens after filing, when changing anything triggers a formal amendment. Each of these gives the comfort of a check without the substance.
A useful test: ask the reviewer to name one thing they verified. If they can say “I matched the equipment number to the log”, the review happened. If they cannot, it did not.
How long should a pair review take?
For a typical hazard report with a few attachments, ten to fifteen minutes is a reasonable expectation, and shorter for simple ones. Longer reports and investigation summaries naturally take more time. If reviews regularly take far longer, the reports may be too sprawling, and that is itself a finding worth feeding back to authors.
Can the reviewer reject a report?
The reviewer should return it with queries, not reject it. The author owns the report and the reviewer owns the questions. If the author and reviewer disagree on whether a concern is valid, the matter can go to the safety manager, but in practice most disagreements dissolve once the evidence is on the table.
How do we start with a small team?
Begin with one document type, such as hazard reports, and one rule: no report is filed until a named colleague has read it against its evidence and replied with either a query or the word “checked”. Run it for a month. Keep a simple tally of what the readers found, grouped loosely as missing detail, inconsistency, unclear reasoning and vague action. After a month, the tally tells you where your reports are weakest, and training can focus there.
What signs show the pairing is paying off?
Watch for qualitative signals. Fewer back-and-forth queries from the committee. Authors who attach evidence without being asked. Reviewers who volunteer for the job. Questions that become sharper, as the team gets used to asking them. You do not need elaborate statistics for a small team to see these changes, and you should be wary of inventing numbers that the data cannot support.
Should the pair review also cover corrective actions?
Yes, and perhaps more than the initial report. Actions are where good intentions go vague. A reviewer reading “improve awareness of ramp hazards” will ask who, by when, and how anyone will know it was done. Turning that vague line into a named task with a date is among the most valuable things a second reader does, and it costs no more than a couple of sentences of discussion.
What is the one-paragraph summary?
Authors cannot see their own gaps. A second reader with access to the evidence can. Make the review a routine step for hazard reports and risk assessments, keep the tone factual, rotate the pairs, return favours, and treat the check as one useful layer rather than a guarantee. Small teams can start next week with a single document type and a short tally of what the checks find.
