Why people follow safety rules only when watched: is it the worker or the system around the rule?

This piece grew out of conversations between practitioners on Reddit about a familiar frustration: workers follow the rule when a supervisor is nearby and drop it when nobody is. The discussion tends to frame this as an attitude problem. We listen to what practitioners are talking about to share what is being discussed and to add to the conversation. Here, that means a better question and a way to test it. The reader will come away with a method for telling whether a rule is followed because of oversight or because it works, and a list of what a safety culture assessment should examine beyond observation counts.

Observed compliance is compliance that depends on someone being present to see it. It is a signal about the system around the rule, not only about the person following it.

Moving from compliance to culture: ask what the organisation built around the rule

Workers have their own duties. Section 7 of the Health and Safety at Work etc. Act 1974 requires employees to take reasonable care and to cooperate with their employer. The employer's duties are wider. Regulation 5 of the Management of Health and Safety at Work Regulations 1999 requires suitable arrangements for the planning, organisation, control, monitoring and review of preventive and protective measures. A rule that is followed only when watched is evidence that one or more of those arrangements is not working.

The useful question changes from "why will they not comply?" to "what makes the rule hard to follow or easy to skip?" The answer usually sits in habit, in how supervisors are rewarded and in how the procedure was written. For a practical route to ownership of safety at every level, see this piece on building a culture of accountability so safety does not depend on supervision.

Three things organisations build around a rule that is followed only when watched

Habit and normalised deviance: when the workaround becomes the job

Normalisation of deviance is the process by which a departure from a rule becomes accepted as normal because nothing bad has happened. A guard is lifted to clear a recurring jam on a pallet wrapping line. Nobody is hurt, so the next shift does the same. Within a month the workaround is how the job is taught. The worker is not defying the rule. The worker is following the local habit, and the rule applies only when a visitor is on the line.

Supervision that rewards speed: what the watcher actually checks

Supervisors respond to what they are measured on. If the shift target is visible on a board and stop-work calls are not recorded anywhere, workers learn which one counts. The rule is then followed when a manager is watching because the manager changes the priorities for that hour. When the manager leaves, the target returns as the priority.

Procedures nobody tested: rules that fail at normal working pace

A procedure written at a desk can add steps that the task cannot absorb. A lone-working check that takes ten minutes on a delivery round with a fixed schedule will be skipped. If the compliant method is slower, harder or less reliable than the workaround, the rule will hold only while someone is there to enforce it.

How to tell oversight-driven compliance from compliance that works

Five patterns account for most of the cases practitioners describe. Each row pairs what is seen on site with a check that separates oversight from workability.

What you see on site	The check to run	What you learn
Compliance is high when a supervisor or auditor is in sight and drops once they leave	Compare night, weekend and cover-supervisor shifts with day shift on the same records	Whether the rule depends on someone watching
Workers describe a quicker way that skips a step	Have the people who do the task walk through it and time the compliant method against the workaround	Whether the procedure works in practice
Supervisors are praised for output and rarely mentioned for stopping a job	Review what supervisors are measured on and how stop-work calls are received	Whether supervision rewards speed over safety
The same deviation appears across several teams and nothing has gone wrong	Search incident and near-miss records for it and check whether anyone has ever reviewed it	Whether a deviation has become normal before it causes harm
The procedure was written at a desk and issued without a trial	Trial it on the live task with the operators who will use it	Whether the rule can be followed at normal working pace

What a safety culture assessment should look for beyond observation counts

An observation count says how often a behaviour was seen. It does not say why. An assessment that looks for causes should cover:

• Whether each rule was trialled with the people who do the work before it was issued
• Whether the compliant method takes longer than the workaround and by how much
• How supervisors are measured and whether production targets sit alongside safety stops
• What happens to a worker who stops a job, based on records of stop-work calls and the response to them
• Whether near misses are reported at similar rates from every shift
• Whether deviations that are found change the procedure or are only logged as behaviours
• Whether management decisions on staffing, resources and maintenance backlog are assessed alongside front-line attitudes

The last item matters because culture surveys often measure front-line attitudes and leave management decisions unexamined. That gives a partial picture of why a rule is not followed.

Why observation programmes change behaviour without changing conditions

Behavioural safety programmes define safe and unsafe behaviours, use trained observers and give feedback. They can reduce unsafe behaviours, raise the profile of safety and increase manager visibility. They do not change the guard, the procedure or the target.

The regulator says as much. HSE's guidance on the limits of behavioural safety programmes treats them as one tool among several. It states that they cannot stand in for sound engineering and safety management systems. It also sets out conditions to satisfy first, including identified hazards, accurate operating procedures and a proper balance between production and safety. The page is written mainly for major hazard sites, but the question it poses applies to any workplace: are the conditions right before the behaviour is measured?

A rising observation score can therefore mean two things. The rule is followed because it works, or the rule is followed because observers are present. Only a check against conditions can tell the two apart.

What to look for in safety culture assessment software for UK businesses

Safety culture assessment software collects survey and observation data and shows where attitudes and behaviours differ across the organisation. For the question in this piece, it should split results by shift, team and role so that observed and unobserved periods can be compared. It should include questions for supervisors and managers, not only front-line workers.

It should record stop-work calls and near misses by source and link each finding to a corrective action or procedure review. Trends over time should be visible. Where responses can identify individuals, UK GDPR applies, so check anonymity thresholds and where the data is hosted.

Examine the system around the rule, not only the worker following it

A rule followed only when watched points to habit, supervision or a procedure that was never tested. Compare observed and unobserved shifts, time the compliant method against the workaround, check what supervisors are rewarded for and look at management decisions as well as front-line attitudes. Each step moves the conversation from blaming workers to examining management systems, which is where the employer's duty sits.

Readers who want to continue the discussion can join practitioners in Ideagen's Reddit community.