Why does a team behave differently from its formal rules?

We often speak of 'the norm' as if there were only one. Research into social influence gives us a more precise vocabulary. Cialdini, Reno and Kallgren distinguished descriptive norms from injunctive norms: what people believe others do, and what they believe others approve or disapprove of. Their experiments on littering suggested that which norm becomes salient can influence behaviour. [1] The study does not investigate warehouse safety, but the distinction helps formulate workplace research questions.

At least four layers are relevant. The formal rule is the documented requirement: for example, inspect a vehicle before use. The perceived descriptive norm concerns how many colleagues you think actually inspect their vehicles. The perceived approval norm concerns which behaviour colleagues and supervisors respect or condemn. A personal norm concerns what the employee personally considers right, independently of group pressure.

All four can differ. An employee may believe checks are sensible, know that management requires them, assume most colleagues skip them and worry that questioning a senior worker will be seen as difficult. The immediate decision is then not simply a test of knowledge. Different choices carry different social and practical consequences.

Cristina Bicchieri's work further examines empirical expectations, normative expectations and whether an individual's preference for a particular action is conditional on those expectations. [2] This matters because a shared habit is not automatically a social norm. People may all skip a step because they face the same impossible time constraint. What appears to be conformity could instead be a common response to bad process design.

So a sound investigation does not stop at 'What do employees do?' It also asks 'What do you think colleagues would say or do if you behaved differently?' When behaviour changes substantially in the presence of respected peers, social expectation may be part of the explanation. When a practice is impossible even without an audience, practical conditions may be more important. These are hypotheses to test, not labels to attach to employees.

The reference group also matters. A corporate message that 'most employees follow the procedure' may not change much when every relevant role model in one shift does the opposite. The group whose opinion counts is often local: a team, craft community, supervisor or a few respected peers. For that reason, broad survey averages can conceal the very mechanism that an intervention needs to change.

Employees do not operate as isolated decision-makers. They want to contribute, be considered competent and avoid disrupting colleagues without good reason. Behaviour therefore has a second meaning. Following a procedure can signal reliability or a lack of trust. Reporting a mistake can signal responsibility or disloyalty. Challenging a senior colleague can mean preventing a problem or threatening someone's status.

This is precisely why words such as safe, efficient and professional cannot be assumed to carry the same meaning everywhere. Management may use 'efficient' to mean avoiding errors and rework. A shift team may interpret it as never slowing production. A training presentation can say 'safety comes first' while daily management focuses almost exclusively on output per hour. Employees see two signals and may conclude that the true standard is what receives attention and rewards.

Research into organisational safety climate shows why these perceived priorities matter. Zohar conceptualised safety climate in terms of shared employee perceptions about the relative importance of safety in the organisation. [3] The presence of a policy is not equivalent to a climate in which safety is taken seriously. This does not establish a universal explanation for noncompliance, but it does suggest that leadership decisions and workflow signals belong in the diagnosis.

Social influence is not always something people recognise in themselves. An employee may sincerely describe skipping a check as an independent, reasonable choice while the behaviour of experienced colleagues quietly defines the default. This does not require a simplistic claim about hidden neurological control. Observed conduct and anticipated approval are ordinary, testable aspects of a social situation. There is no single brain switch for obedience.

The question is which associations make the desired behaviour feel ordinary or unusual in this particular group. For one experienced operator, checking equipment may represent craftsmanship. A newcomer may worry that checking signals inexperience. Another team may interpret a pre-use check as respect for colleagues who will use the equipment next. Do not turn these possibilities into fixed psychological profiles by age, seniority or job title. Compare the actual teams, tasks and circumstances.

The table below presents illustrative hypotheses, not measured Neurofactor findings. The same rule may acquire very different meanings in different groups.

There is another reason to be careful with communication. Repeatedly announcing that 'far too many employees skip checks' can unintentionally tell everyone that skipping is widespread. In a field experiment on household energy conservation, Schultz and colleagues showed how descriptive normative information could produce unwanted changes among some households, while an additional approval cue prevented that effect in their setting. [6] This does not prove that every workplace message behaves the same way. It does warn against making the undesired action sound like the established standard.

The answer is not to pretend that most colleagues already comply. Inventing a majority or hiding safety problems would undermine trust. Instead, identify real, verified examples of good practice and explain why they represent professional teamwork. Show employees what the organisation makes possible and what leaders actually support.

A formal rule becomes a living standard only when people can help each other act on it. That takes more than another poster. Whether someone can question a decision, admit a mistake or request extra time influences what they dare to do in front of colleagues. Edmondson's field research examined psychological safety and learning in work teams. A shared belief that interpersonal risk-taking is reasonably safe was associated with team learning behaviour. [4]

That does not imply psychologically safe teams automatically follow every procedure. It suggests a necessary line of inquiry: can people raise the issue without anticipating embarrassment or retaliation? Morrison's review of employee voice and silence describes why employees sometimes withhold concerns and useful information. [5] In a change programme, this is particularly important. A problem that people cannot discuss safely is unlikely to emerge clearly from a standard satisfaction questionnaire.

Imagine a warehouse operator who discovers that a faulty scanner makes a safety confirmation unreliable. When the operator reports the defect, a supervisor responds, 'We have no time for that now.' Next time, hesitation is understandable. There is no need to infer a mysterious subconscious motive. The concrete consequence has already taught the employee something about what gets rewarded or punished.

Leadership therefore matters through behaviour, not simply communication. Detert and Burris studied managerial openness and employee voice among thousands of employees and managers in a restaurant chain. Their findings highlight the relationship between perceived openness, psychological safety and speaking up. [8] The specific setting limits direct generalisation, but the research supports treating managers' responses as part of the workplace mechanism rather than a separate HR issue.

The revealing moment often comes when somebody does the right thing. What happens next? Are they supported, given time and thanked for noticing a risk? Does the team help resolve the problem? Or is the person treated as an obstacle? The collective response can strengthen or weaken an unwritten rule more effectively than a beautifully worded policy.

This is a fictional composite scenario. It does not describe measured Neurofactor respondents, clients or results. Imagine a distribution centre where employees use powered internal transport vehicles. The safety procedure requires a brief check before the first journey of a shift: look for visible damage, verify critical functions and report defects. The organisation wants every vehicle to start safely, yet the checks are often skipped during the morning rush.

Management's first reaction is understandable: explain the risk again. A reminder appears by the entrance and the team receives additional training. In a knowledge quiz, everyone can describe the procedure. However, a correct answer gives us no reliable reason to conclude that behaviour has changed.

A sound investigation would preserve at least two explanations. Explanation A involves a social norm. Experienced colleagues immediately get moving; newcomers imitate them. Doing the check carries an unwelcome association with inexperience and holding others back. Explanation B involves a practical barrier. Essential checkpoints are hard to access, defects take too long to report and the shift schedule leaves little preparation time. In that case, 'peer pressure' may be only part of the explanation.

Both may operate at once. A senior employee can personally endorse safety but skip the check when a manager asks why the first order has not moved. A newcomer hears the formal instruction yet sees that output is monitored instantly while inspections receive little visible support. The written rule says 'start safely'. The experienced practical signal may feel like 'keep the operation moving'.

What should research do next? Observe comparable shifts where this can be done safely, transparently and lawfully, without covertly building individual performance files. Ask confidentially what reaction employees expect when they complete the check. Establish whether the necessary time, tools and authority are available. Compare perceptions of colleagues' behaviour with appropriately aggregated actual practice.

Only then select an intervention. If there is no time to inspect, the workflow must change. If respected operators establish the default, let them help demonstrate a credible safe routine. If defects are not followed up, fix the reporting and repair loop. A new poster cannot repair a broken scanner or remove social penalties for doing the right thing.

The purpose of this example is not to divide employees into obedient workers and resisters. It is to make a safer system possible, learn what determines everyday decisions and prevent the organisation from blaming people for a contradiction it has created.

Visible behaviour has multiple possible causes. This is a research decision aid, not a validated scoring system or proof of causation.

The final row prevents an important false success story. More completed forms may indicate that more checks are performed. They may also indicate that employees have become better at filling in forms. A punitive system can even reduce the reporting of problems. Measure the quality of the action and the safety of reporting, not simply the quantity of ticked boxes.

The organisation should also question whether the original rule is still workable. A workaround may reveal valuable information about contradictory objectives or inadequate equipment. That does not make unsafe behaviour acceptable. It means a responsible employer must redesign the conditions so that the required safe practice can genuinely be performed.

Step 1 - Define the desired action precisely. Do not begin with 'we want a better safety culture'. Identify who must do what, at what point, under which conditions and with which legitimate exceptions. 'Carry out the pre-use vehicle inspection before the first journey' is an observable action. When the formal requirement is unclear or unworkable, improvement begins with the requirement itself.

Step 2 - Map the setting where the decision occurs. Walk through the process with employees during ordinary and busy periods. Are instructions, materials, equipment, time, authority and support genuinely available? What happens when someone pauses or reports a fault? Document conflicts between formal objectives. Keep the observation proportionate and transparent rather than turning it into individual surveillance.

Step 3 - Measure real actions separately from self-reports. A questionnaire asking whether employees follow a procedure cannot establish compliance. Use appropriate, permitted and aggregated behavioural observation and process indicators. Determine whether completed logs truly represent the required actions. Ticking a box is not the same as inspecting a vehicle.

Step 4 - Investigate empirical expectations. In a confidential setting, ask 'How many colleagues on your shift actually do this, in your opinion?' and 'Who do you learn from when you are unsure how something should be done?' The relevant reference group may be a shift, a professional community or a handful of respected colleagues, not the organisation's entire workforce.

Step 5 - Investigate approval expectations and social costs. Present specific situations: someone pauses the line to check equipment, reports a near miss or questions a senior colleague. What response would employees expect from their peers and supervisors? Ask about appreciation and disapproval. This reveals why 'everyone says the rule matters' can coexist with 'I cannot follow it without losing face'.

Step 6 - Explore associations attached to the action. Use open questions to discover what checking, reporting or speaking up brings to mind: craftsmanship, care, delays, distrust, responsibility, incompetence or team loyalty. Let positive and negative meanings coexist. Implicit association methods may be useful as additional evidence for a defined hypothesis, but reaction-time differences do not diagnose a person's motives or predict individual compliance.

Step 7 - Compare groups and conditions without stereotypes. A night shift may behave differently from a day shift because of supervision, workload, available equipment or local role models. An onboarding group may experience different pressures from a long-established team. Investigate those practical conditions before making claims about age, personality or job category. Ensure the samples are appropriate and the comparisons fair.

Step 8 - Test a limited intervention with a credible comparison. If you change training, scheduling, reward systems, leadership and equipment at once, it becomes difficult to learn which element mattered. Start with the most strongly supported barrier. You might make time available for checks, have respected colleagues model the correct practice and improve follow-up on reported defects. Compare periods or groups where that is feasible, taking operational differences into account.

Step 9 - Evaluate maintenance, side effects and the actual result. Revisit behaviour after a week, over multiple cycles and during peak demand. Assess the quality of the action, the safety of reporting, relevant operational outcomes and the perceived team norm. Look for hidden workarounds, increased workload or suppressed incident reports. Improvement matters only when the required behaviour remains feasible and effective outside a demonstration or audit.

This process is a framework for investigating explanations, not a guaranteed intervention package. Evidence strength depends on measurement quality, sample size, research design and working conditions. A before-and-after comparison can help a team learn, but without suitable controls it does not automatically establish that the intervention caused an observed change.

1. Let expertise set the example. Identify colleagues who are genuinely trusted for their work, not merely their formal title. Invite them to demonstrate the desired practice voluntarily and credibly. The example must be genuine: a staged performance by someone who normally skips the procedure will not build lasting trust.

2. Change what management actually rewards. When supervisors discuss speed daily and safety only after an incident, employees notice. Give attention to preparation quality, safe stopping, problem reporting and corrective action even during busy periods. Do not quietly penalise the person who raises a concern through poorer shifts or negative performance reviews.

3. Provide time and functioning tools. A protocol requiring several additional actions but allocating no preparation time creates predictable conflict. Adjust the timetable, allocation of responsibilities, equipment or sequence of work. Otherwise the workplace may teach people that they must ignore the rule to achieve the expected output.

4. Specify and demonstrate the norm. 'We work safely' is easy to agree with and hard to observe. 'We complete the check before the first journey and report any defect immediately' is concrete. Only use verified information when communicating how many colleagues currently perform the action. A fabricated majority is not a legitimate change strategy.

5. Provide social support without shaming. Explain why the desired behaviour is part of professional teamwork, not proof that some employees are morally superior. Allow space for reasonable questions and learning. Public humiliation can discourage honest reporting even when it appears to increase immediate compliance.

6. Make speaking up workable. Practise short phrases such as 'I think this check is missing. Can we do it together?' or 'I need to resolve this safely before we continue.' Supervisors must respond consistently and respectfully. Psychological safety is not the absence of accountability; it is the possibility of raising relevant concerns without unnecessary interpersonal harm. [4][5]

7. Reduce the social cost of the first move. Do not require the newest member of a team to challenge every experienced colleague alone. Ask trusted senior colleagues to begin the correct practice visibly. Have supervisors protect the time required. This allows a new standard to become a collective routine rather than an individual's act of social courage.

8. Give feedback at the right level. Discuss how the workflow functions and what support the group needs. Prefer proportionate, aggregated indicators over leaderboards that label individual employees as 'unsafe' or 'disloyal'. Check whether the numbers truly reflect safe conduct. A dashboard should support decisions, not become the objective of the system.

9. Change the system when employees identify a real problem. Take concerns about impractical instructions seriously. Verify them and revise the process when that is safe and lawful. An organisation that expects only its employees to adapt, while management refuses to address contradictory expectations, weakens the credibility of its own standards.

Do not implement these nine choices as nine unrelated campaigns. Select the smallest combination that addresses the identified cause, record the expected mechanism and test what happens. That approach is more useful than an attractive general-purpose culture slogan.

Suppose a survey finds that more employees now agree with a formal rule. Has the social norm changed? Possibly, but the survey may simply show that employees have learned which answer is expected. Norm change should also be investigated through expectations about colleagues' actual behaviour, anticipated approval, consistent leadership reactions and ultimately the quality of the required action.

Tankard and Paluck review how people's perceptions of norms can help explain social change and how changing such perceptions may matter. [7] This is not permission to manipulate employees with fabricated statistics. A trustworthy intervention makes real practice visible or creates conditions under which a new practice can develop. It does not simply tell employees that everybody already complies.

Compare at least four outcomes. Actual behaviour: is the action performed properly at the relevant moment? Descriptive expectation: do employees believe their immediate colleagues perform it? Approval expectation: do employees expect support for following the rule or raising a concern? Practical feasibility: are the resources, time and authority genuinely available? Where appropriate, add evidence about the real operational outcome, such as verified errors or incidents, while interpreting reporting patterns carefully.

Beware of perverse incentives. Rewarding a team for 'zero reports' could reduce incidents or just increase silence. Measuring only completed inspection forms could encourage administrative compliance. Attaching personal consequences to confidential association responses will quickly make those responses less trustworthy. The evaluation must therefore look at reporting safety and possible unintended behaviour, not just attractive metrics.

Write success criteria before testing. For instance: critical checks are completed correctly during realistic workloads; faults are followed up; employees increasingly expect the action to be normal and respected; and the new procedure does not create unacceptable delays or additional risks. Without predefined standards, almost any single improvement can be presented as a win.

Ultimately, campaign recall is not the desired outcome. The question is whether employees can perform the required action in ordinary work without having to choose between safety, colleagues and their assigned output. Only then has the organisation changed more than its communication.

A team may know the instruction and still follow another routine. That is not a mystery resolved by an additional poster. People observe colleagues, notice what managers reward, anticipate reactions to speaking up and learn which actions are possible during a demanding shift. Over time, those experiences give a formal rule a practical and social meaning.

If you want to change behaviour, investigate that meaning within the relevant group. Separate descriptive norms, approval expectations, personal convictions and real process barriers. Make the desired action visible and feasible. Ensure managers do not reward the opposite behaviour. Then evaluate whether it is performed correctly when no one has organised a demonstration for the researchers.

A standard does not become reality because it has been written down. It becomes reality when people can act on it, recognise it as part of good work and can rely on colleagues and leaders to support it.

• Associations in behaviour change: why knowing is not doing
• Why can positive change feel like a loss?
• Why do good intentions fade while motivation remains?