Psychological Safety Is an Escalation Control

A delivery lead hears that a dependency may fail. A technical specialist has noticed a control gap. A junior analyst has evidence that the forecast assumption is no longer credible. None raises it in the steering meeting. By the time the concern reaches a sponsor, the organisation is managing an incident rather than choosing among options.

This is often called a communication problem. More precisely, it is an escalation-control problem. A delivery system cannot manage information that does not travel from the person who notices it to someone with the authority to assess, decide, investigate, or act.

Psychological safety matters because it affects that journey. But it should not be treated as a morale slogan or a sentiment score that sits beside governance. It should be designed into how teams invite dissent, receive bad news, assess uncertain concerns, route material issues, protect people from retaliation, and feed decisions back to those who raised them.

Safety does not mean comfort or agreement

Psychological safety is commonly defined as a shared belief that a team is safe for interpersonal risk-taking. NHS England, Looking after your team’s health and wellbeing guide In practical delivery terms, that includes admitting uncertainty, asking for help, challenging a plan, reporting an error, saying that an assumption no longer holds, or disagreeing with a more senior colleague.

It does not mean that every concern is correct, every idea is adopted, disagreement is comfortable, performance standards disappear, or evidence no longer matters. A safe team can have demanding standards and direct challenge. In fact, it needs both: people must be able to surface a concern, and the organisation must be able to test it proportionately and make a clear decision.

The evidence base supports this distinction with appropriate caution. Frazier and colleagues’ meta-analysis aggregated 136 independent samples spanning more than 22,000 individuals and nearly 5,000 groups, examining antecedents and outcomes associated with psychological safety. Frazier et al., Psychological Safety: A Meta-Analytic Review and Extension It does not show that one meeting practice will solve every project problem. It does support treating interpersonal risk-taking, voice, learning, and team conditions as legitimate management concerns rather than as soft extras.

Escalation needs both voice and a control path

Speaking up is useful only if it changes what the system can know or do. Research on voice in action teams describes raising concerns and suggestions as an important part of team communication, particularly where error prevention matters, while also noting that the literature is heterogeneous. Krenz & Burtscher, Investigating voice in action teams

For a project, programme, product, or portfolio, a credible escalation path has six elements:

Element	What it makes possible
Safe route	A person can raise a concern through a normal, alternative, confidential, or formal route appropriate to the issue.
Clear prompt	Meetings, reviews, and decision packs explicitly ask for uncertainty, dissent, changed assumptions, and unresolved constraints.
Proportionate assessment	The concern is triaged without requiring the person raising it to prove the entire case before anyone looks.
Accountable authority	Someone can decide, fund investigation, accept risk, change scope, or escalate further.
Response and feedback	The organisation acknowledges, clarifies, routes, acts, and explains what can be shared.
Learning and protection	Lessons update controls and decisions; retaliation, humiliation, or reputational penalty are identified and addressed.

Remove one of these elements and “please speak up” becomes weak. A concern may be raised but then disappear in a meeting. It may be heard but never reach authority. It may be investigated but the person who raised it never learns what happened. Or a visible negative response may teach everyone else that silence is safer.

Design review practices that invite useful challenge

The highest-leverage changes are often mundane. Add explicit prompts to the moments where commitment increases:

• “What would make this plan fail even if the team executes it well?”
• “Which assumption is most likely to be wrong, and what evidence would tell us?”
• “What evidence, advice, constraint, or approval is missing?”
• “Who sees a consequence that the current decision pack does not represent?”
• “What could a junior, supplier, customer-facing, operations, or specialist colleague hesitate to say here?”

Make it possible to record dissent without forcing immediate consensus. Name the decision owner, the evidence standard, the next investigation or action, and the route for an issue that exceeds the forum’s authority. Distinguish a delivery concern from a formal grievance, misconduct allegation, safety incident, security issue, or protected disclosure; each may require a different route and specialist handling.

NHS England’s Freedom to Speak Up guidance is a useful analogy, not a direct policy template. It emphasises routes, leadership, communication, response, learning, and action as parts of a credible environment for raising concerns. NHS England, Freedom to Speak Up: guide for leaders The delivery translation is not to create a new bureaucracy. It is to ensure a concern has an observable path from signal to decision.

The leader’s response is the real test

Teams learn more from what happens after bad news than from what leaders say before it. A useful response sequence is:

1. Acknowledge. Thank the person for raising it; do not force them to defend their character or motives.
2. Clarify. Separate what was observed, inferred, assumed, and unknown. Ask what decision or harm may be affected.
3. Protect and route. Choose an appropriate confidential, independent, formal, or normal management route. Do not promise confidentiality that cannot be maintained.
4. Assess and decide. Assign a proportionate investigation, owner, decision authority, date, and escalation threshold.
5. Communicate. Tell the person raising the concern what will happen and, where appropriate, what happened. Explain any limits on information sharing.
6. Learn. Update a risk, assumption, decision, control, design, or leadership practice. If no action is required, record why rather than treating the concern as an inconvenience.

Public NHS material illustrates the importance of this response loop. Its internal Freedom to Speak Up update records that people’s confidence was affected not only by whether a route existed, but by senior-leader engagement, response times, visible action, and whether concerns appeared to make a difference. NHS England, Internal Freedom to Speak Up update That is one organisational context, not proof of a universal intervention. The practical lesson travels: a speak-up route without credible response will eventually become a silence system.

Do not use safety to avoid accountability

Psychological safety does not make every escalation valid. Leaders should ask for evidence, test alternatives, seek expertise, protect due process, and hold people accountable for their decisions and conduct. The control is balanced only when people can raise a concern without punishment and the organisation can distinguish weak evidence from a material issue through fair assessment.

Likewise, anonymous channels can be valuable but are not a substitute for normal escalation. An organisation that only learns about delivery risk through anonymous reporting has a management-path problem. Conversely, requiring concerns to be raised only through a line manager may be unsafe when that manager is implicated, unavailable, or unable to act.

Use the escalation-health diagnostic at programme, project, product, portfolio, and functional review points. Test the route, prompt, assessment, authority, response, learning, and protection—not merely whether staff report high psychological safety.

The goal is not a team where nobody feels discomfort. The goal is a delivery system in which the discomfort of raising a credible concern is lower than the cost of letting it remain hidden.

Sources and editorial limits

• Frazier et al., Psychological Safety: A Meta-Analytic Review and Extension and Krenz & Burtscher, Investigating voice in action teams — peer-reviewed context for psychological safety and speaking up; associations and action-team findings do not prove a specific delivery intervention.
• NHS England, Freedom to Speak Up: guide for leaders, NHS England, How do I take action?, and NHS England, Internal Freedom to Speak Up update — public speak-up and response-practice context from healthcare, used as a qualified analogy.

The escalation-control model, response sequence, prompts, and diagnostic are Strategica synthesis. This article does not replace formal whistleblowing, grievance, HR, safeguarding, legal, safety, security, ethics, regulatory, or investigative processes.