Psychological Safety Has to Be Designed Into the Work

Organizations often talk about psychological safety as though it is primarily a matter of encouraging people to speak up. But telling employees that their opinions are welcome is not enough. Neither is putting psychological safety into a policy, conducting an annual survey, or asking employees to rate how safe they feel.

Psychological safety is revealed in what happens when someone actually speaks.

Amy Edmondson and Michaela Kerrissey describe psychological safety as a shared permission for candor: people can take the interpersonal risks involved in asking questions, admitting mistakes, and disagreeing with colleagues. Importantly, that does not mean work should always feel comfortable, that everyone must agree, or that employees should always get their way. Candor can be uncomfortable.

That distinction matters because organizations can easily conflate the appearance of safety with the conditions that enable candor.

A meeting where everyone agrees may look healthy. A problem quietly corrected before anyone notices may appear to be efficiency. An employee who never challenges a supervisor may look cooperative. A workplace survey may even report that employees generally feel safe.

None of those things necessarily tells us what happens when someone says, “I think we’re wrong.”

In knowledge-intensive organizations such as colleges, universities, and libraries, this is also an information problem: when people do not feel able to question, clarify, or disagree, important knowledge never reaches the people making decisions.

For academic librarians, that should feel familiar. Libraries are full of work that depends on professional judgment: collection decisions, instruction, assessment, reference practice, metadata, technology adoption, accessibility, staffing, scholarly communication, and increasingly AI. Much of that work also crosses organizational boundaries, which means librarians are often expected to raise concerns, question assumptions, or explain risks to people with more institutional authority.

That makes psychological safety more than a workplace-culture issue. It affects whether professional expertise actually enters the decision-making process.

Measurement Is Not the Intervention

Surveys can provide useful information. Edmondson and Kerrissey discuss psychological safety measures, but, importantly, they describe the resulting data as material for conversations about improving the work environment.

Polling employees, then, is not the same thing as creating psychological safety.

The more revealing questions come afterward. What happens when an employee questions a decision? When does someone admit making a mistake? When an employee says an instruction does not make sense? When someone asks for help? When someone offers an answer that leadership does not want to hear?

In academic libraries, those moments can be very ordinary. A librarian may question a new platform, identify a privacy concern, challenge an assessment method, object to a workflow that creates accessibility barriers, or point out that a new AI tool is not doing what its marketing claims.

The issue is not whether leadership ultimately agrees. The issue is whether the concern can be raised, examined, and answered without the person being characterized as resistant, negative, or difficult.

Those interactions teach employees far more about the actual culture than a statement telling them that speaking up is encouraged.

Don’t Confuse Communication Style With Competence

This becomes particularly important when thinking about neurodivergence.

People do not all process information or communicate in the same way. Someone may need additional time to formulate a response. Another person may communicate more effectively in writing than in a fast-moving meeting. Someone may ask detailed questions, interpret instructions literally, communicate very directly, or need expectations made explicit.

A psychologically safe organization should be careful not to turn those differences into judgments about attitude or competence.

Carucci makes a related point about workplace behavior more broadly: what managers interpret as personality flaws or performance problems can sometimes be protective responses to an environment in which openness feels risky. Understanding that possibility does not require excusing inappropriate behavior or turning managers into diagnosticians. It requires becoming curious about the conditions surrounding the behavior before deciding what the behavior means.

That distinction is especially useful in academic libraries, where much of the work is collaborative but also interpretive. A colleague who asks repeated questions about a cataloging rule, an instruction plan, an AI policy, or a vendor claim may not be obstructing progress. They may be seeing ambiguity that others have normalized.

Similarly, a colleague who needs time to think after a meeting may still have the strongest analysis. A colleague who communicates directly may still be offering valuable professional judgment. A colleague who prefers written follow-up may be contributing more carefully, not less.

The broader point is simple:

Why are organizations so quick to interpret difference as deficiency?

Silence does not necessarily mean disengagement. Asking for clarification does not necessarily mean someone was not listening. Directness does not necessarily mean hostility. Needing time to respond does not necessarily indicate a lack of knowledge.

An agreement does not necessarily mean agreement.

Neurodiversity makes this especially important because different ways of processing and communicating information can produce different forms of insight. Inclusion is not simply allowing those differences to exist; it is designing processes that enable them to contribute to organizational knowledge.

Design More Than One Way to Contribute

If organizations genuinely want candor, they should not require everyone to demonstrate it in exactly the same way.

Meetings matter, but they should not be the only place where people’s thinking counts. Edmondson and Kerrissey explicitly recognize that important workplace conversations can be spoken or written, synchronous or asynchronous.

That has a straightforward practical implication: provide multiple ways to contribute.

Allow written follow-up after meetings. Circulate significant questions in advance when possible. Give people time to think before demanding an answer. Use shared documents where employees can develop ideas asynchronously. Make expectations explicit rather than relying unnecessarily on unwritten social cues. Ask for dissent rather than assuming silence means consensus.

For libraries, this is especially relevant because librarianship already depends heavily on documentation, reflective practice, committee work, project planning, and written analysis. There is no reason to treat rapid verbal participation as the only legitimate form of contribution.

These practices do not lower standards. They can increase the amount and quality of information available to the organization.

Look at What the Workplace Rewards

Organizations should also examine what happens after someone identifies a problem.

If employees who quietly repair mistakes are rewarded while employees who surface systemic problems are regarded as troublesome, people learn to keep problems quiet.

If the person who always says yes is considered a “team player,” people learn that agreement is safer than candor. Carucci specifically cautions that habitual agreement can conceal concerns employees do not feel comfortable expressing.

If an employee asks for clarification and is treated as incapable, people learn to pretend they understand.

And if someone repeatedly raises the same unresolved problem and eventually stops, the resulting silence should not automatically be interpreted as evidence that the problem disappeared.

Carucci’s discussion of withdrawal is particularly relevant here: managers can mistake silence for disengagement or lack of capability when speaking has come to feel risky.

Academic libraries should pay attention to this because the profession often values collegiality, service, adaptability, and cooperation. Those are useful values, but they can become counterproductive if “collegial” quietly comes to mean “does not challenge decisions” or “adaptable” means “absorbs every new demand without questioning priorities.”

The organization has to reward the behavior it claims to want.

Psychological Safety Is Built in the Response

Edmondson and Kerrissey make perhaps the most important point: psychological safety cannot simply be mandated. It develops through group interaction. They emphasize leaders’ messaging, modeling, and mentoring, including how leaders ask questions, listen, acknowledge uncertainty, and respond to others' input.

That gives leaders a more useful self-assessment than another survey:

What happens when someone actually takes the risk of speaking?

Do I become defensive?

Do I explain why the person is wrong before understanding what they are saying?

Do I respond differently when criticism comes from someone whose communication style makes me uncomfortable?

Do employees have to phrase difficult information in exactly the right way before I take it seriously?

And perhaps most importantly:

What have employees learned from the way I responded the last time someone disagreed with me?

For academic library leaders, that last question matters because libraries are facing significant change: AI, changing student needs, staffing pressures, new systems, accessibility expectations, evolving research practices, and shifting institutional priorities. These conditions make candid professional judgment more valuable, not less.

Psychological safety is not about eliminating disagreement, discomfort, standards, or accountability. Edmondson and Kerrissey argue explicitly that psychological safety and accountability are separate dimensions and that high performance requires both.

Nor should neurodiversity become another organizational slogan. The point is simpler: if organizations genuinely value different ways of thinking, they need working conditions in which those differences can actually reach the decision-making process.

A survey can tell an organization something about its culture.

The organization’s response to people determines whether their knowledge can actually enter the conversation.

Psychological safety, then, is not simply a workplace benefit. It is part of an inclusive knowledge infrastructure. In libraries, higher education, and other knowledge organizations, the goal should be to create conditions in which people can question, clarify, disagree, reflect, and contribute in different ways—and still be heard.