10 Psychological Safety Survey Questions to Build a Better Workplace

It’s Friday morning, and you’re ready to wrap up a successful week.

That is until you receive your latest employee survey results that deliver one loud message: Your employees feel like their ideas aren’t valued.

You’re surprised. Your colleagues are great. You feel like your ideas and opinions are respected by other members of the leadership team.

But now that you think about it, you have heard about some of the snarky comments people make on other teams.

How people feel anxious to speak up with their ideas or suggestions.

You remember someone saying they were afraid to rock the boat during a review meeting, so they just played it safe by asking the usual questions during the brainstorming session.

But what if these employees’ ideas could have triggered big innovative changes at your company?

What if their willingness to present something different encouraged others to contribute their ideas as well and initiate game-changing products?

Research shows that when employees feel comfortable asking questions, expressing ideas, or constructively criticizing their colleagues, their organizations are more innovative, creative, and adaptable.

When employees feel comfortable poking the status quo and they don’t fear negative consequences, they enjoy what’s known as psychological safety.

Developed by organizational behavioral scientist Amy Edmondson, psychological safety refers to “a shared belief held by members of a team that the team is safe for interpersonal risk taking.”

You can think of it as the difference between raising your hand with a question or forging ahead with a task that's unclear, because you’re afraid to be ridiculed.

It’s also the difference between sharing an undeveloped idea with your team, safe in the knowledge that they’ll consider it nonjudgmentally, or staying quiet and moving forward with the status quo, because you don’t want to be harshly criticized.

If you felt psychological safety, you’d be comfortable presenting bold ideas at your annual strategy meeting without weighing the ramifications. Less time would go into thinking about the optics and more time would go into debating new approaches with your team.

Why should you measure psychological safety?

It’s clear that psychological safety is important. Every organization wants more innovative, creative, and collaborative employees. So should you measure it?

Yes. Since psychological safety is correlated to such positive organizational attributes, it’s worth knowing how much of it your company possesses.

And if you plan to improve the amount of psychological safety in your organization, measuring it helps you know whether it’s increased or decreased within your organization.

Measuring psychological safety can also help you understand the cause of other important metrics such as retention. Employee retention is top of mind for many companies, especially as they navigate “The Great Resignation.”

Today’s employees stick with a job when it gives them a sense of purpose, when there’s no micromanagement, and when they can work effectively with their colleagues.

Essentially, they want to work on a great team.

Teams make complex things happen. They take individuals with their own skills, experiences, and resources and weld them into a tool directed towards a specific goal.

But these makeshift tools aren’t always well constructed. Some teams become incredible forces for change and innovation while others spend so much time arguing internally – or fail to speak to each other at all – that they struggle with the task at hand.

As the amount of collaborative work employees do skyrockets, it’s more important than ever for companies to build superstar teams.

This was the reason for Google’s Project Aristotle, a research project that brought together organizational psychologists, statisticians, engineers, and sociologists to figure out what made the perfect team. They analyzed hundreds of Google teams to figure out why some were more successful than others.

After two years of research, they realized that common shortcuts, like only putting introverts with introverts, weren’t any guarantee of team success.

Instead, they had five key characteristics: dependability, structure and clarity, meaning, impact, and, of course, psychological safety.

You can measure the amount of psychological safety through pulse surveys distributed to different teams.

Distribute the following statements to your employees as part of your next pulse survey. Ask them to rank their level of agreement with each statement, and use their responses as your new benchmark for psychological safety in your workplace.

I can make mistakes without fear that my coworkers will hold it against me.

Cautiousness is not a bad trait. Companies want their employees to be mindful when engaging with customers or executing processes in order to avoid costly mistakes. But in some companies, this cautiousness is used in unproductive ways.

Being overly cautious while managing a customers’ sensitive data is prudent and responsible.

Being cautious about making suggestions during a brainstorming session or asking questions when assigned a new task is not.

The difference between these two examples is how “fear” or “cautiousness” supports the primary goal.

In the first example, the primary goal is processing a customer’s transaction without compromising their financial data. So a little bit of fear has a lot of upsides.

In the other examples, the goal is to think big in order to be innovative and to solve problems, or to receive clarification for a project.

In these cases, fear doesn’t support the goal. Fear prevents a person from speaking up with a question or an idea, leading to unexplored options or costly mistakes.

This kind of fear causes people to fall into the “All or Nothing” thinking trap. This thinking trap prevents people from moving forward unless they’ve reduced all uncertainty or feel absolutely confident or there’s no anxiety about a course of action.

It’s a shame, too, because one of the best ways to innovate is to build on ideas, and one of the most effective ways to avoid expensive re-work is to provide clear, upfront expectations.

For this to happen, employees must feel comfortable sharing their half-formed ideas or “silly” questions.

And for that to happen, employees have to feel comfortable making mistakes in front of their colleagues.

My co-workers welcome opinions different from their own.

Homogenous teams can be efficient. People who think similarly or have the same training and professional experience tend to view situations the same way and reach consensus more quickly.

But there’s a downside to this efficiency. When a group of people share the same point of view, they also share the same blindspots.

On the other hand, diversity of thought allows teams to spot risks and even opportunities. The trade-off is more debate and slower decision-making, but in some organizations, such as organizations focused on growth and innovation, this trade-off is worth it.

Of course, creating an environment with diversity of thought is difficult. Even if employees have differing perspectives or opinions, they may not feel comfortable speaking up and standing out.

There are typically two explanations for this: the personality perspective and the situational perspective.

The personality perspective attributes this unwillingness to speak up to an employee’s personality. They’re not the type to speak up, so they won’t, and companies might “correct” for this by hiring more outgoing and outspoken employees.

The situational perspective attributes “hesitancy to speak up” to the work environment. Employees don’t speak up because they feel like they’ll pay a high social cost for their input. Translation: They don’t feel enough psychological safety.

Asking your employees whether their co-workers welcome diversity of thought is a useful way to gauge how much psychological safety exists within your teams and your company at large.

I feel safe to take a risk on this team

You might have high levels of psychological safety organization-wide, but that doesn’t mean it’s evenly distributed across your teams.

Colleagues need to feel comfortable taking calculated risks within their teams. People who take calculated risks weigh the pros and cons of moving forward with a task or project and learn how to flourish with incomplete information.

This kind of savviness relies on your confidence within a team.

Teams that cultivate psychological safety encourage their employees to take risks and as a result, these employees become better at spotting opportunities.

Keep in mind that psychological safety isn’t about giving employees a cushion to land on when they are reckless. Instead, it’s about giving them an environment in which to grow, and growth can sometimes be painful. Great managers encourage their employees to take calculated risks by:

• Sharing strategic objectives: When employees understand a team or an organization’s larger objectives, they have the context to make better decisions and conduct more effective cost-benefit analyses.
• Improving processes: Creating streamlined processes and failsafes encourages employees to be creative and innovative, because they know the mission-critical tasks have a clear operating procedure. For example, a developer can get creative on a project knowing that there are backup files safely stored elsewhere.
• Defining what constitutes a smart risk: Managers gain experience on where they can be bold and where they need to exercise caution. It’s worth sharing these insights with employees, so they can practice.
• Leading by example: Telling employees that they can disagree isn’t enough, especially if you haven’t built enough trust or credibility. During brainstorming sessions, consider sparking discussion by asking your team to think about what they would do if you completely eliminated a specific process or took a controversial approach to a project.
• Rewarding people who take risks: If someone tries something new and it doesn’t work out, recognize their effort and tie incentives to that risk taking, especially if it was a well-calculated risk. This gives fellow team members an example of the kind of risk taking you want to see, and also encourages them to take risks themselves.

My direct manager values my ideas.

The less receptive managers are to employees’ ideas, the less likely they are to share them, and the more likely your organization is to continue doing the same things over and over again. So ask your employees whether their direct managers value their ideas.

In some cases, managers set aside employees’ ideas because they are not aligned with the work the team is meant to do, there’s no budget for the idea, or the idea just isn’t particularly good. But there are ways to make employees’ ideas feel valued even when they can’t always be executed. And it’s worth the time, because it might encourage employees with great ideas to step forward if they see how respectfully those with underdeveloped ideas are treated.

How can you show your employees that you value their ideas, even if you can’t follow through on them?

• Make time: During your one-on-one meetings, give your team members time to chat about any ideas they’ve been thinking about.
• Ask constructive questions: Help your employees flesh out their ideas without demoralizing them. Ask them how they would execute this idea, how much it would cost, and how it would help the team meet its larger objectives, but do so in a way that’s constructive rather than shooting down the thought.
• Coach them: Help your team members come up with well thought out ideas. For example, if an idea’s promising, you might want to encourage that employee to write a business case to justify it and to get time and funding.

People keep each other informed about work-related issues within the team.

Think about the last time you worked within an organization or a team with minimal psychological safety. You probably felt anxious sharing any issues with your colleagues.

If there was a hiccup, you kept it quiet in the hope that you could smooth things over later.

If you could, great. If you couldn’t, you wound up with a bigger problem than you would’ve had if you’d asked for help.

Teams with great communication, especially about work-related issues, tend to have a higher degree of psychological safety. They know that they can reach out for help, and they understand that others will let them know if there’s a problem.

You can always improve the quality of communication within your team by:

• Having an open door policy: Make it clear to your employees that they can always stop by to speak to you or send you an invite for a 15-minute call to discuss an issue.
• Providing constructive feedback: Be mindful of how you deliver feedback. If there’s an issue, focus on problem solving instead of assigning blame.
• Organizing team building activities: Schedule time for your team to have fun instead of just work. This gives your team an opportunity to build relationships and trust in a way that isn’t exclusively tied to work.
• Conducting communications training: Communication is a skill and it can be learned. Consider investing in some communication training or workshops for your teams. Learning about things like the communication process or nonviolent communication can help.

Members of this team are able to bring up problems and tough issues.

Some of the most common problems faced by teams include difficulty managing change, unclear work expectations, and an unfair division of labor.

When there’s limited psychological safety, it’s easy for these issues to get brushed under the rug because team members don’t feel comfortable speaking up.

The result is that these relatively minor issues fester until they become an infectious problem that destroys the entire organism. Employees working on teams with a high amount of psychological safety are able to speak up when something isn’t right.

If something goes wrong because a team member didn’t do their part or provide a crucial piece of information, they can respectfully air their grievance knowing the other will listen with empathy and take accountability.

Working through tough interpersonal issues allows teams to work even better together when managing trickier business problems. Your team can collaborate more effectively.

A few ways you can encourage problem solving on your team include:

• Conducting training on different problem solving strategies and empathetic listening
• Getting out of your team’s way and trusting them to find a solution to their problems while expressing that you’re available if they need guidance
• Leading by example by being respectful and constructive in your interactions with employees

People on this team never reject others for being different.

An inclusive team environment is a strong indicator of high psychological safety.

When you feel accepted and respected by your colleagues, you feel more comfortable opening up and relating to them, which leads to trust.

On the other hand, if you feel like you’re routinely “othered” you’re less likely to want to open up and share what’s known as a “status-confirming disclosure.”

A status-confirming disclosure is a piece of personal information that you’re worried will highlight your difference and lead your colleagues to make stereotypes about you.

Consider this example highlighted in Harvard Business Review in which a black executive felt uncomfortable talking about the concert she saw over the weekend because it was a gospel artist that she worried her colleagues wouldn’t be familiar with.

Her concern was that they would view her as “different” from them and change their perception of her, even if unconsciously.

This is one of the key differentiators between a diverse organization and an inclusive organization.

A diverse organization has people from different backgrounds.

An inclusive organization ensures all of these individuals feel like they can thrive and work without hiding elements of themselves.

It is easy to ask other members of this team for help.

Often, employees don’t want to ask for help because they’re worried they’ll be a nuisance or they’ll reveal their inadequacy.

But asking for help can often save teams more time by avoiding expensive re-work. At the same time, these requests for help may predominantly fall on one team member who winds up overwhelmed by the responsibility and even resentful of the distraction from their work.

Create an environment where people feel comfortable asking for help.

You can balance this with the demands of a busy workplace by getting into the habit of documenting all processes so that employees have self-serve knowledge hubs available to them.

This way, when someone asks for help, their colleagues can politely direct them to the right resource. You can consider:

• Recording all meetings for a project so that any new employees can quickly get up to speed
• Creating systems for frequently repeated processes and documenting them so that employees can refer to easy-to-use manuals
• Creating dedicated experts within the team on specific areas (e.g., invoicing) so people know who to ask for help, they can get useful information, and the task of “helping” is split evenly among the team

Working with members of this team, my unique skills and talents are valued and utilized.

We do our best work when we’re working on things we’re good at.

But managers often forget this when assigning work.

It goes without saying that there will be times when you have to do things you don’t feel like doing.

But generally speaking, it’s a good idea for managers to allocate tasks to their team members based on what employees are best at. This way, employers can tap into their employees’ full potential.

One report from SHRM finds that 85% of employees don’t feel they can live up to their full potential at work because their skills aren’t properly utilized.

Imagine if they had an opportunity to use them. Through regular performance management meetings, managers can get a better idea of what their employees have to offer and use these gifts accordingly.

Understanding where your employees’ strengths lie can also help you undertake more effective upskilling initiatives. If you know what your employees enjoy and you know where your company’s skills gaps are, you can put employees on an appropriate learning path, giving them a chance to feel more engaged through professional development.

No one on this team would deliberately act in a way that undermines my efforts

Many employees spend more time with their “work family” than their actual family.

It makes sense then that having a good relationship with your colleagues would create a more positive and productive workplace.

Not everyone has a great “work family.” In fact, some employees have to deal with colleagues that struggle to cooperate, collaborate, and communicate.

If you rank poorly on this question, you have an opportunity to enhance your workplace culture by working team-building initiatives into your people and culture strategy.

Amy edmondson's original 7-item psychological safety scale

For a research-backed measure, pair Amy Edmondson's original seven-item scale with your own practical survey statements.

Edmondson published this team-level questionnaire in her 1999 paper, Psychological Safety and Learning Behavior in Work Teams, in Administrative Science Quarterly.

The original scale items

If you want to stay close to the academic instrument, use the wording below as written.

Three of the seven items are negatively worded by design, to break up straight-lining. There's a cost, though.

In psychometric studies, reverse-worded items reliably load onto their own method factor, which can inflate a two-factor solution that has nothing to do with the construct you're actually measuring. Keep that in mind before you read too much into any factor split in your own data.

How to use it in an employee survey

The original ran on a 7-point Likert scale (1 = strongly disagree to 7 = strongly agree). If your core engagement questions use a different scale length, keep them out of the same instrument. A 7-point safety block sitting beside 5-point items forces respondents to recalibrate mid-survey and complicates any rescaling later.

Before scoring, reverse-code items 1, 3, and 5 so that higher always means safer: on a 7-point scale, replace x with 8 − x; on a 5-point scale, 6 − x. Then average the seven items into a team score.

The classic mistake is averaging raw responses without reverse-coding. The three negative items pull against the rest, and a genuinely unsafe team ends up looking healthy.

Audit this in your export before anyone briefs leadership on the numbers.

Edmondson built this as a team-level construct, so the meaningful unit of analysis is the group, not the individual. Aggregate to team, department, or business unit, and check that within-team agreement is high enough to treat the mean as a shared property rather than noise around wildly split perceptions.

Watch the two dimensions inside the score

Beyond the method-factor caveat, there's a substantive split worth watching: how safe people feel with the team leader versus with peers and the wider team.

This matters operationally. A team can score well on peer trust while one manager quietly suppresses everything above the line.

Before prescribing a fix, disaggregate the leader-facing items from the peer-facing ones. A mediocre average driven by the leader items is a management-development problem; one driven by the peer items is a team-norms problem, and the two interventions barely overlap.

Benchmarking honestly

External benchmarks are tempting, but the published norms come from specific cohorts, a Japanese healthcare sample or a US legal team, that share neither the task interdependence nor the reporting structure of your own population. Comparing across cohorts on this construct is close to meaningless.

Use your own baseline instead. Run the scale, re-run it after a defined interval, and compare each team against its own starting point rather than a borrowed global average.

How it fits with practical survey questions

The seven-item scale gives you a validated baseline. Your added practical statements locate where the deficit shows up, such as speaking up in meetings, asking for help, admitting mistakes, or challenging decisions.

Keep the two sets separate in reporting, and hold the validated items verbatim so your trend line stays interpretable. The moment you reword a validated item, you break comparability with your own prior waves.

Save local rewording for the practical, action-oriented items.

In a large organization, this pairing lets you track the core construct over time with the standardized items while giving each manager locally worded questions they can actually act on.

A platform like Sparkbay supports this because the wording of both surveys and dashboards is configurable, so you can keep the validated items verbatim while adapting the practical questions. Report access is mapped automatically to your org hierarchy, so each manager sees only their own teams.

Anonymity isn't optional on a topic where respondents are effectively rating their own manager. Results stay hidden below a minimum number of responses, configurable and set to five by default in Sparkbay, so no manager can trace an answer back to a person, and no one can back into an individual by cross-tabbing a small team against demographics.

How to score and interpret psychological safety survey results

Pick one agreement scale and hold it constant across every psychological safety item. Switch anchors mid-survey and your item-to-item comparisons stop meaning anything.

A 1-5 Likert scale works well:

Edmondson's original scale deliberately mixes positively and negatively worded items to break acquiescence bias. The trick only pays off if you reverse-score the negative items before you average anything.

That reversal step is where most hand-scored surveys quietly go wrong. Mark every negatively worded item in your key, and never blend raw and reversed items in the same average - do that and a genuinely safe team will read as unsafe.

Quick sanity check: before reversal, a healthy team should score high on the positive items and low on the negative ones. If both sets move together, your key is broken.

Average the results at three levels

Report at company, team, and item level, since each answers a different question:

• Company level: average all valid responses across the organization.
• Team or department level: average responses for each group, once the group meets your anonymity threshold.
• Item level: review each statement separately to see what drives the overall score up or down.

Item level is where the real story sits. Two teams can both land at 3.8 for completely different reasons.

One feels fine admitting mistakes but won't challenge a senior decision. The other trusts its peers yet won't take concerns to its manager.

A single team number swallows both.

Separate the two safeties: leader versus peers

Safety toward the manager and safety toward colleagues don't rise and fall together, so split your items into a manager-facing set and a peer-facing set and average each on its own.

The failure modes point in different directions. A team that trusts each other but fears its leader hides problems upward - escalation dies while local firefighting looks perfectly healthy. A team that respects its leader but competes internally simply stops helping colleagues. Blend the two into one figure and you'll aim the fix at the wrong target: coaching the manager when the real work is rewiring how the team collaborates, or the reverse.

Use a starting benchmark to read the score

Treat your first survey as a baseline, not a verdict. On a 1-5 scale, a rough first read:

The bands matter less than the traps in reading them. A wall of neutral 3s might be honest ambivalence, or it might be people who didn't feel safe enough to give a straight answer - so read those neutrals against your response rate.

And a high score sitting on a low response rate should worry you: the people who felt least safe are usually the ones who never opened the survey.

Once you have your own baseline, watch the movement rather than the absolute figure. A climb from 3.4 to 3.7 can be genuine progress.

A drop from 4.2 to 3.8 is worth chasing down, even though 3.8 on its own still looks respectable.

Look for gaps, not just averages

Cut the results by function, region, tenure, level, and manager group wherever you have enough responses to keep answers anonymous. In Sparkbay, results stay hidden below a configurable minimum number of responses - 5 by default - and report access maps automatically to the org hierarchy, so each manager sees only their own teams.

The signal usually lives in the gap between teams doing near-identical work. When one operations team scores 4.3 and a comparable one scores 3.1 on "I can raise concerns without negative consequences," the org average quietly buries it.

Read the distribution too, not only the mean. A 3.8 made of tightly clustered 4s is a different animal from a 3.8 made of eager 5s and silent 2s, and each calls for a different response.

Track the share picking 1 or 2 on your key statements. A polarized team - a vocal group that feels safe, a quieter group that doesn't - tends to be more fragile than its average lets on.

Slice too finely and you risk re-identifying people. Cross two filters, region and tenure for instance, and a group can slip below your threshold even when each filter looks safe on its own.

Let the platform suppress those small cells instead of judging them by eye.

Turn the score into an action signal

High-scoring teams are a resource, not just a result. Find out what they actually do - how they run retrospectives, how they handle errors, how they invite disagreement - and make those habits portable.

Mixed-scoring teams don't need "be more open." They need one or two specific behaviour changes from the manager: calling on quieter people by name, or circling back after someone raises a concern.

At-risk teams need quicker, human follow-up - reading the comments, running confidential listening sessions, and working out whether the driver is the leader's behaviour, workload, unclear roles, or a conflict no one has resolved.

Watch the second-order effect of overreacting to one low score. Hand a manager a bad number with no context and the reflex is to hunt for whoever complained, which destroys the very safety you just measured.

Present low scores as a shared problem, never drill below the threshold, and give managers a script for responding openly instead of getting defensive.

A perfect score was never the point. The point is making it safer to raise bad news early, before silence hardens into mistakes, resignations, or decisions nobody thought to question.

How to run the questionnaire anonymously and how often to repeat it

Psychological safety questions only work when employees believe the process protects them.

If people suspect their manager can trace a low score or a written comment back to them, they'll soften the truth. Cleaner data comes from building anonymity into the survey before you send the first invitation, rather than promising it afterward.

There's an uncomfortable paradox here. The teams with the lowest psychological safety tend to be the ones most likely to distrust an anonymity promise.

So your protections have to be visible and mechanical, not a reassuring line buried in the intro.

Set a clear anonymity threshold

Use a minimum response rule for every dashboard, team view, and demographic cut.

In Sparkbay, results are hidden below a minimum number of responses, 5 by default, and you can configure that threshold. No manager can open a report for a small group and see who said what.

The failure mode in large organizations is the intersection trap. Department, location, tenure, and job level each clear the threshold on their own, while the crossed cut, "new hires in one location," quietly isolates three people.

Confirm your tool enforces the minimum on the combined slice, not just on each filter independently, and pressure-test the two- and three-way crosses your HRBPs will actually build.

When a group falls below the threshold, roll it up into a larger group or wait until the next cycle. Don't make manual exceptions for senior leaders, sensitive teams, or "just this once" requests.

A single override signals that the rule is negotiable.

There's a real trade-off. A higher threshold protects individuals but leaves small teams with no data of their own, so decide in advance whether those teams get rolled into their parent group or simply see nothing until numbers grow.

Tell employees exactly how anonymity works

Don't just say "Your answers are anonymous." Explain the mechanics in plain language.

For example:

• Managers see results only when enough people respond.
• Reports hide groups below the minimum response count.
• Open-text comments do not show names.
• Leaders should look for patterns, not try to identify individuals.
• HR will not share raw responses in a way that could expose a person.

Repeat this in the invitation, the reminder, and the results communication. The most common trust leak isn't the survey design.

It's a well-meaning manager who says in a team meeting, "I could tell that comment was from one of you." Brief managers explicitly on what not to say.

Tell employees who can access each level of reporting, too. In a large enterprise, executives, HRBPs, and line managers each see different scopes, and vague messaging invites people to assume the worst.

One nuance worth stating plainly: confidential is not the same as anonymous. If your platform can link a response to a person but promises not to reveal it, call that confidential and explain the safeguard, rather than using a word that implies the data doesn't exist.

Be careful with open-text comments

Comments give leaders the context that scores can't show, but they also carry more identity risk.

Ask employees not to name colleagues, clients, or highly specific incidents in their written comments. You can still invite detail, but steer people toward themes, examples, and work conditions.

When reporting comments, avoid sharing a single quote from a very small team if the wording makes the person obvious. Summarize the theme instead, then include verbatim comments only where the anonymity threshold and context support it.

Writing style itself can deanonymize, and threshold logic does nothing to stop it. A distinctive turn of phrase, a project only one person owns, the only respondent who writes in a second language, any of these is recognizable with no name attached.

Have a second reader review comments for identifiability before you circulate them, and never pass verbatims to a manager who could plausibly name the author.

Choose a cadence that matches the pace of change

Psychological safety doesn't move meaningfully week to week in most organizations. But an annual survey usually leaves too much time between the signal and the response.

A practical starting point is a quarterly pulse. It gives teams enough time to act, and it gives HR four data points a year to compare.

The real constraint on cadence isn't fatigue. It's action capacity.

Measure only as often as your managers can realistically close the loop, because a survey that outpaces your ability to respond erodes trust faster than no survey at all.

If you run a broader engagement survey, keep the psychological safety items inside it and use shorter pulses between full cycles. That gives you a stable baseline and a lighter way to track movement.

Time the pulse deliberately around known events. Measuring the week after a layoff announcement or during a peak crunch tells you more about the shock than about the underlying climate, so either wait for a steadier moment or read the results with that in mind.

Keep the core questions stable

To compare results over time, keep a core set of psychological safety questions unchanged for at least several cycles.

You can add a few rotating questions for current issues, say a new operating model or a return-to-office change. But avoid rewriting the main items every quarter, because small wording changes can shift scores.

Use the same response scale, the same anonymity rules, and the same reporting structure. Consistency lets you spot real movement instead of noise from the survey design.

Sparkbay lets you configure the wording of both surveys and dashboards, which is useful for matching your organization's language. Set that wording once at the start of a tracking period and resist mid-cycle tweaks, since even a shift from "my team" to "my department" can move a score without any real change in experience.

Compare trends, not single scores

One low score doesn't tell the full story. Look at the direction, the size of the change, and the pattern across teams.

A drop from 8.1 to 7.8 on the engagement score may not mean much on its own. A drop from 8.1 to 6.9 across several teams after a reorganization deserves attention.

Compare results in three ways:

• Over time: How has the same team moved across the last 2 to 4 survey cycles?
• Across levels: Do front-line teams, managers, and senior leaders see safety differently?
• Across topics: Are people comfortable speaking up, but not comfortable admitting mistakes?

The level gap is often the most telling reading. When leaders rate safety far higher than their front line, the score is measuring their blind spot, not the team's climate.

Read a rising score with care, too. Psychological safety scores can climb simply because the people who felt least safe stopped responding or left, so check the trend against your response rate and turnover before you celebrate.

In Sparkbay, report access maps automatically to the org hierarchy, so each manager sees only the teams they are responsible for, subject to the anonymity threshold. Leaders can act locally without exposing individual employees.

Close the loop after every cycle

An anonymous survey still needs a visible response.

Share what you heard, what you'll do, and what you won't do right now. Naming what you're deliberately not acting on is what buys credibility, because employees read silence on a known problem as either denial or inability.

Ask each manager to choose one or two actions, not ten. For psychological safety, the actions that move the score are behavioral rather than programmatic: interrupting the same three dominant voices in meetings, responding to a raised risk without visible irritation, admitting the manager's own errors first.

Be honest about the highest-stakes case. If a team scores low on safety and its own manager is part of the cause, that manager can't credibly lead the follow-up alone, so route those conversations through HR or a skip-level leader instead.

Handing a low-safety team back to the manager who caused it, with instructions to "build an action plan," usually depresses the next score further.

Then use the next pulse to test whether those actions changed the employee experience. Over time, that's what turns the questionnaire from a listening exercise into a management system.

Turning your survey results into action

A psychological safety survey only creates value when leaders use the results to change how teams work.

Most published surveys stop at measurement. The harder problem, the one almost nobody writes about, is what happens in the six weeks after the scores land, when attention fades and the results quietly become a slide no one reopens.

Low scores should trigger a clear, time-bound response with a named owner. Not a vague promise to "communicate better." Treat the survey as the start of a conversation, and use that conversation to pick a few visible actions.

1. share the results with the right level of detail

Share the main themes with each team, including strengths and the areas that scored low. Keep the focus on patterns, not individual comments or suspected sources.

A senior leader might put it like this: "Our lowest score was on whether people feel safe challenging decisions. We need to understand where that shows up in meetings, planning, and escalation."

The most common failure is asymmetric distribution. Executives get the full breakdown, but the team that actually produced a low score never hears about it, so nothing changes where the behavior actually lives.

Cascade ownership along with the data, not just the right to look at it.

The reverse risk matters too. Slice results too finely and you can accidentally point to one person, especially in a team of six or seven, which is precisely the kind of exposure that erodes trust.

If you use a platform like Sparkbay, managers only see results for the groups they are allowed to view, and only once enough people have responded to protect anonymity. Report access is mapped to your org hierarchy so each manager sees only their own teams, and results stay hidden below the minimum response threshold, which is five by default and configurable.

That lets teams discuss the findings without the review turning into a hunt for who said what.

2. ask questions before offering answers

Low psychological safety rarely comes from a single event. It accumulates from repeated micro-signals: ideas dismissed too quickly, risks punished, bad news softened, senior voices dominating the room.

Leaders should model curiosity by asking what would make it easier to speak up. A few questions that tend to open things up:

• "What topics feel difficult to raise in this team?"
• "When do we shut down debate too early?"
• "What happens here when someone makes a mistake?"
• "What is one thing I do that makes it harder to challenge me?"

The last one matters most. When a leader invites feedback on their own behavior, they take on the risk of speaking first, and that is what makes it safer for others to name the real issue.

Be honest about the trap here. Ask these questions in a room that already feels unsafe and you'll usually get silence, which a leader can easily misread as agreement.

If no one answers, the silence is itself a data point, not a green light.

When the room stays quiet, switch channels. Collect written responses, run a follow-up pulse, or ask people to bring one example to the next meeting, so the least confident voices are not forced to go first.

3. admit mistakes and name the behavior you want to change

Employees calibrate their candor to how leaders handle discomfort. If a leader defends every decision, explains away every score, or blames the team for staying quiet, the next survey will not tell the full truth.

So name one or two specific behaviors and what you'll do differently. For example: "I realize I often respond to concerns by explaining the decision instead of asking more questions. In our next project review, I will pause before responding and ask what risks we are missing."

This works because it makes change observable. Employees can watch whether the leader follows through in the next meeting, not six months later.

There's a limit worth naming. Admitting mistakes builds safety only when the system around it stays fair.

If people watch a genuine error get punished the week after a leader preached openness, the confession reads as theater and the score sinks further. Safety is inferred from what happens to the person who took the risk, not from what the leader says about risk-taking.

4. set concrete follow-up steps

Choose a small number of actions for each low-scoring area. One or two well-run changes beat a long action plan that no one owns.

Each action needs an owner, a deadline, and a check-in people can see. If employees can't tell what changed after they gave feedback, they'll discount the next survey.

Favor changes to the mechanics of how work happens over changes to how people are told to feel. A structural fix, a fixed speaking order or a standing agenda item, keeps working even when a busy leader forgets to be intentional.

It removes the dependency on a manager's mood in the moment.

And watch the second-order effect of a low score on "challenging decisions." The instinct is to add more open discussion, but if the underlying problem is that objections never change the outcome, more meetings just breed more cynicism. What actually helps is a documented case where an objection altered a decision, and saying so out loud.

5. close the loop and measure again

After a reasonable period, usually one to three months for team-level behavior changes, ask whether the action made a difference. Use a short pulse or a structured team discussion.

Then report back: what improved, what didn't, and what the team will try next. That's the difference between safety as a running practice and safety as an annual survey topic.

Expect a counterintuitive result. Scores sometimes dip on the second measurement.

As people come to believe leaders genuinely want the truth, they surface issues they'd previously kept quiet, so a short-term decline can mean the process is working. A flat score with rising comment volume reads the same way.

For that reason, judge progress over several cycles, against comments and behavior rather than a single number. In Sparkbay this is the engagement score out of 10, tracked over time alongside employee comments, so you can separate a real shift from normal noise between surveys.

The benefits of a robust employee survey platform

A robust employee survey platform such as Sparkbay can help leaders understand obstacles and opportunities and focus their efforts where it matters. It can also track engagement and satisfaction over time to identify which actions are working and which aren't.

Explore employee data any way you want

It is important to be able to segment data based on demographics such as tenure and role. Sparkbay allows leaders to interrogate their data in any way they want to uncover deeper insights about their people.

Empowers action

Surveying employees serves no purpose if leaders do not take advantage of the opportunities they discover. With Sparkbay's automated coaching, managers can act on easy-to-implement strategies. These tools drive action to ensure you meet your employee-centric objectives. 

Provides robust benchmarks

Leaders need to compare their survey results with industry benchmarks to understand how they are doing. Using the right survey platform, you can compare your survey data with others in your industry, similar-sized companies, and top employers. 

Equips managers with dashboards to visualize their results

Manager dashboards help managers visualize their survey results all in one place. With these dashboards, managers can easily track employee experience trends, and identify strengths and opportunities.

Provides a seamless user experience

You don't want your managers to waste valuable time on a clunky, hard-to-use platform. We designed Sparkbay with simplicity in mind. Your managers will master our intuitive dashboards in no time.

Gain access to real-time results

Taking action is the most significant part of your survey strategy. That’s why Sparkbay's real-time results let you expedite your process and propel insights into action.

Using the right survey software

By leveraging the right survey software, you can quickly understand where your employees stand when it comes to psychological safety in the workplace. Sparkbay’s employee survey tool makes it easy for organizations to distribute pulse surveys and track data while keeping all employees’ responses anonymous.

Get in touch with a member of our team to learn how Sparkbay’s survey platform can help you build a better workplace through our employee surveys.

If you're interested in learning how Sparkbay can help you survey your employees, you can click here for a demo.