How to design a DEI survey people actually trust

At a glance

A DEI survey and a general culture survey are not the same instrument.

This guide is about the DEI survey specifically: identity, equity, and belonging questions, where a badly worded item can cause real harm, not just a bad response rate. If you are measuring broader organizational health instead, our culture survey strategy guide covers cadence and business impact.

Most DEI survey advice stops at a list of questions to copy and paste. That is not enough. The harder problems are getting people to answer honestly, protecting anyone who does, and proving the data actually changed something. Here is how to handle all three.

Before you write a single question, decide what decision this data will inform.

A DEI survey that exists to check a box gets treated like one by the people answering it. A DEI survey tied to a specific decision, like revising a promotion process or redesigning onboarding, gives respondents a reason to tell you the truth.

Write that purpose down in one sentence and keep it visible while you build the survey. If a question does not serve that purpose, cut it.

This single habit prevents the most common DEI survey mistake: a long, unfocused instrument that measures everything and clarifies nothing.

This is also where you decide who owns the results. Name the person or team accountable for reviewing findings and driving action before launch, not after the data arrives. A DEI survey without a named owner tends to sit in a shared drive, and employees notice when nothing visibly changes. Most HR teams manage this work from a broader HR survey program, with the DEI survey as one recurring piece of it.

Leadership sometimes treats a DEI survey as a reporting requirement rather than a decision-making tool. Reframe it early: this data exists to find specific, fixable gaps in how people experience advancement, recognition, and voice at your company. A DEI survey that leadership sees as strategic gets funded, staffed, and acted on. One that gets filed under compliance gets neither.

Diversity counts headcount. Inclusion, equity, and belonging measure something different: whether people believe the system treats them fairly and whether they feel safe being themselves at work.

Your DEI survey needs distinct questions for each, because a company can be diverse on paper and still fail every one of these tests.

Use a consistent agree or disagree scale for statements like these, so results are easy to compare over time:

Each statement targets a different mechanism. Belonging is emotional, equity is structural, and inclusion is behavioral. Blending them into one vague "how included do you feel" question makes the results impossible to act on.

Standard demographic checkboxes miss people whose identity does not fit the options you listed. Add an optional self-description field alongside standard categories, and make every identity question skippable. A DEI survey that forces a demographic answer before someone can continue will lose respondents, and it will lose the ones you most need to hear from.

Use these as a starting point, then adjust the language to match how your organization actually talks:

Notice that none of these ask someone to label an experience as discrimination outright. They ask about specific, observable situations instead, which produces more reliable data than a single abstract question ever could.

If you want a fully built starting point instead of writing every question from scratch, SurveyMonkey publishes a belonging and inclusion survey template developed with behavioral science researchers.

This is where most DEI survey guides stop short, and it is the part that determines whether your data is trustworthy at all. Getting this wrong does not just hurt response rates, it can make people who are already vulnerable at work feel more exposed, not less.

Use SurveyMonkey anonymous response settings so individual answers cannot be traced back to a respondent, and say so explicitly in the survey introduction. Our guide to building an anonymous survey walks through the setup step by step.

Anonymity that exists in the settings but is never explained to respondents does nothing, because people answer based on what they believe is true, not what is technically configured.

Ask sensitive identity and experience questions before you ask for demographic detail, not after. If someone answers a hard question about discrimination and then immediately has to identify their department, race, or tenure, they may go back and soften the earlier answer. Sequencing protects candor.

Run a pilot with a group of employees who represent different roles, tenures, and backgrounds before launching the full DEI survey. Watch for three things:

That third point matters more than people expect. A small team where only one person fits a demographic combination is not actually anonymous, no matter what the settings say. Adjust question groupings or suppress small-cell reporting to protect against this.

Page and question skip logic lets you route respondents past sections that are not relevant to their role, tenure, or location, which shortens the survey and reduces frustration.

It also matters for harm reduction specifically: a manager-only question about promotion decisions should never appear for someone who has never managed anyone, since it forces an irrelevant, sometimes uncomfortable guess.

Configure branching before launch, then confirm it during your pilot.

Raw agree or disagree percentages tell you very little on their own. The real signal in a DEI survey comes from comparing groups and reading the open-ended comments closely, because that is where people explain the "why" behind a score.

Filter results by team, tenure, and (where sample size allows) self-reported identity groups to spot gaps that an aggregate score would hide. A company-wide belonging score of 80% can still hide a subgroup sitting at 45%, and that gap is usually the most important finding in the entire survey.

SurveyMonkey Analyze features, including sentiment analysis and thematic analysis, can surface recurring themes in open-ended comments without you reading every response line by line.

A DEI survey that produces a report nobody sees is worse than not surveying at all, because it confirms the exact skepticism you were trying to overcome.

HR professionals report this concern constantly: employees doubt that feedback stays anonymous or leads to real change, and every unaddressed survey reinforces that doubt for the next one.

This is what actually rebuilds trust for the next survey cycle, not a better subject line.

Even well-designed surveys get undermined by a handful of repeatable errors. Watch for these before you launch:

Most of these mistakes trace back to the same root cause: treating the DEI survey as a form to fill out rather than an instrument that has to be tested and refined like any other piece of research.

Designing a DEI survey that people trust enough to answer honestly is a methodology problem before it is a tooling problem, but the right platform makes the methodology easier to execute. SurveyMonkey anonymous response settings, question-level skip logic, and AI-powered thematic and sentiment analysis are built to support exactly the sequencing, anonymity, and analysis steps covered in this guide.

See how to design a DEI survey employees trust enough to answer honestly with SurveyMonkey inclusion and belonging solutions, built on 90+ HR-vetted templates and anonymity controls your team does not have to configure from scratch.

See how to design a DEI survey employees trust enough to answer honestly.

If your organization has not run any recurring employee listening program yet, start smaller.

Start with an employee engagement survey.