An engagement survey is only as useful as the honesty of its answers, and people answer honestly only when they believe nobody can trace an answer back to them. Many surveys promise anonymity and then show a manager a breakdown for a team of three. Everyone in that team knows exactly who wrote what.
This guide covers what anonymity has to mean for the people answering, and the practical rules that protect it.
Anonymity is a property of the data, not a promise
There is a difference between a survey that hides who answered and one that never records it. If the system stores the respondent and then hides the name in the report, the link still exists. An administrator, an export, a support request or a future feature can bring it back.
The stronger approach is to never write the respondent next to the answers at all. The system records that someone was invited and whether they responded, so it can show a response rate, but the answers themselves carry no name.
Ask any survey tool you consider a direct question: is the respondent stored with the answer and hidden, or never stored?
Small groups break anonymity
Even with no names stored, a report can identify people through arithmetic. If a department has two people and one of them is on leave, a “department average” is one person’s answer.
Two rules prevent this:
- A minimum group size. No figure is shown for a group with fewer than five responses. Five is a common threshold because a smaller group lets people guess who said what.
- No subtraction. If one group is hidden but its total can be worked out by subtracting the visible groups from the overall figure, the hidden group is effectively visible. A careful tool hides enough to stop that.
Comments need the same care. A single written comment, shown with the writer’s team, is often recognisable by its wording alone. Showing comments only when several people wrote one, in random order and without team labels, keeps the writing from pointing back at its author.
Wait for the survey to close
Live results are tempting, but they leak. If a manager checks the results while a survey is open, and then checks again after one person says “I just filled it in”, the difference is that person’s answer. Showing anonymous results only after the survey closes removes that opportunity.
Say what happens to the answers before people answer
Trust also depends on what people are told up front:
- whether the survey is anonymous or named, stated at the top;
- who will see the results, and at what level of detail;
- whether written comments may be summarised by AI, decided before the survey opens rather than after.
Changing any of these after people have answered breaks the promise they answered under.
Ask fewer questions, more often
A short survey people finish is worth more than a long one they abandon. A handful of agree or disagree statements, one question about whether they would recommend working there, and an optional comment box is enough to see a trend. Running the same core questions each time is what makes the trend meaningful.
How astraHR handles this
astraHR’s engagement surveys follow these rules by design. In an anonymous survey the respondent is never stored with the answers. No result is shown for fewer than five responses, groups that could be worked out by subtraction are hidden as well, and anonymous results and comments appear only once the survey has closed. Whether comments may be summarised by AI is settled before the survey opens. Surveys are part of the Complete plan.