Anonymity in employee surveys means that responses can never be associated with the person who gave them. There are three basic rules for protecting anonymity: never report results for groups below a set minimum size (5 people in most organizations), avoid combinations of demographic questions that could reveal identities, and strip open-ended comments of anything that could identify the writer. In surveys that collect demographic data or use personalized links, the data may count as personal data, so obligations under KVKK (Turkey's Personal Data Protection Law) also need to be planned before the survey launches.
- In an anonymous survey, responses can never be linked to an individual; in a confidential survey, linking is possible, but only the party running the research sees the responses.
- The things that most often expose identities are small reporting groups, combinations of demographic questions, and open-ended comments.
- The most basic rule is never to report results for groups below a set minimum size (5 people in most organizations).
- In surveys that collect demographic data or use personalized links, the data may count as personal data, so KVKK (Turkey's Personal Data Protection Law) obligations should be planned from the start.
Why does anonymity matter so much?
When employees worry that their answers might be traced back to them, one of two things happens: they either skip the survey or give "safe" answers instead of their real opinions. Both make the results worthless. Especially on sensitive topics such as managers, pay, fairness and culture, the precondition for honest answers is that employees genuinely trust the anonymity.
That trust is built not only through technical safeguards but also through what happened in previous surveys. In an organization where the perception has taken hold that "they found out who gave which answers," it can take years to win back participation and honesty in later surveys. We also discussed this from a participation angle in our post on giving the silent majority a voice.
The difference between anonymous and confidential surveys
The two concepts are often confused, but they are different things:
| Anonymous survey | Confidential survey | |
|---|---|---|
| Are responses linked to the individual? | No, never | Yes, via an ID number |
| Who sees individual responses? | No one | Only the party running the survey (usually an independent firm) |
| Where does demographic information come from? | The employee's own self-reporting | Automatically from the HR system |
| Advantage | Highest level of trust | Fewer questions, more accurate demographic data, individual-level trend analysis across repeated measurements |
| What to watch out for | Employees need to answer demographic questions accurately | The party doing the matching must be independent, and the organization must have no access whatsoever to individual data |
In both models, the organization should see aggregate results only. Telling employees clearly which model the survey uses is far better than having them discover later that a survey labeled "anonymous" was actually confidential.
How do identities get exposed?
Anonymity is usually compromised not by deliberate violations but by overlooked details:
- Small groups: The results of a 3-person team make it easy to guess who answered what.
- Combinations of demographic questions: When department, tenure, age range and gender are asked together, each question may be harmless on its own, yet their combination can point to a single person. A group such as "finance department, over 10 years' tenure, over 50 years old" often contains just one person.
- Inference by subtraction: If a department's overall result for 23 people is reported alongside the result for a 20-person subgroup of the same department, the answers of the remaining 3 people can be calculated.
- Open-ended comments: Comments describing a specific incident, project or person, as well as writing style and habits, can reveal identities.
- Technical traces: Personalized survey links, IP addresses and response times eliminate anonymity if the organization has access to them.
- Paper surveys: Handwriting, and to whom and where a survey was handed in, can reveal identities.
8 measures to protect anonymity
- Set a minimum reporting group size: Do not report results for groups below a set number of people. The common threshold is 5; for sensitive topics such as manager evaluations, discrimination or well-being, it can be raised to 7–10.
- Prevent inference by subtraction: When reporting a group's results, make sure the differences between that group's sub-segments do not expose a group below the threshold. The reporting system should check this automatically.
- Limit demographic questions: Ask only for the breakdowns you will actually use in the analysis. Ask about age and tenure in broad ranges rather than exact values.
- Clean up open-ended comments: Before reporting, remove names, job titles, project names and any details that could identify a person. Present comments to managers grouped by topic, not by individual.
- Block access to individual responses: No one in the organization, including the HR manager and the CEO, should be able to access individual responses. Any technical traces (links, IP addresses, timestamps) should remain solely with the party running the survey.
- Have an independent party run the survey: Knowing that responses are collected by a firm outside the organization noticeably increases employees' trust. We compared this option in our post on whether to choose a research firm or software.
- Use sealed envelopes and boxes for paper surveys: Make sure surveys are handed in to sealed boxes, not to managers.
- Explain the rules in advance: At the start of the survey, state clearly what data is collected, who can see what, and the minimum reporting group size.
What should you tell employees?
An anonymity assurance works when it is clear and concrete. In the survey invitation and on the welcome page, answer these questions:
- Who can see my responses?
- Can my manager find out how I answered?
- What is the minimum group size for which results are reported?
- How will my open-ended comments be used?
- What demographic information is collected, and why?
Saying "Our survey is completely anonymous" is not enough. Saying "Your responses are collected by an independent research firm; no one in the organization can access individual responses, and results are reported only for groups of at least 5 people" is far more reassuring.
What to watch for under KVKK
This section is for general information only and does not constitute legal advice. We recommend consulting your legal advisor for an assessment specific to your organization.
Law No. 6698 on the Protection of Personal Data (KVKK) protects data relating to identified or identifiable individuals. The key points for employee surveys are:
- Anonymous data vs. personal data: The law defines anonymization as rendering data impossible to associate with an individual, even when matched with other data. Truly anonymous data does not count as personal data. However, responses collected via personalized links or matched with HR data, and data that makes a person identifiable through demographic combinations, may qualify as personal data.
- Duty to inform: If the data qualifies as personal data, employees must be told at the time of collection which data is processed, for what purpose, on what legal basis, to whom it may be transferred, and what rights they have.
- Legal basis: The processing condition being relied on should be determined in advance. Because it can be debatable whether consent in an employer–employee relationship is truly freely given, this assessment should be made with a legal advisor.
- Special categories of data: Data such as health, religion and ethnic origin are classified under KVKK as special categories of personal data and are subject to stricter rules. Extra care is needed if such data is collected in well-being and mental health surveys; where possible, frame questions around work-related conditions rather than diagnoses and health status.
- Data processor relationship: If a firm runs the survey, the organization may act as the data controller and the firm as the data processor. The parties' responsibilities, data security measures and how the data will be destroyed at the end of the project should be set out in a written contract.
- Transfers abroad: If the survey platform's servers are abroad and the data qualifies as personal data, the transfer is subject to the conditions in Article 9 of KVKK. This article was amended in 2024; ask the provider which safeguard it relies on.
- Retention period: Determine how long the data will be kept and how it will then be deleted or anonymized.
Pre-survey checklist
- The minimum reporting group size has been set and announced to employees.
- The reporting system prevents inference by subtraction.
- Only demographic information that will be used in the analysis is being collected.
- There is a process for cleaning up open-ended comments.
- No one in the organization can access individual responses.
- Employees have been told clearly whether the survey is anonymous or confidential.
- Under KVKK, the privacy notice, legal basis and, where needed, the data processing agreement are ready.
- It is clear where and for how long the data will be stored.
When planning the survey as a whole, you can work through these steps together with our guide on how to create an employee engagement survey.
Frequently asked questions
Are employee surveys really anonymous?
Yes, if they are designed properly. That requires responses not being linked to personal information, results being reported only for groups above a set minimum size, and identifying details being removed from open-ended comments. Having an independent firm run the survey makes it easier for employees to trust this assurance.
What is the difference between an anonymous and a confidential survey?
In an anonymous survey, responses can never be associated with an individual. In a confidential survey, responses can be matched to the person's demographic information via an ID number, but only the party running the survey sees individual responses, and the organization receives aggregate results only.
What is the minimum group size that should be reported in an employee survey?
Common practice is at least 5 people. For sensitive topics such as manager evaluations, discrimination or well-being, this threshold can be raised to 7–10 people.
Is explicit consent required under KVKK for an employee survey?
It depends on how the survey is designed. Truly anonymous data falls outside the scope of personal data protected by KVKK (Turkey's Personal Data Protection Law). However, if demographic information is collected or personalized links are used, the data may count as personal data; in that case, at a minimum the duty to inform must be fulfilled and the legal basis for processing must be determined. Stricter rules apply to special categories of data such as health. Always consult your legal advisor about your own situation.
Is it a problem if the survey software's servers are located abroad?
If the data collected qualifies as personal data, transfers abroad are subject to the conditions in Article 9 of KVKK. Before you start working with a provider, get written confirmation of where the data is stored and which legal safeguard the transfer relies on.
