A survey without a name field is not necessarily anonymous. A response may contain an email address in free text, a distinctive job title, or a combination of details that points to one person. The promise of anonymity therefore depends on the information and its context, including what other information could be used alongside it.
Confidential research can still know who a participant is while restricting access and use. Pseudonymisation replaces identifiers with codes or similar substitutes, but a retained link back to an individual means the researcher has not made that data anonymous simply by changing the label.
Consider what a reader could recognise
The Information Commissioner’s Office’s guidance on effective anonymisation considers identifiability in context, including information that could be combined and means reasonably likely to be used. It does not frame anonymity as removing names or eliminating every purely theoretical possibility of identification.
For an illustrative workplace study, “the only overnight dispatcher at our northern depot” may be identifying to colleagues even without a name. Replacing that description with “an operations employee” reduces detail, but the surrounding incident may still be recognisable. Review the whole passage and its intended audience.
Recordings deserve particular attention because faces, voices, screens, and background conversation can reveal identity. A blurred name on the first screen does not address every identifying detail elsewhere in the session.
Separate what the study needs to keep
Recruitment and payment may require contact information even when research responses can be collected separately. Design the connection between those records deliberately. A separation that prevents analysts from identifying respondents may improve protection, but describe precisely who can still reconnect the information.
Free-text answers need a different review from fixed-choice responses. People can volunteer details the questionnaire never requested, so a claim about anonymous collection should take that possibility into account. The same applies when several datasets are combined later.
Make a promise the arrangements support
Explain the protection in the consent process before people decide what to disclose. If the team will know their identity but restrict access, describe confidentiality and the actual limits rather than using “anonymous” as a reassuring synonym.
Reassess the material when its audience or purpose changes. A report appropriate for a small research team may need different treatment before wider circulation. Preserve the meaning needed for the finding while removing unnecessary identifying detail, and record any edits to quotations so readers are not presented with altered wording as an exact original.