De-identification
Also called anonymisation, pseudonymisation
Removing or replacing the details that could identify a respondent, so what you hold cannot be traced back to a person.
Anonymous data cannot be linked back to an individual. Pseudonymous data replaces direct identifiers with a code while a linking key still exists, as often happens in longitudinal studies. Describe that distinction when explaining your data collection.
Identifiers are not only names and emails. Free-text answers volunteer details, IP addresses and device fingerprints identify, and small cells re-identify by combination: department plus tenure plus office is often one person. Every demographic question narrows the crowd somebody is hiding in.
Collect only the details you need. Review combinations of demographic fields and open-text answers as well as direct identifiers, and explain what you collect to respondents.
Where it goes wrong
Small reporting groups can make people identifiable, especially when readers already know the population. Review the risk before publishing a breakdown.