The UKAN decision making framework is a helpful tool for sorting through the issues of making data anonymous.
Replacing identifiable information, e.g. names, addresses, reference numbers that may be traceable to an individual, with a random unique identifer in the dataset is one important aspect of making data anonymous, but it’s not the whole story. UKAN’s advice is very good – to “know your data”, and go through each field and consider the potential risks. It’s also important to think about the risks of fields in association with other ones – e.g. one field, like the sex of the individual, might not mean much but it could be risker in combination with another variable. This dataset is a good example of anonymised data – a “couple ID” field has been made for each couple so they can be identified for the sake of data analysis but there are no personal details about them.