What does it take for a research project to be anonymous from the very start?

Anonymous data collection is a good and effective method of data collection, provided it is a methodologically good fit for your project. However, it is easy to make mistakes.

What is not anonymous data even if it may appear that way?

You process personal data if any of the following points apply to your project. You therefore need to obtain consent from your informants and you need to notify NSD Data Protection Services about your project.

• Indirect personal data can have great potential for revealing the identity of your informants. Even if you do not ask for name and contact details, combinations of data such as gender, age, height/weight, occupation, field of study, geographical affiliation, hobbies/sports, family relationships, etc. may make it possible to identify individuals.
• Electronically collected data with hidden ID: for example cookies, IP addresses of PCs and IMEI numbers of smartphones. This can be difficult to control and this is why only ‘Nettskjema’ can be used a survey tool.
• Repeated questioning or testing of the same informant, e.g. before and after an intervention, requires a link key that will allow the identification of individuals in the dataset. Even if you or others keep the link key separate to the datasets, the data is not considered anonymous but de-identified. You will be processing personal data for as long as the link key exists and you need to ensure that all data protection requirements are met. When the project is complete, the link key must be deleted and other indirect personal data must be removed. This constitutes anonymisation of the data material, which can subsequently be presented to the world.
• Third-party data: Identifying data relating to other individuals that the informant may provide during questioning.

What is the best approach to ensure anonymous collection of data?

1. Conduct a non-recurring survey without registering any personal data and without creating a link key. This means you will not know who has responded to your survey. This can be done in two different ways:

• A paper survey without entering names or answering any questions that could directly or indirectly reveal the identity of the informant. You subsequently plot the data in a suitable format, such as Excel. 
• Electronic survey: You need to prevent informants from leaving visible or hidden electronic identifiers behind. Nettskjema includes settings that can safeguard such anonymity and is the only permitted survey solution.

[link]  Create an anonymous survey

Avoid pitfalls!

• Be particularly aware of hidden electronic traces and remember that combinations of background variables can identify an informant (cf. the introduction).
• The fewer people that have been invited to complete a survey, the easier it will be for the person reading the responses to link the answers to an individual. Ensure that you get at least ten responses.
• If any questions need to be answered using text, it may be a good idea to be aware that dialect, slang and wording may weaken anonymity.
• Use closed questions: One way in which to ensure comparable answers, simplicity and neutral language could be to use closed questions with predefined response options.
• If you have created a paper survey, remember that handwriting can be very revealing.

Your informants may be less inclined to participate if they doubt your ability to safeguard their anonymity. They will either discontinue the survey, fail to answer truthfully or moderate themselves.

Anonymous interviews: preventing the collection of personal data

Prerequisite: Voice data is considered personal data because the informant can be recognised based on the audio recordings of their voice even if no other personal data is provided. Audio must not be streamed to or stored on digital devices that are connected to a network.

Applies to: Small projects such as student projects with one/few informants. The topics covered in the interview should avoid sensitive topics to the extent possible.

Approach:

1) Make sure that you obtain informed, verbal consent.

• Draw up an information letter explaining the purpose of the interview for your informants.
• The informant must be familiar with the content of the information letter before the interview starts.
• The informant must provide verbal consent. This is done by turning up for the interview as they will then be consenting to participate.
• Let them know that it is possible for them to stop the interview without providing any reason.
• Explicitly ask the informant not to address third parties by name or position during the interview.
• No lists containing names or other personally identifiable data must be kept, on paper or electronically.

2) Collection of data

• First choice: Write notes on paper during the interview and refine the notes later on.
• Alternatively: Make audio recordings using an analogue device such as a cassette player or analogue dictaphone: i.e. ensure that recordings and playback do not take place using digital devices such as PCs, tablets or mobile phones.
• Audio recordings must be destroyed or deleted immediately after transcription.
• Transcription of interviews can be performed on a PC as ordinary text processing (using Word for example), provided that no names or other personal data are reproduced.
• Any identifiable data that still emerges during the interview despite the measures that have been taken, must not be reproduced in references or transcripts.

3) Analytics and reporting

• Assign nicknames or numbers/letters to the informants.
• Do not reproduce any place names, workplaces or other indirect personal data.
• Describe the interview procedure in your student thesis in order to demonstrate that data protection has been considered as part of the project and that collection has taken place anonymously (method with low data protection burden selected).