Bethany Morgan-Brett

Anonymising data and gaining consent retrospectively

Why read this account

To find out how one researcher who was preparing her doctoral study for archiving approached anonymising the data years after it was originally collected.

The challenge

When Principal Investigator Bethany Morgan-Brett began to prepare her doctoral study for archiving eight years after the original research, she had to devise a system for anonymising the data. She also sought retrospective permission from four respondents whose personal details had been destroyed in line with the Data Protection Act.

About the research

This is a doctoral study examining the subjective experiences of ageing, with a particular focus on mid-life. The research data is located in the UK Data Service collection: Negotiating Midlife: Exploring the Subjective Experience of Ageing, 2006-2008.

A psycho-social approach was used, based on Hollway and Jefferson’s free-association narrative interview (FANI) method. Hollway and Jefferson (2000) was used in conjunction with the raw data from their study in the UK Data Archive collection: Gender Difference, Anxiety and the Fear of Crime that gave rise to the FANI method.

Accessing research notes, original interview schedules and a failed pilot exercise allowed Principal Investigator Bethany Morgan-Brett to understand how the unique psycho-social method had evolved from the need to factor in the unconscious defensive dynamics arising in interviews. The data from this study were invaluable in helping her to think through the unconscious psychic processes of her own participants.

About the data

Morgan Brett interviewed 22 men and women aged between 39 and 58 years old, looking at three specific aspects of ageing: (i) the ageing body, (ii) inter-generational relationships through the life course and (iii) life reviews. Data were collected in the UK between September 2006–January 2008.

This cross-sectional study offers textual data arising from face-to-face interviews: a data list, user guide, interview transcripts and Study Information Data Processing Notes.

Archiving challenges

Data handling

At the time, there was no realistic opportunity to fully prepare the data from the doctoral research for archiving.

For example, transcribers had no specific template to work from, which meant that the transcripts were not consistent in their presentation and would not be shareable. File names were also ambiguous and inconsistent. Although anonymisation was done on a micro level for excerpts from transcripts used in the thesis, the full dataset was more haphazard and not ready to be shared confidentially. Anonymising place names were particularly challenging when trying to retain something of the character of the location and it was decided eventually to widen the description; thus for example Welshpool might
become {Welsh border town}[1]

In 2014, the University of East London awarded funding and ethical approval to bring all Morgan-Brett’s previous research focusing on intergeneration relationships between adult children and their ageing parents together under an umbrella project: Impossible Choice[2], along with the creation of a new set of empirical data. Additional funding was also sought for preparation of data from the doctoral study for archiving. Effective data management is something that should be planned from the outset and revised throughout a research project. Processing an entire collection can be costly and extremely
time-consuming.

  1.  All anonymisation changes were placed within curly brackets {…}. All brackets styles were changed when preparing data for archiving to maintain consistency: rounded brackets to indicate extraneous sounds in the interview (coughs) and square brackets to indicate additional words in the transcript i.e. [end of interview].
  2. https://theimpossiblechoice.wordpress.com

For the purposes of preparing the data for depositing in the Archive, the 2014 funding supported a research assistant working for 100 hours on this project and the researcher herself probably matched this amount of time.

Firstly data were organised and assessed to identify the nature and extent of changes, involving establishing consistency of themes and style. Secondly, clear instructions were given to the research assistant on how to broach systematic anonymisation of all data, administrative as well as research. All names were changed to pseudonyms, ensuring these were consistent with those used previously in publications where applicable. All place names were changed to a generic descriptor, such as {local high school}, {Northern town} or {Company name}. As age was a key area under investigation, all ages remained unaltered. The original intention of keeping a log of all changes to data was adapted for simplicity to recording the first time a change was introduced.

Ethics and consent

During the doctoral research period it became clear that there were confidentiality issues in the transcription process, when the researcher became aware that interviews had been read by the partner of a transcriber. Clearly, confidentiality agreements needed to be extended to those involved in processing the data. (See our template of a transcriber confidentiality agreement)

During the post-doctoral period of collating data in preparation for re-use and archiving purposes, it was obviously necessary to seek retrospective consent to share data. At the time of the PhD data collection, data sharing was considered after only the first four interviews. Subsequent consent forms extended the original statement (that data would be used in the thesis and future publications) to state that it could also be archived or re-used for other research or teaching purposes. The remaining 18 participants signed the form that included this clause.

This meant that Morgan-Brett was re-contacting the four remaining participants eight years after being interviewed. The Data Protection Act requires that
personal data, such as names and addresses, should be destroyed once they are no longer required. Researchers should therefore think carefully about future plans and whether such personal data may be required for the purposes of gaining retrospective consent, for data sharing for example.

As Morgan-Brett had destroyed personal data after the completion of the thesis, access was ultimately more challenging than gaining secondary consent. ‘Janet’ was easily located and agreed to sign the altered consent form. Attempts were made to contact ‘Angela’ through Facebook, but this was unsuccessful. No contact details had been kept for one male and his data are being kept under strict “no release” at the Archive. ‘Ray’ was approached via his place of work, but he had retired. The organisation would not release any contact details, but did provide his surname. He could then be traced by Facebook and happily consented to share his data.

At a glance: Lessons learnt from the project

  • Plan how the data should be managed from the outset and revise this throughout the project.
  • Processing an entire collection retrospectively can be costly and extremely time-consuming.
  • Researchers should think about how the data could potentially be used in the future.

Reuse and additional outputs

During the process of re-checking the data for storing with UK Data Archive, Morgan–Brett re-coded all interviews from her thesis via Nvivo, looking specifically at the theme of inter-generational relationships. This contributed to the Impossible Choice study, focusing on decisions of care for an older parent.

In addition to the data originating from her thesis, there were ten in-depth interviews with adult child relatives (aged 45-68), four interviews with Community Visitors and four care home managers; and eleven additional interviews with older people (aged between 79 and 100) in residential care in two care homes in Essex.

Read more

Read UK Data Service guidance on anonymisation