Consulting a pre-existing advisory group about a research project on ethnicity data

The ‘why’

A challenge of using administrative data for research is that sociodemographic data, like ethnicity, is not always consistently recorded between different public sector organisations. This means that, in linked administrative datasets, the ethnicities recorded for individuals sometimes differ between data sources. As part of my ADR UK fellowship, I investigated similarities and differences in how ethnicity was recorded in a data linkage between the Police National Computer (PNC, Ministry of Justice) and the National Pupil Database (NPD, Department for Education). Before I started writing the final report, I decided to hold a focus group with members of the public on the emerging findings to get their perspective on the results. This was to help me shape recommendations which could be made as a result of the work. 

The ‘who’

I consulted with the National Institute for Health and Care Research Maudsley Biomedical Research Centre (NIHR Maudsley BRC) Race and Ethnicity Advisory (READ) Group. The advisory group meeting was attended by eight members of the READ Group, plus a facilitator. 

The ‘where’

NIHR Maudsley BRC is based in South London. I met with the READ Group online in September 2022. 

The ‘how’

When I first approached the READ Group, the facilitator sent me a guidance document. I then met with the facilitator to discuss what I wanted to present to the group, and what I wanted to get out of the focus group. I emailed the slides which I would be presenting to the group, and the following executive summary, to be circulated to the group in advance of the meeting: 

In this project, I am exploring a new dataset containing linked routinely-collected education and criminal justice data. The purpose of this analysis is to understand how ethnicity is recorded in this dataset, identify possible inconsistencies or inaccuracies, and understand whether these issues might impact our identification of ethnic inequalities. Based on these findings, we might be able to make some initial recommendations to researchers who use these data, and to the data owners who collect this data. 

At the meeting, I spent approximately 20 minutes presenting the project and emerging findings. The remainder of the one hour meeting was used for group discussion. I gave the group the following discussion prompts: 

• What can we recommend to researchers in terms of how to analyse ethnicity fields as they currently stand in these datasets? 
• What can we recommend to data owners in terms of how ethnicity data is collected and shared? 
• Are there further analyses which I need to do in order to make some tentative recommendations? 

Successes

The focus group raised concerns about the associations I presented between ethnicity, educational outcomes, and criminal justice outcomes. The intention of reporting these associations was to illustrate how discrepancies in ethnicity recordings between different public sector data sources could subsequently impact statistical analyses making use of these data. However, the study was not designed to accurately investigate these associations. As a result, the group expressed concern that presenting these associations, however well-intentioned, might mislead readers as to the purpose and meaning of study. Even with caveats, the associations could be interpreted as accurate and meaningful in their own right, and could in turn risk further disadvantaging certain communities. This feedback led me to change my approach to these analyses, ensuring that the focus of the report was only on discrepancies in ethnicity recordings between data sources, rather than how resulting ethnicity variables were associated with educational and criminal justice outcomes. 

This focus group was therefore essential for shaping the findings which I presented in my final report. The group also suggested recommendations which could be made to researchers and data owners regarding the collection and analysis of ethnicity variables in these datasets. These suggestions informed the recommendations which I presented in my final report. 

Lessons learned

Engaging members of the public can highlight possible interpretations of findings which researchers may have overlooked. In this case, consulting with the READ Group prevented me from presenting findings which could have misrepresented the purpose of the project. The group were also essential in helping me to shape the recommendations made to researchers and data owners. 

The practicalities of seeking input from the public can be challenging, but liaising closely with people who have expertise in public involvement and engagement can help overcome these challenges. Meeting with group facilitators in advance of the meeting is important for understanding the dynamics of the group, how the session should be run, and what to expect.  

Planning your public engagement in advance is also essential. I met with the READ Group approximately two months after initially reaching out to them. I had hoped to conduct further consultation with other existing groups, but was unable to do so before the final report deadline. The earlier you can reach out to members of the public to schedule a session, the better (although sometimes the unpredictable nature of research projects can make this difficult).