Using data to drive improvements in race equality in the public sector
Earlier this year, a blogpost made the case for high quality data on ethnicity to understand and address inequalities in the public sector. While we are sympathetic to that call, our research suggests that data quality in itself is not enough.
The context for our research (paywall/academic login) was the call for public bodies to “explain or change” in accounting for disparities in outcomes on the basis of class, gender, and ethnicity in the main. One of the most prominent outcomes of this policy initiative was the government’s Race Disparity Audit. This initiative has resulted in the publication of a significant amount of data on racial inequalities on a dedicated Government website. By publishing data more openly, the Government hopes that public sector agencies will be encouraged to ‘explain or change’ their progress on race equality more effectively.
But what does it mean in reality for staff in public sector organisations to either ‘explain’ or ‘change’ figures and policies around race / ethnicity in particular? To address this question we conducted a study to explore this in more detail. Given the Race Disparity Audit’s aim to use public sharing of ethnicity data as a lever/lens to help or require public agencies account for patterns of inequality and to propose remedies, we wanted to examine whether staff in the public sector feel equipped to do this. Through surveys and interviews with health and higher education institutions, we asked: what do people working in public agencies think about the quality of ethnicity data? Is it helping them to understand discrimination and plan the design of more equitable services?
We identified three main themes in our findings (paywall/academic login). Firstly, respondents from both higher education and healthcare have some concerns about the quality of ethnicity data available to them and, at times, this prevents them from using the data effectively in their work. Problems with data quality relate to inconsistency or poor application of systems used to capture data and associated low rates of disclosure. Problems also relate to limited granularity of ethnic categories that restricts meaningful analysis.
Secondly, only about half (and in some cases less than half) of respondents feel that ethnicity data is effective or very effective in helping their organisation to understand discrimination and plan and delivery services. Some felt this was because of the poor quality of data and restrictions associated with analysis (e.g. challenges undertaking more fine-grained analysis combining factors such as ethnicity, gender and age). Others also felt that the challenge related to improving people’s level of understanding and skill in conducting data analysis and interpreting results.
Thirdly, there is a lack of attention to ‘why’ data is collected. Numerous respondents identified lack of confidence in the process of data gathering as an important reason for low levels of disclosure. There is a lack of trust in whether the data will be confidential, as well as a lack of awareness of how data will be used. Some suggested that the collection of data can also be seen as an ‘end in itself’ and decisions are not always made strategically about which data to collect or how that data will be used to improve services.
Although there were a number of calls from participants for better data, questions still remain about what organisations will do with better quality data on racial inequality in the future. We heard that more open data would lead to better understanding and better policies. Yet, at the same time, we also heard that a number of fundamental challenges exist which are restricting progress on this agenda. There are gaps in technical skills and knowledge required to analyse such data and use it to change practice. But more than this, there are also broader organizational culture and attitudinal issues related to limited willingness, ability or comfort in discussing race issues.
If public agencies are to use data on racial inequality to ‘explain’ those racial inequalities, then they will need to engage with some of the, sometimes difficult, questions that lie behind patterns in the data that they see. Public agencies will need to use data to examine the impact of previous interventions on race equality and to explore, over time, the impact of different interventions. Improving the quality of ethnicity data and providing open access to the data is likely to aid analysis of inequality. But more than this, we need a stronger commitment and investment to examine what works in progressing race equality over longer periods of time (e.g. 5-10 year policy cycles).
Link to research (open access for 14 days from date of publishing this article): https://onlinelibrary.wiley.com/doi/10.1111/spol.12501