Absolutely—I could not agree more. There are valid reasons why it is difficult to disaggregate equalities data. Although the populations of groups might be quite big, they might comprise quite a small proportion of the overall population, hence it is difficult for survey data to capture enough people to know that results are robust. However, there are ways around that. Government statisticians have a system of asking similar questions across a range of surveys to try to increase the sample size for particular equalities groups, which they have used for minority ethnic groups and, I think, religious groups. That is still not ideal, because it does not necessarily allow people to ask very detailed questions—it is just about getting at the broad base of some of the big issues.
Disaggregation is very important and I know that there is a lot of frustration in some stakeholder communities when they are given reasons such as a lack of a big enough sample size; they say that that is not good enough. We know there is a problem with sample size and that we need to sort it out.
To be able to provide good cost-effective future policy to address the gender pay gap, it is important to be able to pick out exactly what is happening through people’s life courses, to see where the issues come up, and, where there have been successes, to understand what works and, in particular, what does not work.
In relation to statistics for minority ethnic groups who might face very different barriers, I note that we often have just a white group and a non-white group, and it is not helpful to badge people in just that way.
I have already talked a lot about disability, which is a word that you see in a lot of mainstream policy documents and strategies, but it is clear that the people writing them do not understand exactly what disability means. It is not just one thing; you cannot just have one disability action plan. Of course there are parts of Government that understand and really get this—the minister and the equalities unit, who are on the next panel of witnesses, for instance—but there is not necessarily a mainstream understanding of disability in the likes of the economic development teams. If data are not disaggregated, the more mainstream analyses do not necessarily capture disability. To misquote someone, “If you don’t measure it, it won’t get done”.
Good data is critical if you want people to understand and be able to create mainstream policy, because it is the language of a lot of Government departments. There are good initiatives that are trying to capture the voices of lived experience, which is important in helping to understand how to interpret the data and where the data misses bits and pieces, which it always will. The participation approach is necessary, but it must be both, not just one or the other.
Investment is necessary to work out where the data can be found. I know that there is work going on to look at administrative data sources and see where disaggregated data on equalities can be used from other sources where it is collected for other purposes. Departments running the big Scottish Government surveys, such as the Scottish household survey and the Scottish health survey, could think a lot more about how to disaggregate their data.
The Scottish Government puts a lot of money into the family resources survey, which is the carried out by the Department for Work and Pensions and is the main data source on incomes and poverty that is used across the UK. It forms the basis of how we measure the child poverty targets in Scotland. It uses a big sample from across the UK and quite a big sample from Scotland proportionally, and there could be some improvements in how the questions are asked and the results are disaggregated. There is potential, but investment is needed. Surveys are expensive, but they are critical to getting understanding and information into mainstream debates.
I could not agree more that it is important to focus on disaggregating data.