Development of the socio-economic background harmonised standard: research findings

This report summarises the findings from our work to develop the current socio-economic background standard.

Policy details

The Government Statistical Service (GSS) Harmonisation team is based in the Office for National Statistics (ONS). The team supports harmonisation of data across the UK. This involves providing bespoke advice and harmonised standards and guidance about how data on different topic areas are collected and presented. The standards and guidance are designed to improve the consistency, coherence, and comparability of statistics.

Producers of statistics use these harmonised standards as a starting point in their data collection process. By using harmonised standards, they can:

• align with other producers of statistics
• increase the usefulness of their statistics
• meet the Code of Practice for Statistics’ cross-cutting theme of coherence

Harmonised standards also allow users to effectively and accurately compare data that has been collected across different datasets. This means we can more easily understand what that data does, and does not, tell us. This ensures that statistics are being used to their full effect for the public good.

About this report

This report presents the research that has been done during the development of the first harmonised standard for socio-economic background.

This report:

• provides contextual information about what socio-economic background is and how it is currently measured
• sets out the reasons why this standard was developed, including the approach we took
• presents the findings from our research, including qualitative and quantitative findings, and shows how the socio-economic standard now looks
• sets out what will happen after the publication of the socio-economic background harmonised standard

As part of this development, we have worked with a variety of government stakeholders to gather feedback on their experiences of using socio-economic background questions. This included asking about areas where stakeholders felt the questions could be improved, and what their data needs were for this topic.

While we will give some information about stakeholder engagement in this report, we will mainly be discussing the research we did with the public to understand their experience of answering these questions.

What we mean by socio-economic background

The Social Mobility Commission defines “socio-economic background” as:

“the term to refer to the particular set of social and economic circumstances that an individual has come from. It permits objective discussion of the influence of these circumstances on individuals’ educational and career trajectories; and it can be objectively measured by capturing information on parental occupation and level of education.”

Socio-economic background is different to the term “social mobility”. The Social Mobility Commission defines “social mobility” as a way of considering the link between a person’s income and occupation, compared to their parents or peers.

By capturing information about people’s socio-economic background and current occupation data users can use the data to:

• investigate trends in social mobility
• provide evidence for policy interventions to help people from a lower socio-economic background

Reasons for this development

Socio-economic background information has been collected by some employers for several years, as reported in the Civil Service recommendations for measuring socio-economic background back in 2018. These recommendations were adapted by the Social Mobility Commission in 2020 in their cross-industry toolkit. This shows the need to measure socio-economic background to understand barriers to social mobility within the workplace.

But while data about socio-economic background may be collected in some workplaces, the Inclusive Data Taskforce Recommendations highlighted it is rarely available within government data and censuses. Socio-economic background was identified as an important concept for understanding inequalities across the population, which makes it highly relevant for policymaking. It has been recommended that a standard should be developed to address the gap in government data, using the work that has been done to measure socio-economic background as the foundation.

Stakeholder engagement

The Harmonisation Team worked with stakeholders who had previous interest in the topic of a socio-economic background harmonised standard. The goal of this work was to understand their data needs for a new standard. Once the 2022 Harmonisation Workplan was live, we increased our engagement and worked closely with government departments, Devolved Administrations, and our Harmonisation Champions to understand their work further. We then created a topic group to ensure that the data user needs were considered throughout the development work for the new standard. This also meant we were able to readily share research findings with stakeholders.

We have engaged with stakeholders on an ongoing basis throughout this project. This has meant that the team could be kept up to date with any emerging information and contexts that may be important to this work. For example, we were able to use information from stakeholders to help us work on futureproofing the questions when considering measures such as free school meals.

Research methodology

As part of the development process for the socio-economic background harmonised standard, we have undertaken research, including:

• desk research into questions already being asked on the topic as well as analysing queries and feedback received to harmonisation
• qualitative research in the form of two rounds of cognitive interviews with the public to gain in-depth data about their experiences of answering our question set designed to measure socio-economic background
• quantitative research to assess responses and accuracy rates for the free school meals question and parental qualification question

This approach ensures a user-centred design approach to question design. This is done so that we meet the data needs of our stakeholders and enable them to get high quality data. We do this by making sure the public can answer the questions as they are intended. Insight from our research will also help us reduce respondent burden.

For example, we are aware of the data quality risks when asking respondents to recall parent or guardian’s occupations when they were 14. This meant that our question development concentrated on both:

• capturing the necessary information to establish a participant’s socio-economic background
• designing the questions in a way to that would help with participant recall

For our sampling in both the qualitative and quantitative methods, we used a recruitment company called “People for Research”. This helped us ensure we got representation from across the UK with a spread of ages, genders, ethnicities, and socio-economic statuses. We had:

• 28 sessions for the first round of cognitive interviews
• 843 responses in the small-scale survey
• 30 sessions for the second round of cognitive interviews

We used inductive coding to analyse the cognitive interviews. This meant we could establish a coding frame, and then use thematic analysis to identify the themes throughout the sessions. For the small-scale survey, we compared response rates for different questions and investigated the percentages.

Research and engagement findings

The report has broken down the research and engagement findings by each question that will be included within the harmonised standard.

The original and current versions of the questions are given in this report so you can compare them. We have also shown the changes to the questions in bold.

School type question

Completion

Most participants found the school type question easy to answer and understood what it was asking, especially when:

• their school type was mentioned in the guidance
• they attended school outside of the UK

We found that participants tended to read all answer options and use a process of elimination if they were unsure about how they should answer. For example, we noticed this with participants who had attended school in Scotland or Northern Ireland.

Some participants noted the age range in the question stem. To reduce burden, we added the word “secondary”.

Understanding

When asked what “state-run or state-funded school that was non-selective” meant to them, participants understood it to be a school:

• which was not for a particular group or faith
• where there were no fees to attend the school
• that was run by the local authority or the government
• that was the ‘standard school’ they would think of

Most participants understood what “academic ability” meant in the guidance, but there was less clarity around what was meant by “a faith school”. This led to confusion for participants whose schools were affiliated with a faith, but may not have selected pupils based on faith. This was found to be particularly relevant for participants from Northern Ireland, but we did not find this led to incorrect reporting.

We added guidance to the “independent schools” response option to give “private schools” as an example of what we mean. We discovered that respondents used the terminology “private schools”, so we recycled this wording according to user-centred design principles.

 

Free school meals question

Completion

Participants found this question easy to answer, regardless of whether they did or did not receive free school meals. This question did not apply to participants who finished school before 1980, but these participants still found the question easy to answer.

Understanding

Participants had a shared understanding of free school meals, where they felt it meant that children were provided lunches by the school as their parents were on low income or were on specific benefits from the government.

Participants also understood the term “eligibility” to mean meeting a set of criteria. The first round of cognitive interviews showed that the term “eligible” confused participants who may have been eligible for free school meals but did not receive them. We ran a quantitative survey to test the effect of changing the wording from “eligible” to “received” as a potential solution for this. The survey showed there was a significant portion of participants that could identify that they were eligible for free school meals but did not receive them. Based on these findings, we kept the terminology as “eligible”.

We made some other small wording amendments to the question, in line with accessibility and best practice. The phrase “at any point during your school years” was changed to “at any time between the ages of 11 and 16” to reflect recent policy changes meaning that children in reception and primary schools now receive free school meals in Wales. Our second round of cognitive interviews showed that this change worked well and did not result in any burden for participants.

 

Employee, self-employed, or not working question

Completion

Overall, this question was successful in capturing the employment status of the main or highest income earner in the household when the participant was 14 years old. Participants reported few issues when answering this question and understood it well. Many participants said the question was easy to answer and were confident with their responses.

Understanding

It was not immediately obvious to some participants who the “main or highest income earner” in their household was, as their parents earned a similar salary. But this did not affect the findings and participants were still able to respond accurately.

This question worked well in both rounds of testing, so few wording changes were made in line with accessibility and best practice.

 

Number of employees question

We tested 2 versions of this question as part of the development of a new harmonised standard for socio-economic background. One version was routed for participants who selected that their main or highest income earner was an “Employee”. The other version was for participants who selected “Self-employed with employees”. Combining the questions would have created a longer, more complicated question stem. So, we decided to create two different questions to reduce respondent burden. The fundamentals of the question remained the same for both.

Completion

Initial findings from this question were mixed. Some participants found it easy to recall the size of the company or number of employees because they knew the company was large or small. Other participants found this question harder as it was not immediately obvious what size the company would have been. This burden mainly affected participants who were answering the “employed” version of the question.

Understanding

Some participants had difficulty understanding what the question was asking. They felt that the word “people” or “employees” was missing from the question stem, which led to confusion around what the question was asking about. We added the word “people” to the question stem to reduce respondent burden.

Some participants were unsure whether the question was asking about the number of employees at the location where the main or highest income earner worked, or at the whole company. This would have drastically changed the answer for some participants. We added guidance after the first round of testing to help clarify the meaning of the question in the “employed” version.

Some participants found the guidance useful and reported that it helped them understand the question better. But there were also several participants who were still unsure of how to correctly answer the question, despite the fact the guidance had helped their understanding somewhat. This suggests that the use of guidance helps reduce confusion, but any further burden would be because of the data requirement itself.

 

Supervisor question

Completion

Overall, participants had a good understanding of the supervisor question and were confident in answering it. Feedback from participants showed that the question was easy to answer for a wide range of circumstances. This means that the question was widely applicable and did not just cater towards a certain group or occupation.

Understanding

Some participants used the guidance that defined what the term “supervise” means, but others did not. When we asked participants about their thoughts on the guidance, they commented that they felt the guidance was helpful, clear and easy to understand, regardless of whether they had actually used the guidance to help them answer the question. Some respondents said the guidance had helped them confirm what they thought supervisor meant.

Some participants had questions about the phrase “day-to-day” in the guidance. This was because they felt that you do not have to supervise people day-to-day to be their supervisor. For some, this changed how they would answer the question. We amended the wording from “day-to-day basis” to “regular basis” because of this. This change worked well in testing.

While most participants understood the meaning of the term “supervisor”, there was some confusion over the extent of supervision that the question was looking for. Some participants were unsure whether the supervision had to be in a formal capacity, or if mentoring or voluntary supervision would apply. We added the word “formally” to the question stem to make this clearer.

 

Parental occupation question

This question will be asked of respondents who have indicated that their main or highest income earner was any of the following:

• an employee
• self-employed freelancer without employees
• self-employed with 1 to 24 employees

Completion

Experiences with the parental occupation question were mixed. Some participants found this question easy to answer because they were able to see the job title in the example, or they were very confident in about how the occupation would be categorised. Some participants found it more difficult to answer the question. These participants had more difficulty because the specific role they were searching for did not appear in the guidance, or because the occupation did not fit neatly into the response options.

Understanding

Participants often used the examples in the guidance to understand what the different answer options meant and establish the most likely option. Despite any burden, most participants were able to answer accurately to the best of their knowledge.

The original parental occupation question that we tested contained 8 occupational classes. The first round of cognitive testing showed there were too many answer options, which made the question burdensome. To lessen this burden, we combined categories to reduce the answer options to a five-class model. We tested this new, reduced model through the quantitative survey and found it worked well, with most participants selecting the correct answer option. Our recommendation is to keep the reduced set of answer options. Combining the response options has made no difference to the outputs that can be gathered by data processors and users.

The findings from this survey identified where participants often choose the wrong occupational category. We have made the following changes to reduce the risk of incorrect reporting in future:

• engineers were often misplaced as the current guidance only referenced “civil or mechanical” — reporting was improved by replacing this with “engineering professions”
• managerial roles were often put into options to reflect the industry, rather than the “managerial” option — we changed the guidance to reflect all levels of managerial roles which helped with reporting

Some burden was experienced by participants who reported that the main or highest income earner in their household was a self-employed business owner. We reviewed the data need and respondent pathway through the question set and determined that this question could be removed for those who were self-employed with 25 or more employees without affecting data quality. The question will be kept for those who were self-employed with 1 to 24 employees.

Generally, the occupational categories were well understood. In cases where they were not, participants were able to understand the meaning of the category through the examples in the guidance. As this did not affect response accuracy, we recommend keeping the names of the answer options in their current form.

 

Parental qualifications question

Completion

The parental qualifications question worked well in both rounds of cognitive testing. Most participants found the question easy to understand and were able to recall with relative ease whether their parents or guardians held a degree or degree-level qualification, or not. Participants who did not know their parent’s qualifications, correctly selected “don’t know” as the option.

Understanding

There were mixed findings on the experience of participants when their parents held qualifications that were older or were achieved outside of the UK. Some participants were confident in their answer when converting the qualification. This was the case where they knew how it would equate to the current UK education system. Participants were less confident where their parents had more uncommon qualifications or if they were unsure on overseas equivalents. In these cases, the guidance helped them to answer the question.

The question was often understood as asking whether their parents had attended university or not. This led us to think about whether it would be necessary to redesign this question to ask whether either of your parents or guardians attended university. But this change would have failed to capture respondents whose parents achieved other degree-level qualifications. On this basis, we would not recommend changing this question. Only minimal wording changes were made, in line with accessibility and best practice.

 

Conclusion

Summary

Our research and engagement activities mean we are confident that these questions can capture socio-economic background of respondents in the UK. This has included qualitative and quantitative research, as well as regular engagement activities with a wide range of stakeholders.

We have published the new socio-economic background harmonised standards alongside this research report. The new harmonised standard gives details of the recommended questions, as well as output guidance and interviewer-led mode questions.

What happens next

We will:

• review the harmonised standard in five years’ time to assess how it has been implemented across the Government Statistical Service
• continue to support our users with any queries that they have and support them in using the standard

Contact us

If you would like more information about the socio-economic background harmonised standard, please contact the Harmonisation team at Harmonisation@statistics.gov.uk.

Find out more about harmonisation.