Sample, quota, and fieldwork dates

We commissioned YouGov to recruit a nationally representative sample of 1,000 adults from the global majority population aged 18 years or older in Great Britain via an online panel.

The fieldwork ran alongside a nationally representative sample of 7,006 adults from the overall population[1], which contained global majority respondents. Respondents were invited to take the survey through email if they met the survey criteria.

The fieldwork took place from 23 November to 6 December 2022. The survey was carried out online and lasted approximately 15 minutes. 1,165 started the survey, and 1,016 completed it, resulting in a completion rate of 87%.

All figures, unless otherwise stated, are from YouGov Plc. NCVO analysed the data independently.

Merging datasets and weighting

Merging data

After the fieldwork, we combined the data from the two samples to achieve the largest possible number of global majority respondents for analysis.

The first stage was to merge the overall population dataset (n=7,006) with the global majority dataset (n=1,000). This provided a total sample, within this merged dataset, of 1,454 global majority respondents (454 from the main sample and 1,000 from the boost).


This research adopted a standard approach to combining and weighting a boost subgroup sample with a nationally representative survey dataset.

Certain subgroups in the population are less likely than others to take part in surveys. These groups have a risk of being underrepresented in the data. Weights are applied to correct these biases.

Weighting is applied at the combined data level. This is because we have a more accurate profile of all UK adults on which to base the weighting targets, compared with profiles of the global majority population alone. This approach involves less extreme weighting than weighting within the global majority sample alone. It therefore produces a higher effective sample size for analysis of this audience.

After merging the two datasets, we weighted the total combined dataset (n=8,006) using the same weighting targets as used for the overall population sample. We included an additional ethnicity profile in the weighting targets.

Table 1 below shows sampling quotas for the demographics. The survey data was then weighted to the marginal region, socio-economic group, age, gender, and educational-level distributions. All the percentages presented in this report are based on weighted data.

Find out more about how YouGov sets quotas and weighting.

Filtering responses to the global majority

To analyse and report on global majority figures, we filtered the combined weighted dataset to global majority respondents only (n=1,454). This is the basis for analysis in this report.

Within this, questions about formal volunteering were filtered to formal volunteers only (n=392).


We recognise that there are different ways we can look at the data and make comparisons to draw insights into global majority volunteering.

This report uses the overall population as a benchmark, rather than comparing the global majority sample with white respondents only. We decided on the approach to make comparisons between

the overall population and the global majority as a way of understanding how the experiences of the global majority may differ from the overall ‘average’ experience, and why.

This overall population sample includes some of the global majority survey respondents, as it is a nationally representative sample. This sample was used as the basis for the main Time Well Spent 2023 report.

We decided to use the same overall population dataset for comparison with the global majority findings so that readers of both reports would see a consistency in the numbers and percentages reported.

This approach enables us to draw insights through two ‘lenses’. First, the global majority in and of itself. Second, a comparison between the global majority population and the overall population. Those who work with volunteers can consider how to adapt their overall approach, while also addressing the specific needs of global majority volunteers.

While the findings reported are through these two groups, as part of our analysis, we conducted some spot checks comparing global majority with white respondents to understand whether this resulted in different findings. While there were some differences, these were minor and did not result in any changes to the overall findings.

Demographic questions

The research followed the standard definition and methodology used in the census and other research to classify demographic groups. The key classifications are described below.


In asking respondents about their ethnic group, YouGov’s online panel used the grouping used by the Office of National Statistics in the following way.

Respondents were allowed to select only one option from the above.

Socioeconomic group

Socioeconomic group (also known as ‘social grade’) is a classification based on the occupation of the chief income earner of the household, with six categories. Information is collected about their current or last job, so that all respondents (except those who had never worked) are coded.

The National Readership Survey has more information about social grade definitions.

There are six classification categories:

  • A: Professional occupations
  • B: Managerial and technical occupations
  • C1: Non-manual skilled occupations
  • C2: Manual skilled occupations
  • D: Partly skilled manual occupations
  • E: Unskilled occupations

In this report, we group respondents into two broad categories, ABC1 (non-manual occupations) and C2DE (manual occupations and people not working).

Other socio-demographic analysis variables

These are generally taken directly from information collected by questionnaire when people join the YouGov panel. Those that do change, working status etc., are asked on a frequent basis to keep updated. The principal variables are:

  • gender
  • age
  • ethnicity
  • highest educational qualification obtained
  • working status
  • disability.

For disability, we use the following definitions.

  • Disabled: Reported day-to-day activities being limited in some way because of a health problem or disability which has lasted, or is expected to last, at least 12 months.
  • Non-disabled: Reported no limitations to day-to-day activities because of a health problem or disability which has lasted, or is expected to last, at least 12 months.

Questionnaire development

The same questionnaire was used for our research of the general public and the global majority. Before developing the entire questionnaire, we carried out a scoping phase, to help shape the research and its focus.

Part 1

We conducted a rapid review of existing literature, and previous and current national surveys on volunteering, to review the existing evidence base on volunteering and identify knowledge gaps.

The review focused particularly on volunteering trends during and after the covid-19 pandemic.

Part 2

We carried out five interviews with external stakeholders across the voluntary sector. This included researchers and sector leaders familiar with the topic of the global majority and volunteering.

Our aim was to understand their current priorities, interests, and challenges in engaging with this group.

We also engaged with volunteer managers at events held by NCVO and consulted internally with REACH Network, NCVO’s staff network of global majority colleagues.

Part 3

From this scoping phase, we identified a number of priority areas, which formed the basis of the questionnaire. We drew on existing survey questions where relevant – especially where these questions had previously undergone cognitive testing.

We also asked experts to review the draft questionnaire; this involved a broad range of stakeholders including researchers, volunteer managers and other voluntary sector experts. These reviews helped us ensure our questions were clear and relevant. They also helped us prioritise questions, given the limited number of questions we could include in the survey. As the bulk of the 2019 survey was replicated, cognitive testing was not carried out.

As part of this process, we decided which questions from the 2019 survey to retain and to amend for comparability between the two surveys. The key questions that we amended are covered in the volunteer values questions section below.

Methodology of specific analysis

Significance testing

As outlined in the introduction, the primary aim of this research was to understand volunteering behaviour and experience of the global majority population. When comparison is made between two datasets or two subgroups within a single dataset, significance tests were conducted (i.e. two-sample test of proportions) to measure if the differences observed were statistically significant.

Significance testing is a statistical process which shows whether differences between sub-samples, or over time, are likely to be real (as opposed to being an artefact of the confidence intervals which are usually inherent in any sample survey). Significance testing was conducted at the 95% confidence level, which means that we can be 95% confident that the differences observed are genuine differences and have not just occurred by chance.

Significance testing for differences between subgroups was conducted using a two-tailed t-test, based on the proportion of respondents giving the response in question in each group. The unweighted base sizes and rounded percentages were used as the input values for significance testing.

There was some overlap between the overall sample and the global majority sample. This means the two samples are not fully independent. It is possible to include the degree of overlap when significance testing to take account of this. However, we treated the two samples as independent when testing significance, so as not to over-estimate the likely significance of any apparent differences.

In total the overlap made up 7% of the combined sample sizes of both surveys (454 out of 8,006) so the overlap is fairly small, and any adjustment would make relatively little difference. In addition, given the nature of the sample (quota, rather than random), significance testing is solely indicative of differences.

Given that adjusting for overlapping samples will return a smaller difference as being significant than tests for independent samples, it was decided to use significance testing as if for two independent samples, to err on the side of caution.

A full range of significance tests on volunteer data were re-run, this time including a factor to account for the small overlap. This did not alter which differences were returned as significant.

Key driver analysis

Key driver analysis (KDA), also known as relative importance analysis, essentially looks at a group of factors and weights their relative importance in predicting an outcome variable using regression. KDA shows what factors or variables can drive another independent variable.

Specifically in this research, we conducted KDA to see the importance between a range of factors – volunteer demographics and experience – and volunteer satisfaction (see key factors influencing volunteer satisfaction).

The analysis works by running multiple correlations between these factors. These variables included the age, gender, and socio-economic group of the volunteer as well as positive and negative aspects of their volunteering. KDA produces a percentage for how much each of the attributes explains any difference in volunteer experience.

The table below shows the key factors influencing volunteer satisfaction for global majority volunteers.

Volunteer values questions

We used the following process and methodology to develop a question on volunteer values. We added this new question to help us understand what volunteers and non-volunteers value if or when they give unpaid help.

The key findings from Time Well Spent 2019 showed eight key factors that comprised a quality volunteer experience.

  • Balanced: It doesn’t overburden those who volunteer with unnecessary processes.
  • Connected: It gives people a sense of connection to others, a cause, and/or an organisation.
  • Enjoyable: It provides enjoyment and people feel good about what they are doing.
  • Flexible: It takes into account how volunteers can give their time and fits around their circumstances.
  • Impactful: It makes a positive difference.
  • Inclusive: It is welcoming and accessible to all.
  • Meaningful: It resonates with people’s lives, interests and priorities.
  • Voluntary: The volunteer has freely chosen to do it.

Based on this, we developed 10 statements exploring these values.

  • ‘It is important to be given the opportunity to influence the development of the group, club, or organisation.’
  • ‘It is important to feel recognised enough for the help they give.’
  • ‘It is important to feel like they belong to the group, club, or organisation.’
  • ‘It is important to feel the group, club, or organisation is making a difference.’
  • ‘They value a culture of respect and trust in the group, club, or organisation.’
  • ‘They want people from a wide range of backgrounds and cultures.’
  • ‘It is important for them that they are not pressured to give more time.’
  • ‘It is important for them to meet people through the group, club, or organisation.’
  • ‘It is important for them that they enjoy giving unpaid help.’
  • ‘It is important for them that they feel less isolated as a result of giving unpaid help.’

To understand the values of non-volunteers, we asked all respondents to imagine a person who is giving unpaid help and to rate each statement on a scale of 1 to 6, with 1 being ‘they are not at all like me’ to 6 being ‘they are very much like me’. Finally, a mean score for each statement was calculated to compare the values.


  1. Analysis and key findings of the overall population sample were published in the main Time Well Spent 2023 report in June 2023.

This page was last reviewed for accuracy on 28 November 2023