Time Well Spent: A national survey on the volunteer experience

Sample

The sample was drawn from the YouGov panel, which is designed to yield a representative sample of adults aged 18 or over in Great Britain. The responding sample is weighted to the profile of the sample definition (see below) to provide a representative reporting sample. The total sample size was 10,103 adults.

Weighting

The survey data were weighted to the marginal region, social grade and age/gender/ educational-level distributions, as set out below in Table A1.1. All the percentages presented in this report are based on weighted data. Details of weighted and unweighted bases for standard demographics are shown in Tables A1.3 and A1.4, at the end of Appendix 1.

Questionnaire development and testing

Questionnaire development was informed by a scoping phase to help shape the research and its focus. This was comprised of two parts.

1.We conducted a rapid review of existing literature, and previous and current national surveys on volunteering, to look at the existing evidence base on the subject area and identify knowledge gaps. The review included the Community Life Survey72 (the current survey on volunteering trends), Helping Out73 (the previous national survey which explored the volunteer experience) and Pathways through Participation74 (looking at how people’s involvement changes over their lifetime).

2.We undertook 18 telephone interviews with stakeholders across the voluntary sector to understand their current priorities and interests in relation to the volunteer experience to help define the focus of the survey. In addition to telephone interviews, we also engaged with stakeholders at events where volunteer managers were present.

From this scoping phase, we identified a number of priority areas, which formed the basis of the questionnaire development. We drew on existing survey questions where relevant – especially where these questions had previously undergone cognitive testing. Expert reviews of the draft questionnaire were also carried out; this involved a broad range of stakeholders including researchers, volunteer managers and other voluntary sector experts. These reviews were used to ensure relevance of the questions and their responses and helped us to prioritise questions, given the limited number of questions which could be included in the survey.

Throughout the questionnaire development phase, different versions were tested with some members of the general population to check clarity and interpretation, focusing particularly on new questions.

Further stakeholder engagement

In order to ensure that the research reflected the needs and interests of those engaging with volunteers and that it generated insights that would be practical and useful, we engaged with a variety of stakeholders (eg volunteer-involving organisations and networks) throughout the research process, not just during the questionnaire design phase.

During the initial analysis phase, we conducted a workshop in July 2018 with a small number of stakeholders to feed back on and discuss early findings. We then carried out a more formal set of three workshops in September 2018, two in London and one in Leicester, and engaged with over 70 stakeholders, to present some of the emerging findings from our research and provide an opportunity for people to discuss and debate the implications of these for practice and policy. These were used to inform the ‘Conclusions and implications’ section of the report. Stakeholders represented a broad range of organisations, including smaller organisations, and were from a variety of sectors.

Data collection and response

The survey was conducted using an online interview administered to members of the YouGov UK panel of 800,000+ individuals who have agreed to take part in surveys. Fieldwork was undertaken between 4 and 15 May 2018.

Emails were sent to panellists selected at random from the base panel sample. The email invited them to take part in a survey and provided a generic survey link. Once a panel member clicked on the link, they were sent to the survey that they were most required for, according to the sample definition and quotas (the sample definition could be ‘GB adult population’ or a subset such as ‘GB adult females’). Invitations to surveys do not expire and respondents can be sent to any available survey.

Because of the allocation to different surveys according to sample quotas, it is difficult to calculate a ‘traditional’ response rate. We do have information on dropout; 11,247 started the survey, whilst there were 10,103 final respondents, a response rate of 90%.

Sampling errors and statistical significance

No sample precisely reflects the characteristics of the population it represents, because of both sampling and non-sampling errors. In a random sample, where every adult has an equal and independent chance of inclusion, it is straightforward to calculate the sampling error of any percentage and a confidence interval for the true population percentage, which helps determine whether differences between two percentages are statistically significant.75 However, simple random sampling is almost never used in practice, because of time and cost; most sample designs are more complex.

As noted above, our sample is a mix of random and quota sample from the YouGov panel. With any complex design such as this, the sampling errors are larger than for a random sample of the same size and depend not just on the percentage and sample size but also on how that percentage response is spread across the different types of people in the sample. To estimate that greater sampling error, various measures are used.76 YouGov estimate the efficiency of their weighting design, with the weighting to the target distributions shown in Table A1.1, to be 88%.

In general in the report, we discuss findings where the differences between groups are statistically significant at the 95% level. On the occasions where we draw attention to a finding that is not statistically significant, or that is based on a small sample size, we normally comment on that.

Table and figure conventions

The following conventions are used for tables and figures throughout the report. 1.When findings based on the responses of fewer than 100 respondents are reported in the text, reference is made to the small base size. Such findings are not generally included in charts. 2.Percentages equal to or greater than 0.5 have been rounded up (e.g. 0.5% = 1%; 36.5% = 37%). 3.Due to the effects of rounding and weighting, percentages will not always add up to 100%.

Variables (including definitions)

Socio-economic and demographic analysis variables

A number of standard variables have been used for the analyses in the main part of this report. The key ones are described below.

Social grade

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. For more detail of individual groups see: http://www.nrs.co.uk/ nrs-print/lifestyle-and classification-data/social-grade/ (accessed January 2019).

There are six classification categories:

• A Professional etc. 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 them 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 and to that extent are more self-explanatory. The principal ones are:

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

For disability, the following definitions are used.

• 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.

Volunteering analysis variable

One of the key variables underpinning the report is the extent to which people have volunteered through a group, club or organisation, over their lifetime and recently.

Where the term ‘volunteering’ is used, this refers to formal volunteering through groups, clubs or organisations, which is the focus of this report.

It does not include the more informal ways of giving time and helping others outside groups, clubs or organisations.

Whilst ‘volunteering’ is used throughout the report, in the survey respondents were not asked if they had volunteered. Instead, they were asked whether they had been involved with any groups, clubs or organisations and then whether they had provided unpaid help to any groups, clubs or organisations, prompted by a list of activities (see questionnaire). This method, following that used in the Community Life Survey, was used to capture the full range of volunteering activities, some of which may not otherwise be recognised by respondents as volunteering.

For the analyses in the report, we group people into the following categories:

• recent volunteers, who have volunteered at least once in the last 12 months
• lapsed volunteers, who volunteered between one and three years ago
• those who have volunteered in the past but more than three years ago
• those who have never volunteered through a group, club or organisation.

We also refer to the frequency of volunteering, generally by the following:

• frequent volunteers, who volunteered at least once a month
• occasional volunteers, who volunteered less frequently than once a month.

Other analysis variables

A number of other variables are used throughout the report for analysis. These focus primarily on: how people volunteer and who they volunteer for.

The majority of these are self-explanatory, but it is worth taking note of the following definitions.

• Employer-supported volunteering: volunteering which is done either during working hours (with the time given by employers) or organised by employers; not including schemes for giving money.
• Civil society/third sector: a charity, voluntary organisation, community group, faith-based organisation, social enterprise, non-profit organisation (eg local sports club, environmental group, befriending scheme).
• Public sector: a public service, body or institution (eg NHS, local council, school, library, police).
• Private sector: a private company, corporate, business, profit-making organisation (eg private nursery, private museum, private health organisation, private care home, theatres).

Regression analysis techniques77

Regression analysis aims to summarise the relationship between a ‘dependent’ variable and one or more ‘independent’ variables. It shows how well we can estimate a respondent’s score on the dependent variable from knowledge of their scores on the independent variables. It is sometimes presented as supporting a claim that the independent variables cause the phenomenon measured by the dependent variable, but this is not correct; causality can only be inferred through special experimental designs or through assumptions made by the analyst.

All regression analysis assumes that the relationship between the dependent and each of the independent variables takes a particular form. In linear regression, it is assumed that the relationship can be adequately summarised by a straight line.78 Logistic regression is an alternative form of regression, more suitable for variables such as ours, which fits an S-curve rather than a straight line; the impact on the dependent variable of a one-percentage-point increase in an independent variable becomes progressively less the closer the value of the dependent variable approaches 0 or 1.

The statistical scores most commonly reported from the results of regression analyses are as follows.

A measure of variance explained

This summarises how well all the independent variables combined can account for the variation in respondents’ scores in the dependent variable. The higher the measure, the more accurately we are able in general to estimate the correct value of each respondent’s score on the dependent variable from knowledge of their scores on the independent variables.

A parameter estimate or coefficient

This shows how much the dependent variable will change on average, given a one-percentage-point change in the independent variable (while holding all other independent variables in the model constant). The parameter estimate has a positive sign if an increase in the value of the independent variable results in an increase in the value of the dependent variable, and a negative sign if an increase in the value of the independent variable results in a decrease in the value of the dependent variable.

It is possible to compare the relative impact of different independent variables; those variables with the largest estimates can be said to have the biggest impact on the value of the dependent variable.

Regression also tests for the statistical significance of parameter estimates. A parameter estimate is said to be significant at the 5% level if the range of the values encompassed by its 95% confidence interval is either all positive or all negative. This means that there is less than a 5% chance that the association we have found between the dependent variable and the independent variable is simply the result of sampling error and does not reflect a relationship that actually exists in the general population.

Details of the regression analysis carried out on overall volunteer satisfaction and the likelihood of continuing to volunteer

A large set of variables were included in the regression models, organised into blocks. Table A2.2 lists the blocks and the variables included within them: demographic factors; type of volunteering; experience of recruitment, induction and training; and respondents’ opinions about the positive and negative impacts and experiences that volunteers had experienced.

We used a block-wise forward selection method of entry.79 With this method the dependent variables are grouped into blocks, based on psychometric consideration or theoretical reasons,80 and a stepwise selection is applied. Each block is applied separately while the other predictor variables are ignored. Variables can be removed when they do not contribute to the prediction. With this method we were able to identify which variables within a block were contributing to the equation and which could be ignored, before adding further block(s).

Since the order of entry can have an impact on which variables will be selected (with those entered in the earlier stages having a better chance of being retained), we began with the demographic variables and then other more ‘factual/ objective’ variables, such as type of volunteering, before moving on to more ’subjective’ variables, such as experience of volunteering. The analyses were carried out on all recent volunteers, that is those who have volunteered in the last 12 months.81

Findings from the logistic regression

The key findings have been described in sections 5, 6 and 7 of the report and confirm indications earlier in the report that it is some of the statements on how volunteers feel about their volunteering experience which are most strongly associated with overall satisfaction and propensity to continue.

Table A2.1 lists all the factors (and their coefficients) found to be significantly and independently associated with satisfaction with volunteering or with the likelihood of continuing to volunteer. In Table A2.1 (a) we present the logistic regression where the dependent variable is ‘being very or fairly satisfied with volunteering’ (as opposed to being very or fairly dissatisfied).

In Table A2.1 (b) the dependent variable is ‘being very or fairly likely to continue giving unpaid help over the next 12 months’82 (as opposed to being very or fairly unlikely to continue). A positive coefficient indicates a higher score while a negative coefficient indicates a lower score. For categorical variables, the reference category is shown in brackets after the category heading.

Looking at the satisfaction model (Table A2.1 (a)), most of the positive and negative associations between overall satisfaction and aspects of how volunteers feel about their volunteering experience have been reported in section 5.5.

Two other rather different factors also featured: volunteering outside the UK (with increasing likelihood of volunteering outside the UK associated with less likelihood of being satisfied) and having received role-specific training (with increasing likelihood of having received such training associated with greater likelihood of being satisfied).

In addition, there was also a small but significant negative association with one of the four wellbeing questions ‘Overall, how happy did you feel yesterday?’

Looking at the model for the likelihood of continuing to volunteer (Table A2.1 (b)), most of the positive and negative associations between overall satisfaction and aspects of how volunteers feel about their volunteering experience have also been reported in section 5.5.

Again, two other rather different factors also featured: sometimes volunteering alone (with increasing likelihood of volunteering alone being associated with a greater likelihood of continuing to volunteer) and a reference check being carried out as part of the recruitment process (with checks having been carried out being associated with a greater likelihood of continuing to volunteer).

Looking at the two sets of factors alongside each other, three statements are significant factors for both overall satisfaction and likelihood of continuing: culture of respect and trust; made me feel I was making a difference; enjoyment (I enjoy it).

It is perhaps not surprising that some of the same factors are associated with both satisfaction and continuing to volunteer. In an attempt to more clearly separate influences on satisfaction and on continuing to volunteer, we repeated the ‘continuing to volunteer’ regression model with overall satisfaction included as an additional variable. The results were, however, very similar. The model shown below, therefore, does not include overall satisfaction – see Table A2.1 (b).