The Scottish Government

« Previous | Contents | Next »

Listen

ANNEX 2 METHODOLOGY

This annex begins by setting the context of the Scottish Household Survey and later describes the study methods employed to meet the objectives outlined in Chapter 1.

Scottish Household Survey

The Scottish Household Survey ( SHS) was introduced in 1999 and is a continuous cross-sectional survey which gathers information from a large sample of Scottish households. Its aim is 'to provide representative information about the composition, characteristics and behaviour of Scottish households, both nationally and at a more local level. (Scottish Executive, 2005b)

The SHS gathers information from the highest income householder or their spouse/partner and follows this by gathering information from a random adult in the household. Results from the SHS are reported annually.

Since the survey was introduced, a number of changes have been made to its structure and content. From 2005, questions relating to volunteering habits were asked of the whole sample in quarter one and quarter two and of half of the sample in quarter three and quarter four, compared to the whole of the sample in 1999/2000 to 2003/2004.

Furthermore, a range of demographic related variables have changed to take account of shifts in local and national policies and circumstances. In terms of the volunteering suite of questions, the way volunteers are identified has changed as too have the options within questions. From 2005, additional questions were added on the attractions and barriers to volunteering.

These changes have informed the methods for this study and a description of these follow.

Study Methods

The study programme involved a four stage programme of work:

Stage 1: Preliminary review of data

Stage 2: Statistical analysis of data from the Scottish Household Survey 1999-Q3 2006

Stage 3: Statistical analysis of data from other sources

Stage 4: Focus groups and telephone interviews with volunteers, former volunteers and non-volunteers.

A summary of each of these stages follows.

Stage 1: Preliminary Review of Data

A preliminary review of the quantitative data relating to volunteering in the Scottish Household Survey 1999 to quarter three 2006 was conducted.

Using the SHS Lite dataset, common demographic variables (relating where possible to the random adult) across all SHS years were identified to ensure that any comparisons between survey data would be consistent and allow for valid time series analysis.

We also used the SHS Lite dataset to identify 'volunteering indicators', that is, those variables relating to questions in the SHS that identify a respondent as a volunteer, and 'volunteering variables' - those variables relating to questions that gather information about the volunteering undertaken by the respondent, such as type, intensity and reasons for stopping and re-starting.

Variables selected are listed in Annex 4.

A number of processes were undertaken to ensure the validity of data reported on. Each of these processes is described below.

Data Cleaning

Prior to analysis, it was necessary to 'clean' the data and remove cases or values which would invalidate the results.

Of the 15,395 cases in SHS 2005, 5,235 individuals were not presented with the volunteering question (this is due to the fact that from quarters 3 and 4 of 2005, questions about volunteering were asked only of half the sample, - see paragraph 2.4) and as such the volunteer status of these cases is unknown. Therefore, these cases were omitted from all calculations.

Variables relating to 'total number of hours volunteering' from 1999 to quarter three 2006 were included in the values field of the dataset and as such resulted in anomalies when performing correlation analysis. To overcome this, all variables with a value of over 97 hours were set to 97 hours.

In the time series analysis, anomalies in data relating to total number of hours spent volunteering were cleaned. This included removing cases reporting 'less than zero hours' volunteering and setting cases reporting more than 97 hours volunteering to '97 hours' to ensure consistency across all years' data. In the 2006 dataset, all numerical values over 97 were set to missing.

Weighting of Data

All data relating to the random adult (that is, all variables focusing on random adult and starting with 'r') were weighted using the variable IND_ WT to take account of the different probabilities of selections of the random adult in households (there is less chance of a random selection of adult within a smaller household than a larger household).

Rationale for Analysis Models

For the purposes of this study, models of analyses used were cross-tabulation, multivariate analyses and correlations (where ratio scales [natural ordering] are used).

Cross-tabulations provide information on the absolute levels of prevalence of volunteering and have been reported on using 95% confidence intervals.

Multivariate analyses assess the relative influence of demographic factors on the outcomes of interest and have been used to report on the prevalence of volunteering and intensity of volunteering.

It had originally been intended to conduct regression analysis to identify factors associated with the type of voluntary activity engaged in (for example, fundraising, committee work and so on). However, question Vol6 includes 18 specific categories relating to types of voluntary work. Therefore, modelling the probability of an individual participating in just one activity would in itself generate 17 tables of regression estimates. The situation is further complicated by the possibility that an individual may engage in a combination of multiple types of volunteering. Consequently, the number of unique combinations of volunteering activity is potentially huge. Given this, this analysis would have been of little value, as a result of the very large volume of results it would have produced, and the consequent difficulty of interpretation.

Further descriptions of the models of analyses used in this study are presented in Annex 3.

Other Issues

A range of demographic variables were selected to develop a profile of volunteering based on the 2005 survey. The variables selected included those which would allow an analysis of the key groups in the volunteering strategy (those from lower socio-economic backgrounds, the unemployed, the long-term sick and disabled, the poor and those lacking in formal qualifications).

Questions relating to educational and professional qualifications were only asked of those aged from 16 and not retired. Thus, the information presented in this report relating to qualifications excludes volunteers who are retired.

As the 60-74 age group accounts for on average 22.4% of volunteers, and considering that volunteers who were retired were not asked about qualifications they hold, data regarding qualifications has not been gathered from a large proportion of volunteers. Therefore, this analysis does not provide an accurate picture of the qualifications held by volunteers.

Further, the variable identifying social class of respondents (h_nssec2) was missing for over one third of the 2005 valid sample (10,160) and was therefore excluded from the analysis.

Stage 2: Statistical Analysis of Data from the Scottish Household Survey 1999-2006

This stage was based on the first two study objectives: to carry out a detailed analysis of volunteering; and of the questions which identify potential attractions or barriers to volunteering. The outcomes from this stage were envisaged as:

• a profile of volunteers
• time series analysis of volunteering 1999 to 2006.

Profile of Volunteering

To develop a profile of volunteers, this stage began by considering common variables identified in Stage 1 and those that were relevant to the study but unique to a particular SHS year(s).

A new suite of volunteering questions was introduced in 2005. To ensure consistency in reporting, the current profile of volunteering was based on 2005 data. A list of variables used in 2005 to identify volunteers, develop a demographic profile of volunteers and of volunteering habits can be found in Annex 4.

The variable identifying volunteers was used to isolate the 'volunteering group' from the 'non-volunteering group' and analyses of demographic variables for each group was conducted (cross-tabulations and multivariate analyses).

Cross-tabulations and confidence limits were run for the type of volunteering undertaken (type of organisation volunteered for, type of organisational activity, and type of volunteering activity) against a range of demographic variables to further develop the current profile of volunteering.

A similar process was employed for variables relating to the intensity of volunteering (number of hours spent volunteering and frequency of volunteering). Cross-tabulations were run for frequency of volunteering against relevant demographic variables. Average hours spent volunteering were calculated against a number of relevant demographic variables and multivariate analysis was conducted on number of hours spent volunteering.

Attractions and Barriers to Volunteering

The question 'have you ever volunteered?' appeared in both the 2005 and 2006 surveys and questions following this were asked to participants reporting that they had volunteered at some point (not just in the past 12 months as the question identifying current volunteers was based on).

Therefore, comparisons between participants in 2005/2006 answering questions that followed could be drawn. The questions followed gathered information on reasons why a participant had stopped volunteering and reasons that could encourage a participant to re-start volunteering.

Results on attractions to volunteering drawn from 2005 and the first three quarters of 2006 are presented in Chapter 7. Results on reasons for stopping volunteering drawn from the 2005 survey only are presented in Chapter 8.

Time Series Analysis of Volunteering 1999 to 2006

Time series analysis began with calculating prevalence of volunteering from 1999 to 2006.

A profile of volunteers based on similar demographic variables used in the 2005 profile of volunteering was developed and each SHS year (from 1999 to 2006) was analysed.

Frequencies for the type of organisations volunteered for were run and reporting was based on 1999 to 2004 as variables were consistent. From 2005, the categories within the type of organisation variable were more extensive and so were excluded from this analysis.

Frequencies for type of activities were run for SHS years 1999 to 2004 as variables were consistent. Again, from 2005 the categories within the type of activity variable were extensive and so were excluded from the time series analysis.

Average hours per month and 95% confidence intervals were calculated using means and standard error of means calculations.

Stage 3: Statistical Analysis of Data from Other Sources

Survey data was gathered from the Home Office Citizenship Survey 2005 (the Citizenship Survey is now the responsibility of the Department for Communities and Local Government, and reports based on the 2005 data were published by the DCLG). Data from the Citizenship Survey was compared against the current profile of volunteering in Scotland developed as part of this study, and differences in prevalence rates were reported.

Analysis of the questions presented to participants of the Citizenship Survey, including language used and the way questions are framed, was also analysed against the volunteering questions presented in the SHS. Data from Volunteer Development Scotland's Annual Digest of Statistics 2006 was also included in this analysis.

Stage 4: Focus Groups and Telephone Interviews with Volunteers, Former Volunteers and Non-Volunteers

We undertook qualitative research (focus groups and telephone interviews) with current volunteers, lapsed volunteers (people who have volunteered in the past but no longer do so) and non-volunteers (people who have never volunteered).

The table below summarises the focus groups we undertook in April and May 2007.

Table 1 Summary of Focus Groups Undertaken April - May 2007

We had intended to run another focus group with non-volunteers on the Isle of Skye but only one person attended. We interviewed this person as an individual. We supplemented this with telephone interviews with five non-volunteers on Skye.

We had also intended to run two focus groups with former volunteers. However, despite contacting a large number of former volunteers identified through the SHS and enlisting the help of voluntary organisations, it proved prohibitively difficult to recruit participants for these focus groups. Therefore, it was agreed that we would undertake telephone interviews instead. We completed 20 telephone interviews with former volunteers in total - 10 in Fife and 10 in the Borders.

Therefore, in total the following numbers participated in our research:

• Current volunteers (Kirkcaldy and Skye): 16
• Former volunteers (Fife and the Borders): 20
• Non-volunteers (Skye and Glasgow): 20

We selected these locations to ensure a suitable mix of urban and rural locations.

Recruitment of Participants

Focus group participants were recruited through a mixture of respondents to the Scottish Household Survey who had given their consent to take part in future research, and local organisations who were able to recruit suitable participants on our behalf.

As an incentive, participants were given a £10 retail gift voucher and their travel expenses were reimbursed to a maximum of £10 per participant.

Telephone interviewees were all Scottish Household Survey participants who had consented to being contacted for future research.

Focus Group and Telephone Interview Topic Guides

Topic guides used in our focus groups and telephone interviews can be found in Annex 9.

« Previous | Contents | Next »