« Previous | Contents | Next »
Listen
WELL? WHAT DO YOU THINK?
ANNEX A. DETAILS OF THE RESEARCH METHODOLOGY AND QUESTIONNAIRE DEVELOPMENT
The main sample for the study was a clustered design, using addresses taken from the small user section of the Postcode Address File (PAF). The sampling frame was stratified geographically by local authority and then by electoral division. Within each division, there was a further stratification by type of Enumeration District (ED). A total of 96 EDs were then selected as sampling points with a probability proportional to the number of addresses. All parts of Scotland including the Highlands and the Islands were included in the sample frame.
In each sampled ED, 30 addresses were selected, clustered by using every fifth address after a random start point. Where there were insufficient addresses within a selected ED to generate the necessary sample, the cluster was extended to include an adjoining ED of similar profile. At each address, interviewers had to screen for eligibility to remove non-residential properties from the sample. Where, for example, more than one flat was found at the exact sampled address, one unit was sampled at random using a Kish Grid of all units (whether occupied or not). The same process was used in the selection of one household in multi-household addresses. Respondent selection was again based on the Kish Grid, from a listing of all resident adults aged 16 and over. A copy of the paper Contact Sheet for the main survey is included in a
Annex D to this report.
An additional sample was drawn to boost the otherwise very small number of ethnic minority respondents in the final data - only 10-30 such interviews were anticipated. The 1991 Census data was analysed to identify those EDs with the highest non-white populations and, from a list of the top 30 such locations, every second was selected to form the booster sample - one more was later added. All of these EDs had, at the time of the Census, at least 25% of their population from ethnic minority groups - the average was more than 40% and the general expectation was that the proportions would have grown during the last decade.
In the booster EDs, 30 addresses were again selected and the sampling process was very similar to that used on the main study, although white-only households were screened out of the process. Again the ethnic minority booster Contact Sheet is included in this volume.
The questionnaire was initially designed by an advisory group, including representatives from the Scottish Executive Health Department, the Health Education Board for Scotland, Glasgow University, Edinburgh University, the Public Health Institute Scotland and the National Anti Stigma Campaign. The draft was further developed by NOP and a version was agreed with the Executive before piloting.
The tight timetable meant that the three pilot interviewers were briefed and debriefed over the telephone by the NOP executive team. Each of these fieldworkers carried out one full day of interviewing in different parts of Scotland including inner city, suburban and rural locations - the 16 pilot respondents were a broad cross-section of the adult population and included several ethnic minority adults along with a number of people with personal experience of mental health problems.
The pilot survey led to a number of changes to the questionnaire which was approved by all parties and finalised for operation via Computer Assisted Personal Interviewing (CAPI). The average interview length on the main stage of the survey was 30 minutes. The final questionnaire is shown in
Annex E in this report.
« Previous | Contents | Next »