5.3 Phase 1 – Semi-structured interviews
5.4.1 Questionnaire design and structure
To some extent, the findings from the interviews in phase 1 influenced the construction of several sections in the questionnaire. The interviews were able to provide a more detailed and insightful picture into public understanding and evaluation of RE technologies and surrounding context, but they could not systematically test associations between different beliefs.
The questionnaire comprised 9 pages of questions (plus one cover page), which were grouped into 6 sections. Each section is briefly described below, but detailed information about questions (e.g. wording etc.) can be found in the questionnaire in Appendix 5.3.
Further information on how these questions were used in the analysis can be found in chapter 8.
Section 1: Views about energy technologies
o First associations with ―renewable energy‖ (open-ended question). Answers were content coded as discussed in section 6.4.
o Positive and negative feelings towards RE (5-point scales)
o Favourability towards 10 energy technologies (5-point scale, ―never heard of‖ option). This question was replicated from Spence et al. (2010b)5. o Five matrix questions (agree-disagree scales) about specific beliefs and
characteristics of five RETs: solar energy, biomass, tidal energy (including one question about the Severn Barrage development) and wind energy. The
‗characteristics‘ statements were developed from previous literature and the P1 interviews. Some of the statements were common to all five
5 Spence et al. (2010b) conducted a UK national survey investigating public perceptions of climate change and energy futures from January to March 2010. The results for their national (N=1822) and Welsh sub-samples (N=260) will be a useful comparison for the findings from the Cardiff survey. The researcher (Christina Demski) was involved in the design of parts of the 2010 national survey and therefore had access to the original data set.
83 technologies; whereas others were specifically developed for a specific technology (e.g. ―Wind energy is very noisy‖).
Section 2: Views about climate change (mostly replicated from Spence et al., 2010b) o Belief in CC: yes, no, I don‘t know answer options
o Belief in human causes of CC
o Concern, worry and anxiousness around CC (4-point scales)
o 11 statements about impacts and nature of CC (agree-disagree scale). Five items measured scepticism and uncertainty around CC (e.g. ―most scientists agree that humans are causing climate change‖) which are used in further analysis (see chapter 8).
Section 3: Views about energy security
o Concern (4 point-scale; + ―I don‘t know‖ option) for ES aspects was measured using 10 items (2 items per hypothetical construct: dependency, cost, reliability, vulnerability of supply (chain), future supply).
o Belief in gas/oil and coal running out (answer options from ―in the next 10 years‖ to ―never‖ including ―I don‘t know‖)
As mentioned in previous chapters, measuring public concern over ES is relatively novel, therefore a new scale was created as part of this thesis. Appendix 5.9 summarises the very limited existing research and presents the thinking behind how the scale was created (including some piloting). It should be noted that Spence et al. (2010b) also used an ES scale derived from the same piloting; however, the scale used in this research has been altered to further include more items (see chapter 8 for details). The use of this scale and any possible further development are discussed in the following chapters.
Section 4: Views about energy futures
It was decided to include a range of new and exploratory questions which were designed as a result of findings in the interviews.
o Preference for ES or CC goals was measured using a 5-point scale were 1-2 indicates a preference for ES, 3 indicates equal importance of both and 4-5 indicates a preference for CC.
o Respondents were asked to provide their ideal and realistic RE contributions to the energy mix in 20 years. These were coded on a 10-point scale representing 0-100% contributions.
84 o 5-point scales were used to measure support and opposition towards 5 energy source developments in your area (solar panels, wind farm, coal power station, nuclear power station, biomass plant).
Section 5: Views about the environment in general
o Environmental worldview was measured using the New Ecological Paradigm6 (NEP) scale devised by Dunlap & Van Liere (1978) and widely used in the field (e.g. De Groot & Steg, 2008; Poortinga & Pidgeon, 2003).
Following Hawcroft & Milfont‘s (2010) recommendations, the full 15-item scales was used with a 5-point response scale. For specific theoretical constructs measured see Dunlap (2008) or Appendix 8.17.
Section 6: Demographic measures and comments
o Gender, age, educational qualification, voting intention and working status.
Space was also provided for additional comments.8
Four questionnaire versions were used because the sections on CC and ES were counterbalanced to avoid influencing responses in section 4. The questions on positive and negative feelings towards RE in section 1 were also counterbalanced. The presentation of sections and questions was kept simple with clear instructions for all questions. All questions required participants to tick boxes only (Likert-type scales mostly), with the exception of the first question which asks respondents to list their first three thoughts when thinking of RE.
A cover page provided the respondents with basic and clear information about the questionnaire. This included number of questions, estimated time to take the survey, instructions on how to complete the survey, and reassurances that no special knowledge is required. It also included clear information on how to return the questionnaire. The Cardiff University Logo was clearly displayed at the top.
6 The NEP was originally called the New Environmental Paradigm scale, a name still commonly used.
7 The 7-even numbered items in the scale were re-coded so that higher scores indicated pro-environmental responses (Appendix 8.1).
8 129 respondents left comments. Some respondents critiqued the questionnaire (badly worded questions, ambiguous questions, not enough on nuclear power etc.) while others were very positive about it (e.g.
thought-provoking, well constructed questionnaire etc.). Perhaps unsurprisingly, some respondents expressed frustration at being constrained by the answer choices and wished to provide more complex or nuanced opinions. Respondents also commented about their concern over the planet, environment and climate change, and discussed their favourable views about various RETs and need for behaviour change (and other views on energy futures). A few expressed concerns about energy security (e.g. should be less reliant on other countries etc.) and others expressed their views on climate change being a natural phenomenon (or scepticism about its existence).
85 5.4.2 Sampling strategy and procedure
A total of 1,963 questionnaires were delivered during May and June 2010 in five areas in Cardiff (using the principles of stratified random sampling as much as resources allowed, Robson, 2002). The aim was to obtain a representative sample of the population within Cardiff. Every effort was made to achieve a sample as close to nationally representative as possible in terms of age, gender, and socio-economic background, including varying levels of education9. Sampling was restricted to residential houses and relatively safe areas.
Nonetheless, the five areas in which sampling took place were chosen carefully using data from the 2001 census.
Details about the five Cardiff sampling areas can be found in Table 5.3 and their locations are displayed in Figure 5.1. Areas were primarily chosen based on approximate social grade10 as measured in the 2001 census. Initially, the aim was to select two lower output areas per middle output area11 (e.g. in Area 002, areas 002A + B were chosen); however it soon became evident that not all streets were accessible or suitable and in some cases neighbouring areas were also included. Table 5.3 describes the main characteristics of residents in each sampling area. Hence areas A and B were chosen because the census data indicated a high level or residents from social grades AB and C1. Areas D and E were chosen because they have a higher percentage of people characterised as social grade D and E. Finally area C was chosen because it indicated a quite equal spread across social grades including social grades C1 and C2. The areas were also chosen because they represent residential areas in both central and suburban Cardiff (avoiding known student areas). The chosen areas also represent slight oversampling of lower social grade areas because it was expected that these would show lower response rates (Dillman, 2007).
9 It was not possible to measure the social grade of respondents and therefore these demographic factors were used as indicators instead.
10 Approximate social grade definitions (see www.neighbourhood.statistics.gov.uk for more detail) AB: Higher and intermediate managerial / administrative / professional
C1: Supervisory, clerical, junior managerial / administrative / professional C2: Skilled manual workers
D: Semi-skilled and unskilled manual workers
E: On state benefit, unemployed, lowest grade workers
11 The 2001 census data is split into geographical areas for ease of analysis called Super Output Areas, further divided into Middle and Lower Output Areas for comparison of key statistics in each area. Lower output areas have a population of around 1,500, Middle output areas have a population of around 7,200.
86 Figure 5.1. Sampling areas for the Cardiff survey (P2). ©Google – Map data ©2011 Tele Atlas
Table 5.3
Cardiff survey sampling area characteristicsa Sampling
area Lower layer census
areas (2001 census) Profile of area Area A
Thornhill 002A + B AB or C1 social grade, suburban, low unemployment, single/married, highly qualified, high home ownership Area B
Pontcanna 034 B +D (+ one
street in 034C) Mostly AB and C1 social grade, urban, low
unemployment, single (some married), highly qualified, mostly home ownership (some rented)
Area C Llandaff North
023 C +D + E + F Quite evenly spread across all social grades, but highest C1, D and E, urban, some unemployment, single/
married, lower qualifications, some home ownership/some council housing
Area D
Fairwater 027D + 031 A + B +
C Mostly D and E and C1 (C2) social grade, suburban, some unemployment, single/married, high % with no
qualifications, some home ownership/some council housing
Area E
Llandrumney 015 A + B (a few
houses in 007C) Mostly C1 and E social grades (little AB), suburban, high unemployment, single/married, high % with no
qualifications, some home ownership/some council housing
aInformation taken from www.neighbourhood.statistics.gov.uk; based on middle output areas.
Area A
87 It was deemed unsafe for the researcher to meet residents directly; therefore a part-postal procedure was used (questionnaires were dropped through the letterbox and either collected or mailed back). In most streets, every second house was provided with a questionnaire; with a few exceptions12. The procedure for data collection consisted of three major phases and follows that recommended by Dillman (2007). For example, pre-notification and reminder stages are used, which have been found to increase response rates. All materials (e.g. notification postcard, welcome and reminder letter) can be found in Appendix 5.4-5.6.
1. Pre-notification: An information postcard informing the resident(s) of the forthcoming survey was distributed, a week before the actual questionnaire was distributed. It informed possible respondents that their address has been chosen to take part in a Cardiff University survey on energy and environmental issues. It briefly introduced the survey and the value of their contribution. A list of addresses was careful recorded.
2. Distribution of questionnaires: A week after pre-notification, the questionnaires were distributed to the previously recorded addresses. The delivered A4 envelope included a welcome letter, which informed the participants of the reasons for the research and any information usually required on the consent form (e.g. anonymity). Respondents were asked to return the completed questionnaire either by using a pre-paid envelope or by leaving it outside their front door using a provided plastic bag that is then collected on a specified day. The collection date was specified on the front of the questionnaire. Each questionnaire had a unique (random) 4 digit number assigned to it, which was linked to an address. A letter in front of the number indicated the sampling area (e.g. A4213 would have been distributed in area A). All collected and returned questionnaires were taken off the address list. The addresses that did not return a questionnaire received a reminder (see below). After the reminders were delivered, the list of addresses was destroyed to anonymise the data.
3. Reminders: Approximately two and a half weeks after the initial questionnaire was distributed, a reminder and a second copy of the questionnaire were distributed to all those addresses that had not yet returned their questionnaires. At this point respondents were asked to use the prepaid envelope to return the completed questionnaire.
12Sometimes more houses were skipped if they were inaccessible at the time, for example, if dogs were present.
88 The aim was to deliver a total of 2000 questionnaires, 400 in each area. However the real number of distributed questionnaires deviated from this mainly because return to some addresses after notification was not possible due to safety concerns (e.g. in one case there was a disturbance on the streets, blocking about ten houses). The next section will outline exact distribution and response rates in each area. Questionnaire distribution and collection took place over a 6 week period and split in half, doing 1000 distributions a week. The data collection phases for each of the areas can be viewed in Table 5.4.
Table 5.4
Summary of data collection phases in May/June 2010 for the Cardiff survey Week 1
aCollection did not take place on the 31st May 2010 because it was a Bank Holiday, therefore the second phase of collection was pushed back to Tuesday/Wednesday.