4 Research Methods
Objective 5: Develop engaging and suitable research methods for data collection within the home environment
4.7 Research Methodology: Fixed vs. Flexible
Research methodology is categorised into those with a fixed design and those with a flexible design. Fixed designs are theory driven (Robson, 2011, p.66) whilst flexible designs may be driven by the data collected in order to formulate new theory. A summary of the main research strategies are given below, based on a combination of Gray (2009) and Robson (2011):
Experimental research: An approach that measures the effect of change on variables (a fixed design approach).
Ethnography: An approach seeking to capture, interpret and explain how people live and think, which is largely observational (a flexible design approach).
Phenomenological research: An approach that uses largely unstructured methods in order to gain detailed information of people’s experiences and perspectives (a flexible design approach).
Grounded theory: An approach that aims to generate new theory from the study results. It is particularly useful in new areas or those that lack descriptive theory (a flexible design approach). It is a systematic but flexible strategy. When using this approach, researchers should have a competent level of knowledge in the area, without being too immersed in the literature.
This research takes an approach akin to grounded theory. This approach supports the use of multiple qualitative methods (Gray, 2009, p.22‐23) with a strong focus on the use of interviews, where participants are selected using a purposive sampling strategy (Robson, 2011, p.279‐280).
4.8 Research methods
There are a number of research methods available that suit a qualitative, inductive, flexible approach. Qualitative research (Robson, 2011, p.279) commonly uses interviews for data collection and this method formed a large proportion of the data collection for this research. The main research methods used within this research are discussed below and are then depicted in Table 8 in Section 4.8.6.
Chapter 4: Research Methods
‐ 84 ‐
4.8.1 Interviews
Interviews involve the researcher asking questions and receiving answers from the person being interviewed (Robson, 2011, p.278). According to Stanton and Young (1999, p.87), ‘the main advantage of an interview is its familiarity to the respondent as a technique and this, combined with the face‐to‐face nature, is likely to elicit more information, and probably more accurate information. In addition, because it is administrated, the interviewer can pursue intriguing lines of inquiry.’ Interviews are a well‐known and well‐used data collection method, used in a range of fields and disciplines. The method is very flexible and adaptable, meaning the interviewer is able to modify questions as the interview progresses (in less structured approaches) (Robson, 2011, p.280). Interviewing is a method of collecting rich data in both qualitative and quantitative forms. The face‐to‐face nature of most interviews (other than online or telephone interviews) means that the interviewer is also able to identify non‐verbal cues from the participant, adding to the richness of the data (Robson, 2011, p.280‐281). The ability to be a good interviewer is a skill and requires training and experience to obtain all relevant information from participants. Careful preparation of interview questions is also important and a pilot study is an important component of the approach (Gray, 2009, Robson, 2002, Stanton & Young, 1999).
There are three different types of interview, structured, semi‐structured and unstructured (Robson, 2011, p.279‐280):
Structured: Comprises a planned set of questions, in a particular order.
Semi‐structured: The questions are planned but the order is flexible. The wording of questions may be changed and inappropriate questions can be removed.
Unstructured: The interview is informal and although there is a particular topic of interest, there are no set questions.
In all types of interviews, the validity, which is whether the findings are really measuring what they appear to measure (Robson, 2002, p93), can be addressed by ensuring that all interview questions address the main research question or meet
Chapter 4: Research Methods
‐ 85 ‐
the research objectives. However, the less structured the interview is, the more difficult it is to ensure validity, as the control the researcher has on the direction of the interview is reduced (Gray, 2009, p.375). Due to the time consuming and expensive nature of interviews, it is usual for a reduced sample size to be used, therefore lowering the external validity (Gray, 2009, p.376).
Reliability is also important, meaning the results of a study would be the same if the study was carried out again, because the results are not dependent on variable factors in the study design (Gray, 2009, p.376).
Interviews are the primary data collection tool used in this research, reported in Chapters 5, 6, 7 and 8 and they are carried out with householders, technology developers and other academics. Due to the in‐depth nature of the interviews that are reported in this thesis, particularly in Chapters 5 and 6, the sample size was limited. However, care was taken to ensure the research and validity of the interviews was not compromised, by ensuring the participants were carefully selected and the questions not only met the research aims but that the interviews could be replicated by using a semi‐structured protocol and question list.
4.8.2 Focus group [Workshop]
A focus group is an interview that takes place with a group of people and therefore is subject to the same advantages and disadvantages listed above. However, the added component of several people in attendance means the number of questions are limited, as discussions can take some time and the process requires effective moderation (Robson, 2011, p.279‐280). However, through discussion, it is possible to identify the most important topics to the group and the shared views within the group. This also means that extreme views are ‘weeded out’, which is not possible within an interview (Robson, 2011, p.279‐280).
In this research, the focus group method was used in the form of a workshop, where members of the CALEBRE project were gathered to discuss the role of UCD within the project through a number of activities. It has been described as a
Chapter 4: Research Methods
‐ 86 ‐
workshop because of the different activities involved and to emphasise the interaction and development of ideas within the session.
4.8.3 Case Study
Robson (2002, p.79) describes the case study method as ‘development of a detailed, intensive knowledge about a ‘single’ case, or of a small number of related cases’. Typically a researcher selects a single or small number of related cases within an area of interest. Data can then be collected via a range of techniques e.g.
interview, observation etc. Case studies are most often used qualitatively but can be used as both a qualitative and quantitative method (Gray, 2009, p.169, Robson, 2011, p.135). It is particularly useful when the researcher is trying to uncover a relationship between a phenomenon and the context. It is a useful method for answering ‘how’ questions (Gray, 2009, p.247).
In this research, a case study approach was used for investigating how energy saving technologies are developed and by using the case example of a heat pump.
This is discussed in Chapters 6 and 7.
4.8.4 Diary Study
Diary studies take the form of a self‐administered questionnaire when used as a research method (Robson, 2002). They provide the means of recording data from participants over time, without the intrusion of observation. There is potential for this information to be less reliable, as the participant is responsible for interpreting questions and recording data successfully (Robson, 2002).
This method is used within this research to record hot water use in households (Chapter 6). In this case, it allows the recording of data without intruding on the households’ daily routines. In particular, the measurement of hot water temperature would be very difficult to perform with a researcher present, due to obvious privacy and ethical issues.
Chapter 4: Research Methods
‐ 87 ‐
4.8.5 Survey
Surveys are used mainly to collect quantitative information and are described by Robson (2002, p.230) as, ‘The collection of a small amount of data in a standardised form from a relatively large number of individuals’. The main issue with this method is that there is no way of knowing whether respondents are being truthful in their answers. In addition, surveys often have a relatively low response rate and the validity of a survey will vary depending on the questions asked (Robson, 2002).
However, when conducted properly, with sufficient piloting, surveys can provide an effective and efficient way of collecting data from large numbers of people.
Surveys are used in this research to evaluate the principles of cross‐disciplinary working proposed in Chapter 8.
4.8.6 Requirement Trees
Requirement or objective trees are a useful way of stating and displaying objectives, including demonstrating the ways objectives relate to each other. The process of achieving the requirement tree is also useful for clarifying objectives and for helping the design team to reach a consensus (Cross, 2008, p.81‐82). Requirement trees aim to convert non‐technical, non‐quantified customer information and requirements into a set of product characteristics (Wright, 1998) and were therefore identified as an appropriate tool for translating and displaying insights from the user research into user requirements in this research. Requirement and specifications are a familiar tool to technology developers as they are commonly taught in engineering schools (Claridge & Berns, 1999).
Requirement trees enable the designer or UCD practitioner to organise and identify the relationships and interconnections between all objectives. When beginning at the top of the tree, every new box answers a ‘how’ question, however, when beginning the process at the bottom of the tree and working upwards, every new box answers a ‘why’ question (Cross, 2008, p.81).
Requirement trees are used in this research to facilitate the translation and communication of the user information collected in Chapters 5 and 6, into a form
Chapter 4: Research Methods
‐ 88 ‐
that was usable by technology developers. In the case example, this method was used to produce user centred design specifications for the heat pump to ensure user needs were represented in the design process. The process in which they are used is described in Chapter 6.
4.8.7 Research methods used in combination in this thesis
In this research, a combination of interviews, surveys, literature reviews, workshops and diaries have been used, as demonstrated below in Table 8.
Table 8: Research methods used in parts 1, 2 & 3 of the thesis
Interviews Workshop Survey Requirement
Trees Diary Study
It is often necessary or recommended to use triangulation in a study, where a combination of research methods is used (Wilson, 1995, p.1‐40). Gray (2009) points out that all methods have their individual strengths and weaknesses and the use of methodological triangulation can help to balance out these weaknesses. In addition, when using the same participants over an extended study period, the use of different methods can help to keep the participants engaged and interested.
Chapter 4: Research Methods
‐ 89 ‐
4.9 Sampling
Robson (2011, p.270) describes the idea of a sample as linked to the idea of population which is as ‘all cases’. It is often unrealistic to deal with an entire population, therefore a sample is a selection of the population. There are two types of sampling strategy, probability and non‐probability. In a probability sample, the probability of the selection of each respondent is known whilst in non‐probability samples, it is not known (Robson, 2011, p.270). Examples of types of probability samples (as listed in Robson, 2011, p.270‐277) are: Simple random sampling, systematic sampling, stratified random sampling, cluster sampling & multi‐stage sampling. Whilst examples of non‐probability samples are: quota sampling, dimensional sampling, convenience sampling, purposive sampling & snowball sampling.
For this research, non‐probability samples were used. The studies required participants to have particular characteristics and therefore the populations were specified and chosen to ensure these people types were included. For example, in the home improvement study, purposive (otherwise known as purposeful or judgment (Marshall, 1996) sampling was used, where the researcher actively selects the most useful sample to answer the research question, which may be based on a number of variables (Marshall, 1996). In this study (Chapters 5 & 6), the participants were required to live in Leicestershire and be an owner occupier of a solid walled house, therefore, the population only included those who could meet this criteria.
Specific sampling strategies are stated within the study chapters.
Chapter 4: Research Methods
‐ 90 ‐