Research design essentially refers to the plan or strategy of shaping the research, or as Hakim puts it:
Design deals primarily with aim, purposes, intentions and plans within the practical constraints of location, time, money and availability of staff. (1987, p.1) Just as there is a wide variety of views as to what research consists of (compare the positivist and interpretive positions assessed in Chapter 1), and great differences in actual practices as to what people research and how they do this, so there are alternative perspectives of what the process of undertaking research should actually look like.
The research process
Perhaps the first key point to note is that all research projects embody an argument.
For some (such as those using a broadly quantitative approach) the argument will be structured in the initial stages of the research. This general approach may be referred to as a theory-then-research method of constructing arguments (this approach is outlined in more detail later in this chapter).
For research that utilises a more qualitative approach, the argument proceeds incrementally, and is constructed in the course of the research itself – the research-then- theory approach (see later in this chapter). In such emergent research strategies, many of the questions, aims, and formulations of problems in the research will be developed in the data collection phase, interacting with the researchers’ initial ideas or hunches; thus the final design emergesthroughout the research. As Hakim states:
The builder, and the materials he has available, takes a stronger role in the design than in the usual architect-designed study. (1987, pp.37–8)
Related to this idea is the notion that there are different models of the research process. In the following pages, we shall look at two commonly used models.
Sequential model
This model suggests that research passes through specific stages in pursuit of (at least tentative) answers to stated research questions. In this model, the research process is considered to take a relatively fixed, linear path, with a clear start and end. Gill and Johnson (1997) provide an example of this sequential model (shown in Figure 3.1).
Identify broad area
Select topic
Decide approach
Formulate plan
Collect information
Analyse data
Present findings
FIGURE 3.1 THE SEQUENTIAL MODEL OF RESEARCH (GILL AND JOHNSON 1997, P.3)
Cyclical model
Another common representation portrays the research process as cyclical. Here, many of the same aspects of the research process are included, and in much the same order. Usually the cycle begins with a problem, and finishes in a tentative generalisation. This marks the end of one cycle and forms the start of another.
This process may continue indefinitely, reflecting the progress of scientific knowledge. In this respect, the process is said to be iterative. There is an implication, however, that the process might be entered at a number of points, and that the experience of later stages might lead to a reinterpretation or revisiting of earlier stages.
There is therefore no determinable chronological sequence to the process of research, and often there is an overlap between the different aspects of the research cycle.
As we shall see in Chapter 7, this is particularly the case for qualitative research, where data collection, analysis, and problem formulation are closely bound up with each other. For instance, analysis of early interview data may lead the researcher into revising her or his line of questioning in later interviews. Thus, in the course of the research process, questions may be reconsidered, revised, or even discarded as a result of earlier research. Figure 3.2 provides an illustration of the cyclical model.
So, the pursuit of knowledge is not necessarily as straightforward as the sequential model would imply. The reality for the practising researcher is that research will not always follow a clear and logical path, and at times it may even appear somewhat chaotic.
Problem
Theory Generalisation
Data analysis Research design
Hypothesis
Data collection Measurement
FIGURE 3.2 THE CYCLICAL MODEL OF THE RESEARCH PROCESS (FRANKFORT-NACHMIAS AND NACHMIAS 1996, P.20)
The place of theory in the research process
We have already seen that theory is at the heart of both the sequential and the cyclical models, although it is not necessarily the initiator of the process of research. There are two contrasting views of the relationship between theory and the research process.
Deduction
Usually we are seeking to explain some particular type of action or process, and often there are theories that have been developed already with the purpose of trying to do just that. You might find it useful to see if the theory can explain the action or process that you are interested in. In this way, we follow a deductive approach to our research, in which the theory defines what we look at and how we look at it (see Definition 3.1).
Induction
Sometimes, we come across a situation in which we are not attempting to test how useful a particular theory is, but instead are seeking to understand a particular phenomenon, and through this, trying to build up an explanation of it. In this process, often rather loosely termed induction, we begin with a rather general research problem, and in the course of collecting and analysing data, we look for common themes or patterns in the data. Ultimately, we aim to draw some conclusions about the issue we are investigating, and perhaps to develop a tentative theory of our own (see Definition 3.1).
Definition 3.1 Deduction and induction
Deduction: If you begin with theory, and use it to explain particular observation(s), this is known as deduction. Theory is applied in order to deduce explanations for the data. Basically, deduction begins with the construction of a theory or model, research is designed around the model, and data collected explains or refutes the model. This approach is often referred to as using the hypothetico-deductive method,
associated with the theory-then-research strategy, in which:
1. Theory is consulted, and then guides the formulation of specific research questions.
2. Research questions are constructed as propositions, or hypotheses, which are then tested with empirical data.
If the data that has been collected demonstrates that the theory is lacking in some way, we may conclude that it has been falsified in its
(Continued)
(Continued)
present form, and needs to be revised. This process of falsification forces us to look at data in other ways to improve theories.
Induction: Contrary to deduction, induction moves from a set of observations to a theory, and is at the core of social scientific theory development. Induction allows a theory to be constructed from emerging patterns in the research data. It is associated with an analytic-inductive method, which is part of the research-then-theory strategy:
The person doing such research assumes that he does not know enough before beginning his study to identify relevant problems and hypotheses in the organisation chosen for study, nor to recognise valid indicators of the theoretical variables in which he is interested. He believes that a major part of his research must consist of finding out what problems he can best study in this organisation, what hypotheses will be fruitful and worth pursuing, what observations will best serve him as an indicator of the presence of such phenomena as, for example, cohesiveness or deviance. (Becker and Geer, cited in Burgess 1982, p.239)