2. RESEARCH METHODS
2.4. RESEARCH INSTRUMENTS AND INTERPRETING DATA
2.4.1. Qualitative or quantitative methods
Although qualitative and quantitative methods are associated with the different paradigms for research, many of the tools for gathering data may be used for obtaining either type of data. Choices of methods to use for research depend upon the field of research, as well as the underlying paradigm. The researcher must be aware that different methods inherently bring assumptions and these have implications for their use. The main objective is that a method should be appropriate for helping to provide a good understanding of the particular area of the research. A variety of methods may be employed, each for providing different forms of evidence for a particular facet of the area under investigation (Flood and Jackson 1991).
Quantitative research is concerned with deriving facts, originating from the natural sciences, based on mathematics and statistics, analysis uses statistical methods to give a result that has a prescribed degree of correlation between variables. When results are derived from a representative sample of the population, the effect is said to be applicable to the whole population. On the other hand, qualitative research is concerned with deriving meaning, the how and why of effects, which may not necessarily result in findings that can be extrapolated more widely to the whole population (Brown and Dowling 1998).
Positivist research in education and information systems typically takes the form of experiments or quasi-experiments, either in the field or in laboratory settings. Experimental research relies on the control and manipulation of variables enabling the effects or differences to be measured. This represents a cause and effect model used in
the natural sciences. The main difficulty with this approach is being able to isolate the dependent variables (Bright 1991). Quasi-experiments are in the “field”, so experiments are in the working or learning environment, rather than in a laboratory. This overcomes the bias of choosing particular subjects for the experiments. However, there are ethical considerations in experimenting with subjects as they work, particularly if a control group are deprived of a facility that disadvantages them (Robson 1993).
One feature of the positivist approach is that results are empirical and that the procedures clearly state how results were arrived at and would enable another scientist to follow them and, hopefully, arrive at the same conclusions, i.e. validity (Crossan 2003). Whereas positivist experimental findings may be reproducible in the laboratory, it is less likely that experiments in real life situations will be reproducible, as variables that cannot be controlled may play a part.
The debate is whether we should search for results in the laboratory or in the field, i.e. in strictly controlled conditions or in the day to day workplace, and this centres on the validity and reliability of the research. In the laboratory the variables can be controlled, but the environment is not natural, so dealing with human subjects in a laboratory setting may not be realistic. The alternative of field experimentation will not yield quantitative results free from extraneous variables, but the findings may provide qualitative data, which can make the research more relevant. The context within which the experiments take place is also important, it may not be possible to extrapolate the results to a different context, and much of the research into team working has been based on teams of students in the college setting, assuming that findings there are applicable to the business setting, e.g. (Alavi 1994; Gatlin-Watts et al. 2007).
Research concerning new technology has been rooted in the positivist paradigm, e.g. Jones and Marsh’s investigation of trust development using computer supported collaborative working (Jones and Marsh 1997). Computing artefacts have been proved to “work” satisfactorily according to technical specifications using quantitative methods. Only recently has the effect upon users of these systems been considered, spawning the area of study of Human Computer Interaction and refining information
systems development methods (Bellotti and Blandford 1996; Agre 1997; Alm 2003). Within IS research Kaplan and Duchon (1988) advocate a move from quantitative to more qualitative methods, although positivism might be regarded by some as superior (Fitzgerald and Howcroft 1998), because it is more readily verified.
There are examples of positivist educational research, although mainly in schools, which use quantitative methods almost exclusively (Gomm and Woods 1993; Paulus 2005), or a study into group working, which is quantitative (Bahli and Buyukkurt 2005). Brown and Dowling (1998) suggest that educational research should be reflexive and interrogative, and suggest “3 R’s” of educational research: Reading, Processing and Writing. Their work is essentially positivist, following quantitative methods. However, a combination of quantitative and qualitative was used for a study in the HE sector on building online learning communities (Hill and Raven 2000). The issue of using solely quantitative methods in educational research is now being questioned, e.g. by Eisenhart and Towne (2003).
A post positivist approach can use quantitative and qualitative methods, sometimes called critical multiplism, being rigorous, precise, logically reasoned and supported by evidence, but not just what can be observed (Shaw 1999). Multiple perspectives help define goals and research questions, define methods and analytical techniques and interpret results (Fielding and Schreier 2001). There are now several studies where positivist research has been applied to operationalise the variables, then interpretive methods used to consider the variables in context, e.g. (Schrire 2006), so combining quantitative and qualitative methods is now acceptable research practice.
Surveys, questionnaires, interviews, focus groups, and observation can have a quantitative and qualitative component, and may be used for any data gathering activity. Whereas quantitative methods do play a part in some social research, such as market research, they are of limited use where the sample size is small, as is often the case in information systems or educational research. The results of quantitative questionnaires may not truly represent the whole population if one were to try to extrapolate the results. However, the results from such questionnaires may provide useful information, when combined with other means of collecting data as in
triangulation, which is explained in section 2.4.3. The tools for gathering data that were considered for this research and their features are outlined in the next sections.