Qualitative and quantitative research designs are more similar than different
There has been a traditional divide between qualitative and quantitative research, and nothing can start, continue, or inflame an argument among research theorists than to say, with fundamentalist glee and certitude, that “My research design is better than yours.” However, this paper is not an exercise in fundamentalism. Nor is it meant as an exhaustive discussion of research methodology. Instead, this is a brief look at the topic where the arguments and content are kept purposely simple because this type of discussion can quickly become a morass of jargon.
The argument put forward here is that the distinction between qualitative and quantitative research may have had validity at the turn of the 20th century, but as ideas about research have continued to evolve and develop the distinction has become more historical than actual.1 Whether research is qualitative or quantitative, the techniques are far more similar than they are different and by maintaining the myth of incompatibility researchers may miss important ways of finding answers to their research questions.2
The reasons often forwarded for why qualitative and quantitative research are fundamentally different generally reduce to four areas: (1) Research methodology; (2) Context, values, and involvement; (3) Data, analysis, and participants; and, (4) A common error. Each of these areas is taken in turn and the assumptions exposed.
Research methodology is most often described as the overall philosophy underpinning research, whereas research methods are the practical guidelines or techniques used to produce research.3 Research methodology is covered here in just enough depth to debunk the differences between qualitative and quantitative research that are commonly stated. Those differences can be described as: realism vs idealism; causality vs interpretation; and, hypotheses vs description.
A very basic definition of realism is, “Things exist only in the real world” and, therefore, anything that cannot be observed through the senses is of no consequence. On the other hand, a basic definition of idealism is, “Things exist only within the mind” and, therefore, are open to interpretation.4,5 Realism is stated as the concept underpinning quantitative research, while idealism is the concept that is said to underpin qualitative research.5 But is this correct?
A brief thought about both definitions shows that a reasonable person can come up with examples whereby the basic definitions do not hold. For example, “Do thoughts exist?” – Yes – Therefore the realism definition lacks completeness. “If humans disappeared in the morning, would the world still exist?” – Yes, it existed before humans and it would be conceited to think it would end just because humanity ends – Therefore the idealism definition lacks completeness.
There are, of course, a multitude of definitions for what is realism and idealism.6 Why then is quantitative research said to be “realist” and why is qualitative research said to be “idealist”? It basically comes down to the assumptions which are made about the nature of reality and since philosophers have been arguing about the nature of reality for thousands of years, without coming to a conclusion, it is not likely that researchers will arrive at a conclusion any time soon. So the same argument of idealist versus realist is rolled out based on no real evidence.5-7 There is an alternate, however, where some authors have suggested that whether a researcher uses qualitative or quantitative techniques they are in fact most likely to be critical realists, i.e. some of our perceptions accurately represent the world as it is and some of our perceptions do not represent the world as it is.8
The second part on the philosophy of research is causality vs interpretation. Up until the early 20th century many scientists searched for causality, also known as cause and effect. However, in the 1930’s with the ideas of quantum mechanics and relativity gaining more ground, the Newtonian, mechanistic view of the world changed.1,2,9 Suddenly, scientists could no longer be sure of causality because observation of a phenomenon could change the nature of that phenomenon. As Kryburg has said, it became “questionable to what extent causality is of scientific interest.”10 And so the quantitative side no longer look for causality but deal in probabilities and correlations, and the predictive value of these. Even so, causality as the purpose of quantitative research is still put forward as being a stumbling block between qualitative and quantitative research.11
As for qualitative research, the emphasis is on the interpretation of how social reality is constructed or the cultural meaning of phenomena experienced by those who are in a study.12,13 However, the assumption that quantitative data does not need interpretation but simply manifests itself through mathematical means while qualitative research only requires interpretation is not true. Quantitative and qualitative researchers need to make judgements about their data in order to elicit new meaning, extract alternative meanings, and interpret results based on previous research. Any research without interpretation is simply disaggregate.14,15
This leads directly in to the area of theory and, by extension, hypotheses and deductive reasoning vs description, and also by extension, inductive reasoning. The qualitative camp state that data collected can only describe the situation as it is and that no theories can be developed. Where then does this leave the branch of qualitative research called Grounded Theory, whereby theories are developed based on the data collected?16 From the quantitative camp the argument goes that first you need to have a theory and from that develop a hypothesis that is to be tested. Yet there are examples of where this is not so. If you look at surveys, a quantitative technique, there is no need for a hypothesis or a theory – the point of a survey is to find out information and not to test a hypothesis.17
Both sides use a combination of induction and deductive reasoning. In fact, it is not too difficult to see that induction and deduction are parts of the same process (Figure 1).18 For example, in epidemiology diseases are observed, patterns of disease are detected, tentative hypotheses are postulated about the underlying cause or causes, and theories of the disease are formulated: all inductive reasoning. From there, the theory is tested based on exposure and non-exposure, results are observed, and the theory of the disease is rejected or not rejected: all this is deductive. So why should there be a limit to our understanding based on to one type of reasoning over the other?
Context, Values, and Involvement
The second set of difference between qualitative and quantitative research can be summed up as context, values, and involvement. In the quantitative camp, research is supposed to be conducted independent of context, be free of societal or cultural values, and the researcher is detached from, or not involved, in the process. In the qualitative camp the research is said to be context dependent, societal and cultural values are present and explicitly stated, and the researcher is involved in the process.5,11
However, all research has a context. Quantitative research can attempt to control for this by limiting the context through controlling variables but in some quantitative techniques, such as developmental studies, this is not possible so that the context of the research becomes more important. Qualitative research does not attempt to control for context and it is through the context of the research that the research gains value, however, qualitative research can not always be said to happen in naturalistic setting, e.g. focus groups.2,19 Therefore, whether a researcher has a qualitative or quantitative focus, they approach the problem by creating a controlled environment, either purposely or as a natural consequences of their actions, in order to accomplish their research which is then extrapolated to a more complex environment or real world situation.8
On the topic of societal and cultural values, no research is value free.11 Researchers bring their own ideas, influences, and personalities into the research project: after all researchers are human. Where quantitative research has learned from qualitative research is that these things are present and need to be accounted for in the way research is conducted. In fact, there has been a movement in health research to publically register randomised controlled trials before they begin, so that the procedure, influences, and the veracity of the results can be publically determined.20
Equally, qualitative and quantitative researchers are deeply involved in their own research. It is more a matter of when this occurs. For the quantitative researcher their involvement is notionally suspended as the data is collected, i.e. the researcher does not influence or attempt to interact with participants in any way that may affect the results. But this has more to do with introducing as few biases into the research than a lack of wanting to be involved. Meanwhile, the qualitative researcher is notionally involved the whole way through their research but it is not necessarily the case in large projects where the researcher may not be involved in all or any interviews.2,5
Data, Analysis, and Participants
The third group of differences put forward are that qualitative research uses words as the data, thematic analysis of the data, and has few participants whereas quantitative research uses numbers as the data, statistical analysis, and has many participants. Again, on the surface, this appears true but with a little digging the distinction is difficult to maintain.
Quantitative studies can have one participant, i.e. a case study or single-subject designs.17 Likewise, the reason qualitative studies can have fewer numbers of participants is a matter of thematic saturation, i.e. participants are recruited until themes presented by participants have been voiced or exhibited on at least two occasions.12 In this way, recruitment to qualitative studies is based on the information retrieved from interviews and observation rather than on a predetermined sample size.
On the data and analysis side, the distinction is a matter of precision rather than use of different data and analysis. If the research requires a high degree of precision, then numbers and statistical analysis may be the requirement. However, if the degree of precision is not as important and the views of the participants are of more value, then the use of words and thematic analysis, or another qualitative analysis technique, are more useful. Also, where the subject being studied is too complex to reduce it to quantitative data, then it is better to allow that complexity to stand and to analyse the data in a qualitative manner. All this is true whether the research technique employed was primarily qualitative or quantitative.2,15
A Common Error
There is one common error that is made by both camps and it has been alluded to in the previous sections without spelling it out in detail: Both camps assume that there is only one facet to the other. In other words, the qualitative camp appear to assume that there is one, stereotypical quantitative design and the quantitative side appear to assume that there is one, stereotypical qualitative design. However, both camps have a number of research designs at their disposal (Figure 2).
Figure 2: Research Designs
With such an array of research designs to use, it seems strange to begin a research project by stating that design “X” is the one to use before fully determining the question being asked. Furthermore, a research objective, purpose, or question is normally stated in a way that is independent of the research method employed.15 Therefore, a better strategy is to concentrate on the question that is being asked and from there to determine which is the best research design or designs to answer the question.21,22 As an analogy, choosing a research design first and then working on your question is like choosing a car as your form of transportation and then deciding that you would like to go on an overseas trip. Surely the better strategy is to decide where you want to go and then decide which is the best way to get there?
This is not to say that a single person should be or can be an expert in all areas of research, or will not have a preference for certain types of research.2 However, a researcher should have enough knowledge to know that different research designs exist and what they can achieve. Also, a research project may require more than one research design to satisfy the research question or questions being asked. Such a realisation could lead to greater co-operation between researchers and more comprehensive results being obtained.
This common error is also present in the assumption, largely health research arena, that there is one hierarchy of evidence that will satisfy all research questions. Normally the hierarchy is stated as meta-analysis at the top of the hierarchy, followed by randomised controlled trails, and so on (Table 1). The assumption being that methods further up the hierarchy are better and produce better results than those at the bottom.23 But is this correct?
Table 1 – Hierarchy of evidence for quantitative studies
If the example is a quantitative-type question like, “What evidence exists for ultra-sound to be used to speed up the recovery from soft tissue injuries?” Then there is an argument that the hierarchy is appropriate. The researcher tries to find meta-analysis papers, randomised controlled trail papers and so on to answer the question. But what if a different questions is asked such as, “How do primary carers from different cultural backgrounds cope long term with family members who have an acquired brain injury?” then the researcher’s hierarchy may consist of generalisable studies, conceptual studies and so on.24 Thus the hierarchy of evidence depends on the research question being asked.
Researchers and funding bodies need to be more flexible in their understanding of the appropriateness of a research design for a particular research question and whether that question is worth asking and investigating. After all, evidence based practice, and the call for medicine and other health professionals to base their practice on the best evidence available, includes the best diagnosis and treatment currently available as well as the “thoughtful identification and compassionate use of individual patients’ predicaments, rights, and preferences in making clinical decisions about their care.”25
Although this paper argues that qualitative and quantitative research are far more similar than they are different, that is not to say that there are no differences. All research designs have their strengths and weaknesses and it is up to the researcher to be aware of those strengths and weaknesses.
Finally, one of the major reasons for the continuing divide between qualitative and quantitative researchers is that qualitative and quantitative research are still taught as being fundamentally different.22 However, since this is not true, now is the time to teach the variety of research designs to choose from and to base that choice on the research question.
|Crowe, M., Sheppard, L. - Invited Editorial: Qualitative and quantitative research designs are more similar than different. The Internet Journal of Allied Health Sciences and Practice, Volume 8 Number 4. Oct 2010.|