THE FUNDAMENTAL PURPOSE OF SCIENTIFIC RESEARCH

Expressed at the most fundamental level, the purpose of research is to enable us to understand our world better, that is, to describe accurately how things are, explain why things are as we find them, and predict future observations in terms of these explanations. There are other ways of understanding the world,, including guessing, faith, religion and belief in magic. Scientific research is distinguished from these in that it is carried out in a manner which does not just aim to describe and explain, but explicitly realises the danger of false observations and explanations, and sets out to minimise the opportunity for these. Scientific research is therefore carried out within a framework of principle and action which is designed to maximise the probability that our observations and explanations represent the world as it really is, and conversely, to exclude false descriptions and explanations to the greatest extent possible.

Scientific research therefore is built around the following framework:

• An observation is made
• A question is asked
• A suggested explanation is put forward to explain that observation
• Evidence is acquired to determine whether the suggestion holds up or not
• The evidence is evaluated and a judgement made.

Thus scientific research is based on a structured framework of sequential observation, explanation and experimental support. When research is carried out properly, it is likely that the observations and explanations we produce are probably correct (although absolute certainty is never possible). In contrast, poorly conducted research has a high probability of producing false results, in other words, observations or explanations which do not represent the world correctly. It is far easier to perform bad research than good research, and it is essential to understand why research has to be so carefully and insightfully conducted.

Not all research follows the order of activities described above exactly as shown. There are different types of research problem with different methodological approaches to answer them. They are all however based on the principle of of making observations which are accurate, and, in many cases, providing explanations which are valid. Bad research in contrast has a high probability of producing inaccurate observations and faulty explanations. All good research is characterised by a methodological approach which specifically recognises the opportunity for false observations and false explanations, and is designed to minimise the risk of these.

Clinical research in particular lends itself to the use of different methodological approaches to answer different types of research question. These are often complementary, allowing one to understand a particular set of circumstances better through a combination of approaches rather than reliance on one approach alone. Often the approach is sequential. A common example is where a descriptive study is performed first in order to establish exactly what the observations are that we would like to explain; this can then be followed by an experimental study to determine whether the explanation we come up with is actually supported by subsequent observation (i.e. testing) or not.

Different approaches to the same question may enable a richer, more complex understanding of a situation than any one approach alone. For example, consider the following situation: A number of doctors report that their patients do not appear to be satisfied with the results of a particular form of treatment. We therefore form the opinion that this treatment may not be optimal. (It is however at this stage still just an opinion, subject to bias and quite possibly incorrect.) Consider the following two research studies carried out to throw more light on this situation.

Study 1

• We ask ourselves: Are patient satisfied with the results of treatment?
• The study is carried out as follows. One hundred patients are provided with a questionnaire and asked to answer the question: Are you satisfied with your treatment? (Tick YES or NO).
• We find that ninety of the patients tick NO.
• Our answer to the question therefore is: As a group, patients are not satisfied with their treatment.

That however is all that we can say. Based on this study, we have no idea why they are not satisfied, making it almost impossible to design interventions to improve satisfaction.

Study 2

• We ask ourselves: How do patients experience their treatment?
• This study is carried out as follows: Eight patients sit down for a four-hour structured interview with the researcher. They have every opportunity to describe how they experienced the effects of treatment, what it meant for them, and, if they were not satisfied, the opportunity to share with the researcher why this is so.
• Given that only eight patients were studied it is almost always logistically impossible to question the larger numbers of subjects in this degree of detail), we cannot be sure of the exact proportion of patients who are not satisfied with their treatment.
• What we now understand with a high degree of reliability is how patients actually experience the treatment, and what the factors are that appear to explain their lack of satisfaction.

These two research studies are therefore complementary. Neither is unnecessary as they provide two different types of information.

• The first gives us a definitive answer as to whether patients are satisfied or not. The result confirms that there is indeed a problem, that it affects a large majority of the patients, and tells us that it is justified to seek to improve treatment. It does not however suggest what those improvements should be.
  
• The second provides us with a wealth of detail to understand what underlies the dissatisfaction. The second, having identified in great detail what the problems are, allows us to decide what sort of solutions are most likely to provide relief.

Thus the two together contribute to a solution. It would in fact now make sense to follow-up with a third study:

Study 3

• Motivated by Study 1, we change the treatment regimen to address the issues raised by the participants in Study 2.
• We then provide the changed treatment to a further 100 patients.
• We ask ourselves: Is the second group of patients more satisfied with their treatment than the first group?
• We think they will be, and wish to find evidence if this is so or not.
• We find that 90 patients now express satisfaction with their treatment.
• We therefore conclude (1) that the explanations for the dissatisfaction we came up with as a result of Study 2 are probably correct, and (2) that the solutions we came up with are effective.

Employing different methodological approaches have led to a deeper understanding of the problem we wish to study. The first and third studies are quantitative and exemplify research within the positivist paradigm: the first is observational and the third interventional. The second study is qualitative in nature and exemplifies research within the interpretivist or constructivist paradigm. Furthermore, we can go further and started to classify the reasoning which underlay the research in each study:

Study 1. The research question was: Are patients satisfied with treatment? The research paradigm is positivist and the research quantitative. The aim was to determine whether patients are satisfied with their treatment. The objective was to determine the proportion of patients who are satisfied with their treatment. The method was based on an observational study.

Study 2. The research paradigm is interpretivist and the research is qualitative. The research question was: How do patients experience their treatment? The aim was to develop an understanding of how patients experience their treatment. The objective was to record and analyse the patient experience in a small group. The method employed a focus group interview.

Study 3. The research paradigm is positivist and the research quantitative. The research question was: Is the changed treatment better than the original treatment in terms of patient satisfaction? The hypothesis was that the changed treatment is better than the original treatment in terms of patient satisfaction. The aim was to determine whether the changed treatment is better than the original treatment (in other words, to test the hypothesis). The objective was to determine the proportion of patients who were satisfied with the changed treatment and compare this with the proportion of patients were satisfied with the original treatment. The method was based on an interventional (or experimental) study.

CONCLUSIONS

The conclusions to be drawn from this page are:

• Scientific research has a very serious purpose, that is, to describe things as they really are and not as we think they are, guess they are or would like them to be, and to do the same for the explanations we put forward to explain how and why things are as they are.
• There are a range of different ways of conducting scientific research, each is suited to a differently-phrased research question and taken together they complement each other.
• A sequence of studies (such as those in our example above), where the studies logically complement each other or follow each other, provides a far deeper understanding of how things are than do single studies. (This is essentially the difference between the scope of a Masters project and a PhD project: the first is typically based on a single study, whereas the latter comprises a number of studies which, taken together, provide greater insight into the solution to an overarching research question than do individual studies.)
• A critical step in conducting good research is to define very clearly incorrect the research question, hypothesis (if there is one), aims and objectives, as the most appropriate method is entirely dependent on what these are.

We can now begin to discuss the basic building blocks of research, such as research paradigms, the principles behind research question, hypothesis, aims and objectives, a broad classification of methodologies, and the principles of the scientific method.

Next: Research Paradigms