Article first published online: 4 SEP 2002
Journal of Advanced Nursing
Volume 43, Issue 3, pages 219–220, August 2003
How to Cite
Watson, R. (2003), Editorial. Journal of Advanced Nursing, 43: 219–220. doi: 10.1046/j.1365-2648.2003.02701.x-i1
- Issue published online: 14 JUL 2003
- Article first published online: 4 SEP 2002
Scientific methods are the only credible way forward for nursing research
What is the place of scientific methods in nursing research? It is very hard to imagine that anyone could argue against them or that anyone could argue exclusively for another approach to investigation in any field. Science rests upon the ability to measure. Measurement is the antidote to subjectivity: it settles arguments and suggests specific, new lines of inquiry. Subjectivity lies at the heart of non-scientific methods, and subjectivity is what scientific methods expressly aim to avoid.
By scientific methods I mean, essentially, anything that involves measurement, objectivity and rigour, including randomized controlled trials (RCTs), surveys, taxonomies, systematic reviews and so on. Anything where the method used can be clearly explained, repeated and potentially, replicated. If these three features – explicability, repeatability and the potential for replication – are missing from a research approach, then that approach is not scientific; and that approach, I contend, has little, if any, place in nursing research. Therefore, all forms of research involving subjectivity – such as phenomenology, grounded theory and ethnography – are not scientific.
To explain something means to tell someone else precisely how you did something. Scientific and non-scientific research can both pass this test. Explicability is crucial in science in order that the next feature may be apparent, repeatability: that is, the ability of someone else, based on your explanation, to do exactly the same as you did. Scientific research passes this test; non-scientific research does not. Only if something can be repeated is there any chance that the findings might be replicated or, indeed, refuted. A further feature of scientific research is that another investigator can make changes to what you did in order to improve upon what you did, deliberately, to find out something new. Non-scientific methods – which I'll come to later – can be explained but are impossible to repeat since they rely upon, even glorify themselves in, subjectivity. The chances of replication are negligible. As for refutation, I am not sure that it has ever happened in non-scientific research. Qualitative findings appear to be accepted without question; used to build a subsequent study; and a whole field of ‘knowledge’ builds up without ever being challenged.
Measurement is only appropriate if it is reliable and valid. In scientific terms, validity is predicated upon reliability: if you are not measuring the same thing each time you make a measurement, then what you are using to measure what you are purporting to measure simply cannot be valid. However, those with a predilection for non-scientific methods make vague claims about reliability and validity – and even get them the wrong way round. For instance, I have heard qualitative researchers say that it is possible to have validity in qualitative research, without having reliability. This is simply not possible. If a method is not reliable it cannot be used by two people, or on two different occasions by the same person, to produce the same results. Surely this should apply to qualitative methods every bit as much as it applies to scientific methods? If the method or measurement does not produce the same result for you and me, then how can the result mean the same thing to you and me? How can the results be valid? Subjectivity is acceptable up to a point: it stimulates debate and leads to new lines of enquiry. However, why dress up a method in quasi-scientific terms if it does not produce the same result for two people?
Explanations aside, it strikes me more and more that, in non-scientific research, reliability means ‘I think the same thing on two separate occasions’ and something is valid simply because ‘I think it’. Such an approach to research, which is totally internally valid, has absolutely no external validity, and external validity should be the gold standard for all research. Without external validity and the hope of generalizability – even to a limited group of subjects, as with the RCT – there seems little point to the research.
What are the alternatives to scientific methods? A whole range of qualitative methods are on offer: phenomenology, ethnography, postmodernism, discourse analysis, grounded theory, and all the rest. But, really, they're all the same in the sense that they are all subjective: they start with loaded dice. There is nothing more dangerous than your own opinion in research, especially when it is untamed by objective, reliable and valid methods. The most extreme examples are qualitative research methods taking an ideological label. For example, when did a Marxist researcher not come to the conclusion that their research demonstrated class oppression? and likewise, when did a feminist researcher not come to the conclusion that patriarchal oppression was not at work? Those of us adhering to scientific methods are accused of ‘positivism’ and ‘reductionism’. It is a fact, however, that scientific methods have not been tried and found wanting in nursing research: instead, they have been found difficult and not tried.
There is yet another criticism to be parried: that of decontextualization. Those of us using scientific methods not only take an unpopular stance, we apparently ‘decontextualise’ by following the scientific method. Can research be carried out any other way? The postmodernist-phenomeno-Marxist-feminstic-discourse analyst rightly tries to contextualize everything – not a criticism in itself – but claims that research cannot be decontextualized: there is nothing to be learned about the whole that cannot be learned from the whole and not from the parts – we are more than just the sum of our parts. This is like going to a football match and trying to work out the result by investigating how happy the spectators are. I daresay a great deal could be learned, but why not just measure the number of goals scored by each side and add them up? Give or take a few visually challenged referees and linesmen, it's a fairly reliable test of who won, and it's utterly valid once the referee has blown the final whistle. Making such a measurement tells us the result in an instant; it cuts through spurious and subjective opinion; and it reaches a conclusion long before the non-scientist has arrived at the Nirvana of ‘saturation’.
I admit that non-science may have its uses. For example, non-science can tell us what suffering is like; however, only science can tell us how to alleviate it. Science has cured diseases, while non-science can only look for new ways to deconstruct the experience. Science has taken us to the moon, and back, while non-science can only dream of what the moon might look like. If nursing is to make progress then we must decide if we wish to head for the moon or look at it from a distance. What applies to other fields of investigation involving human subjects – medicine, psychology and sociology, where the scientific method has served us well – applies equally well to nursing. We must not set ourselves aside, we must compete for our place at the high table of credibility and the only way in which to do that is to carry out nursing research using scientific methods.
This editorial is based on the author's contribution to the debate on the motion that ‘Scientific method is the only credible way forward for nursing research’ held at during the Royal College of Nursing of the United Kingdom Research Conference at University of Manchester Institute of Science and Technology, Manchester, UK on 10 April 2003.