Openness and clarity are essential in research reporting

Abstract

As nursing research continues to gather pace, we need occasionally to reflect on the crucial matter of how authors describe methods and methodology in published articles. As the Editor of INR, it seems to me that this aspect of nursing research is sometimes not well taught. Many manuscripts submitted to INR must be rejected because of lack of clarity or flaws in presentation. For example, we continue to receive many articles in which the author(s) describes the sample as ‘convenient’, but then proceeds with sophisticated but completely inappropriate statistical testing. This problem is so common that we have an INR standard letter to authors that says (I paraphrase): ‘It is recognised that in nursing research it is often difficult if not impossible to randomise a sample. Nevertheless authors should describe in detail the rationale for their sample selection from a total research population (e.g. all the staff nurses in three hospitals) and describe the participation rate both in raw numbers and percentages’. If a sample is truly ‘convenient’11 A useful summary of convenience sampling applied to research can be found at http://www.experiment-resources.com/convenience-sampling.html (retrieved 04.03.2011) where the following definition is provided: ‘Convenience sampling is a non-probability sampling technique when subjects are selected because of their convenient accessibility and proximity to the researcher.’ The summary continues that the most obvious criticism about convenience sampling is its sampling bias, because the sample is not representative of the entire population under consideration. (and this is actually rare) then the characteristics of the total population from which it is drawn cannot be known, and comparative evaluations through the application of statistical tests are inappropriate. In individual cases this is a decision that calls for the expertise of a specialist statistician. Once authors have clarified their methods of selection it is frequently evident that they have given the matter considerable thought but have not described the process clearly enough when ‘writing up’ the manuscript. In a previous Editorial I wrote about another problem, that of ‘salami slicing,’ i.e. when authors duplicate parts of the same study in different publications (Robinson 2009). This situation threatens to become more serious as nurse researchers increasingly gather large data sets in a single study. If they then proceed to publish on different aspects of the research, they must show clearly how they have disaggregated the raw data. Otherwise the evidence may be counted twice by other researchers who read the articles. The result may be a significant distortion of how much evidence exists on a particular subject. The findings of meta-analyses across a number of studies also may be distorted when authors are not aware that some results are being reported twice, or even several times over. In my earlier editorial, I cited an article by Ben Goldacre focusing on the need for transparency and openness in research reporting (Goldacre 2009). In a recent article published in The Guardian, he gives three examples of research misinterpretation which occurred because all the evidence was not made available, or may even have been deliberately withheld (Goldacre 2011). He repeats an earlier theme that we tend to trust authority and rarely question the evidence behind the statements of published authors. Yet, he says ‘once you lose trust’ (in an author) ‘any timesaving benefit from reading a précis of their work has been annihilated’. Writing for publication demands considerable skill. Authors must craft an article that is succinct yet still provides essential detail on how the study was conducted and the findings obtained. If authors fail to provide this information, an article may mislead rather than inform with new knowledge. In the previous Editorial [58 (1), 3] I welcomed Dr Susan Jordan to INR's Editorial Board as the ‘first person who is not a nurse’. This was, of course an error on my part, for which I apologise. In fact, the Editorial Board has two other members who are not nurses: Professor James Buchan of Queen Margaret College, Edinburgh, and Linda Carrier-Walker, ICN Director of Development and External Affairs. A long-time member, Professor Buchan contributes useful analyses on the nursing labour force and supports many authors through his insightful manuscript reviews. Neither Professor Buchan, Dr Jordan or Mrs Carrier-Walker are appointed as ‘honorary nurses’ but for the different perspectives that they bring, which are based on a good understanding of the nursing context.