WebmedCentral Comments
Back to comments
Submited on: 04 May 2012 05:22:12 PM GMT
Published on: 05 May 2012 12:12:52 PM GMT
In an open access-site such as we enjoy not have walls or shields made by hiding behind a computer screen, anonymous is not being accountable for their public review. Many of us joined this esteemed group because we tired of reviewers hiding behind a politics and or a cloke simply disguishing cowardess. This review, I reject because it is anonymous, the very issue all the popular Journals laden with structure but definitively lacking access to reviewers chosen without names or access.
When I attended graduate schools, I defended every thesis and I always signed my name to my work, and I am pleased to sign it to this comment.
Dr. Bill Misner Ph.D. {Emeritus}
drbill@omnicast.net
- competing interests: None.
- Invited by the author to make a review on this article? :
- Experience and credentials in the specific area of science:
Statistics: None to declare.
- Publications in the same or a related area of science: No
- References: None
All site content, except where otherwise noted, is licensed under a
This editorial article (NOT a review article) admonishes medical researchers to be diligent in the use of biostatistics in the design and analysis of epidemiological research. Drs. Sathian and Sreedharan cite a number of sources from the last 30 years which have demonstrated statistical inadequacy in the medical literature and remind us that the appropriate application of biostatistics can help investigators make the most of their research data. The topic of this article is continually relevant; as more new and sophisticated statistical tools are developed it becomes more important than ever to ensure that they are applied appropriately and that sound conclusions are drawn from them. Though the article is a timely and well-intentioned reminder of the importance of biostatistics in research, the article could use improvement and clarification.
As Drs. Sathian and Sreedharan are listed as being from Nepal and the United Arab Emirates respectively, English is most likely not their native languages. As such they should be lauded for a well-written article overall. However, there are instances in which the grammar could use clarification as the intended meaning is unclear. Most notably in the Introduction, page 2, half-way down the right-hand column: “Result part became poor because of the lack of knowledge in appropriate test for the analysis of data and the coding of data.” While this sentence is awkward, it is still understandable; the one that follows however is unclear: “If the researcher is not aware about the proper research design in descriptive studies, case control studies, cohort studies and clinical trials better to terminate the study rather than reporting clinical trials in the methodology part and the study will be a hospital based observational study.” Given these and other, more trivial errors, subsequent drafts would benefit greatly from careful proofreading by an expert in written English.
More important than grammar is the explanation of variable types towards the end of the Introduction section. Here the authors identify two types of variables, which they label “categorical” and “numeric”, while providing examples of each. These classifications are somewhat tangled however, as there is overlap and ambiguity in the descriptions of the types. It seems that there may be some confusion between numerical scale types (nominal, ordinal, interval, ratio) and numerical variable types (nominal, ordinal, ranked, discrete, continuous). The article could be improved by clarifying these concepts.
Finally, the article advises that the type of data at hand should determine the type of analysis employed. While this is certainly a sensible suggestion in as much as the choice of analysis should be constrained to those methods which are appropriate and valid for the data, I would point out that the analysis should truly be driven by the research question and the resultant hypotheses of the study.
Overall Drs. Sathian and Sreedharan have reminded us of an important aspect of biomedical research, namely our collective responsibility to use statistical analysis in a responsible and appropriate way. The authors point out that not only will this lead us to draw better conclusions, but it will also assist us in learning the most that we can from our data. That is sound advice for scientists of all nations.