RESEARCH DESIGN AND STATISTICAL ERROR OF ARTICLES SUBMITTED IN A PUBLIC SECTOR MEDICAL JOURNAL - A RETROSPECTIVE SURVEY
Design and Statistical Error of Articles
Keywords:Biomedical research, Statistical error, Statistical methods
Objective: The objective of this study was to identify the quality of statistical procedures used, the accuracy of study design, sampling and relevant statistical analysis
Study Design: Retrospective Survey study.
Place and Duration of Study: Army Medical College from June 2014 to Jan 2015.
Methodology: Three hundred and forty two original research articles published in Pakistan Armed Forces Medical Journal were reviewed. The selected research articles were all original articles. Case series and case reports were excluded.
Results: Among 342 articles, the major error was that the objectives, results and conclusion were not in accordance with each other in 115 (33.6%) articles. In 48 (14%) articles study design was incorrect and does not matched with the objectives. In 26 (7.6%) articles sample size was not appropriate as per the guidelines of journal. In 12 (3.5%) articles sampling technique used for sample selection was not mentioned. In 62 (18.1%) articles type of variables was incorrectly defined. In 6 (1.8%) articles descriptive statistics were not given at all. In 13 (3.8%) articles inferential statistics were not given articles. In 60 (17.5%) articles figures and tables were either not given or if given were incorrect. The pattern of statistical errors was almost similar in different years with a little bit difference but that difference could not attained any statistical significance (p = 0.460).
Conclusion: In order to maintain the standards of the medical writing in research with authentication the teaching of statistical methods should be enhanced.