A comparative analysis of quality of reporting statistics in two Indian journals

Abstract

Purpose: The importance of statistical analysis in medical research papers is ever increasing, hence, evaluation of statistical validity is crucial when evidence based medicine is highly valued. Studies with poor methodological quality and poor statistics cannot prove or disprove study hypothesis with certainty. This study was designed to evaluate, analyze and compare the reporting of statistical methods and errors in articles published in Indian Journal of Pharmacology (IJP) and Journal of Association of Physicians of India (JAPI). Materials and Methods: All original articles published in IJP and JAPI from January 2009 to September 2014 were reviewed and evaluated by using a checklist which included type of statistical test, common errors, etc. The statistical software used for analysis of data in these articles were also reviewed. Results: Three hundred articles (IJP=154; JAPI=146) were reviewed. The most commonly used statistical test in IJP was one-way ANOVA (53.8%) as compared to Chi-square test (50.6%) in JAPI. The statistical software used for analysis was mentioned in 43.5% and 50.7% articles published in IJP and JAPI respectively. The most commonly used software was GraphPad Prism (66.4%) in IJP and SPSS (67%) in JAPI. Statistical errors as per the checklist were more common in JAPI (63.5%) as against 49% in IJP. Use of mean+SE instead of Mean+SD was the most common statistical error in IJP (51.9%) whereas failure to mention the type of 't' test was the most common error (38%) in JAPI. Conclusion: Statistical errors are common in IJP as well as JAPI. To elevate the quality of articles published in Indian journals, every article must be sent for statistical review.