“There are three types of lies – Lies, damn lies and statistics”
– Benjamin Disraeli
Statistical analysis forms a critical part of biomedical research since it forms the basis for scientific claims. The validity of research could be hampered due to misunderstanding or misuse of statistics which does not restrict to data analysis but can include in all the stages of research starting from designing to reporting of the study. Ignorance or lack of adequate knowledge in statistics can lead to unethical research publication due to the errors in statistics or misinterpretation of results.
Frequent errors in study design, sample size, wrong statistics method, and data interpretation can influence the outcome of the study. Statistical errors in published papers have been a topic of discussion in biomedical research and were observed as early as 1966 in a paper by Schor and Karten. They reported that only 28% of paper published in ten medical journals were statistically acceptable. The common design error is to have a small sample size and the inability to observe a reliable and important result.
A study on the statistical error and misuse in endodontic microleakage studies revealed that nearly 41% of articles had used inappropriate statistics. They also concluded that the use of appropriate methods will change the conclusion in 19% of articles published. Another study reported that nearly 51% of the articles published in ten dental journals between 1995 and 2009 had at least one statistical error.
Data interpretation and reporting also play an important role in biomedical research. Randomized trials are considered as gold standard for the evaluation of treatment. However, when a trial report no significant result, interpreting and reporting confidence interval (CI) allows the reader to assess whether the treatment does not truly have a meaningful clinical effect. A nonsignificant result in a clinical trial might lead to premature abandonment of beneficial treatments, if not interpreted properly. A recent systematic review observed that only 17 out of 99 randomized trials interpreted the CI with nonsignificant results.
Peer review shapes the statistics in the published paper which ultimately powerfully influences statistics methods used in future research in the scientific community. This forms only one component of the ecosystem, however, there are other approaches to address this issue is by (a) encouraging the researchers to get trained in statistics and improving the statistical education, (b) seek help from a statistician from the stage of designing and planning of the research, (c) inclusion of statistical experts in the journal editorial panel. A survey among difficulties experienced by endodontic researchers has shown that 80% of researchers had difficulty in performing statistical analysis and 77% of researchers took help from statisticians. It was also observed that researchers with more than 21 years of experience had lesser difficulty in statistical analysis. By encouraging and incorporating/increasing the statistical training period in the curriculum will reduce the statistics misuse/errors. The adoption of statistical review in biomedical journals has been observed at least from the early 1970s. A recent study observed that only 23% out of 364 of top medical journals used statistical review for all articles. They also indicated that statistical review leads to important changes in 50% of manuscript and adds considerable value to the paper.
In India, we have nearly 300 dental colleges and the volume of research output is massive. It is vital that the young researchers need to be aware of the consequences of misuse of statistics. Thus, endodontology journal has taken the initiative of addressing this serious concern by including a new section in every issue for the year 2023 under the title “Importance of Statistics in Endodontic Research.” Experts having knowledge in dentistry as well as statistics will share their experiences which will help the researchers. The journal also aims in creating a panel of ad hoc reviewers for statistics which can enable a better research output. Researchers should understand that statistics is a powerful tool and absolutely indispensable to produce meaningful research.
1. Altman DG. The scandal of poor medical research. BMJ 1994;308:283–4.
2. Schor S, Karten I. Statistical evaluation of medical journal manuscripts. JAMA 1966;195:1123–8.
3. Altman DG. Statistics in medical journals. Stat Med 1982;1:59–71.
4. Lucena C, Lopez JM, Pulgar R, Abalos C, Valderrama MJ. Potential errors and misuse of statistics in studies on leakage in endodontics. Int Endod J 2013;46:323–31.
5. Kim JS, Kim DK, Hong SJ. Assessment of errors and misused statistics in dental research. Int Dent J 2011;61:163–7.
6. Gewandter JS, McDermott MP, Kitt RA, Chaudari J, Koch JG, Evans SR, et al. Interpretation of CIs in clinical trials with non-significant results:Systematic review and recommendations. BMJ Open 2017;7:e017288.
7. Hardwicke TE, Goodman SN. How often do leading biomedical journals use statistical experts to evaluate statistical methods?The results of a survey. PLoS One 2020;15:e0239598.
8. Alim-Uysal BA, Goker-Kamali S, Machado R. Difficulties experienced by endodontics researchers in conducting studies and writing papers. Restor Dent Endod 2022;47:e20.