Intended for healthcare professionals

Views And Reviews Acute Perspective

David Oliver: Let’s argue about statistics

BMJ 2016; 355 doi: https://doi.org/10.1136/bmj.i5649 (Published 21 October 2016) Cite this as: BMJ 2016;355:i5649

Statistics for scientific validity of studies

Talking statistics by politicians is one part of the story only. They would just like to highlight their achievements in terms of certain outcomes which they are interested and if there is a negative part of the outcome, that is not likely to be highlighted. But for the sake of scientific validity, statistics is of importance provided appropriate statistics are used. Reviewers of scientific studies will not accept the data unless it is supplemented by appropriate statistics. However, to translate the scientific findings in front of policy makers one has to make simple statements related to the outcomes of the study which can be understood by the lay public including politicians who are not well versed with statistics. However, sometimes, too many statistics in scientific articles may be difficult for other readers. Hence, limited use of statistics, wherever appropriate, should be done. Sometimes, statistical findings may not be acceptable at all when data are loaded with bias and it might sound implausible biologically despite statistics showing an association. This should be interpreted with caution and the data need to be reevaluated.

Competing interests: No competing interests

25 October 2016
Mongjam Meghachandra Singh
Director-Professor, co-author: Assistant Professor
Reeta Devi
Maulana Azad Medical College, New Delhi
Department of Community Medicine, MAMC, New Delhi; School of Health Sciences, IGNOU, New Delhi