Abstract <p>Drawing on many years of experience in editing for biomedical journals and lecturing on scientific article formatting for PhD students at Moscow State University, the author analyzes the current state of statistical analysis of biological, biomedical, and medical research results, in particular, data obtained in studies aimed at discovering drugs that supposedly slow aging and suppress the development of age-related diseases. He emphasizes that, in many cases, such analysis is either not performed at all or is performed incorrectly. Authors often analyze data that does not follow a normal distribution using the parametric Student’s <i>t</i>-test. Moreover, many researchers, having failed to obtain reliable results (<i>p</i> &gt; 0.05), are increasingly replacing the term “significance” in their articles with “trend,” which seems completely unacceptable. Although this problem affects not only biomedical but also other natural science research, it is especially important for biomedicine, as it is at stake in human life and health. It is proposed that greater attention should be paid to this issue by both reviewers of relevant articles and editors of the journals in which they are published. Furthermore, the importance of teaching the basics of statistical processing of research results to undergraduate and graduate students in biological, biomedical, and medical institutes, as well as relevant university schools, is emphasized.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Statistical Analysis in Biomedical Research: From Significance to “Trends”?

  • A. N. Khokhlov

摘要

Abstract

Drawing on many years of experience in editing for biomedical journals and lecturing on scientific article formatting for PhD students at Moscow State University, the author analyzes the current state of statistical analysis of biological, biomedical, and medical research results, in particular, data obtained in studies aimed at discovering drugs that supposedly slow aging and suppress the development of age-related diseases. He emphasizes that, in many cases, such analysis is either not performed at all or is performed incorrectly. Authors often analyze data that does not follow a normal distribution using the parametric Student’s t-test. Moreover, many researchers, having failed to obtain reliable results (p > 0.05), are increasingly replacing the term “significance” in their articles with “trend,” which seems completely unacceptable. Although this problem affects not only biomedical but also other natural science research, it is especially important for biomedicine, as it is at stake in human life and health. It is proposed that greater attention should be paid to this issue by both reviewers of relevant articles and editors of the journals in which they are published. Furthermore, the importance of teaching the basics of statistical processing of research results to undergraduate and graduate students in biological, biomedical, and medical institutes, as well as relevant university schools, is emphasized.