Discrimination of a frequently neglected statistical error in medical scientific research design

The author's Chinese name is Bi Jingfeng; Segment;

Author Shandong University of Traditional Chinese Medicine; Chengdu University of Traditional Chinese Medicine;

Literature sources are "Zhen Shi Traditional Chinese Medicine" and "Research on Medicine and Materia Medica in Li Shizhen", the email address of editorial department 10, 2008.

Journal honor: summary of the main contents of Chinese core journals: CJFD, the source journal of ASPT.

Medical statistics; Scientific research design;

In the past clinical research design, statistical errors are common, but they are generally easy to find and correct. The author recently consulted relevant medical research papers and found that there was a statistical error with a high rate of erroneous application, even many statistical professionals were no exception. Example: A researcher studied the effect of drug A on hyperlipidemia fatty liver rats, and designed the following experimental scheme: establish hyperlipidemia fatty liver rat model, intervene with high, medium and low doses of defatted capsules, and observe its effect on blood lipid of fatty liver rats through blood biochemical examination. Results: Jiangzhi capsule can obviously reduce the blood lipid of fatty liver rats, and the difference is significant compared with the control group (P < 0.05). Conclusion: Qu Zhi capsule has definite therapeutic effect on fatty liver in rats. In this design scheme, the researchers analyzed the variance of high, medium and low dose groups, methionine tablet group and natural recovery group according to the statistical method of multiple factors and one level. Looking carefully at the relationship between the treatment groups, in fact, this study mainly involves two factors: drug A treatment and methionine tablets treatment, while the high, medium and low dose groups of drug A are three levels of drug A, not three factors equal to methionine tablets. Table 1 comparison of blood lipids of rats in each group (X-S) mmol L- 1 group TC TG HDL-C recovered naturally 2.100.152.320.310.933 0.070A Low dose 2.06538.0000000000001 ...

DOI CNKI:SUN:szgy . 0.2008- 10- 139

Common statistical errors in medical research papers

The author's Chinese name is Li;

Author Baicheng Infectious Disease Hospital; Baicheng, Jilin;

Literature sources: Journal of Jilin Medical College, Journal of Jilin Medical College, editorial office mailbox, 02, 2007.

Journal honor: the journals included in CJFD, the source journal of ASPT.

Keywords medicine; Scientific research papers; Statistical error;

Statistical methods are often used in scientific papers to process, sort out and analyze data, so as to qualitatively or quantitatively explain some theoretical or experimental results. In this paper, the wrong statistical methods in some medical journals (1999~2000, 8 national journals ***60) are summarized and analyzed, so as to remind scientific and technological workers to use statistical methods reasonably, accurately describe, estimate, compare, predict and analyze, minimize the wrong application of statistical methods and improve the writing level of scientific and technological papers. The data of 1 is unreliable. Some data samples are small, some authors choose unrepresentative experimental objects, and there are many human factors. Some authors judge the results according to their own subjective expectations. Worse, they sometimes change the experimental data, resulting in large errors in some experimental results. 2. Lack of scientific statistical methods. There are many statistical methods, such as ratio, composition ratio, development speed, significance test method and so on. Sometimes improper calculation methods will directly affect the results or cause misunderstanding. For example, the connection and difference between rate and composition ratio are often misunderstood, and some authors only look at superficial phenomena without drawing conclusions through statistical methods. 3. The input of unified measurement lacks normative science. Only by calculating the statistical data correctly can we correctly reflect the real situation of things, but if the calculation is improper, there will be false appearances or wrong results. If we compare the non-standardized data, because the internal structure of the two groups of data is different, the knot ...

Docnki: Sun: JLDS.0.2007-02-0 18

Common statistical errors in medical papers and their countermeasures

The author's Chinese name is Yang Yunhua;

Authors: Tianjin Institute of Medical Science and Technology Information, Tianjin 300050;

Literature source: Chinese journal of medical scientific research management, editorial office mailbox, 02, 2004.

Journal honor: the journals included in CJFD, the source journal of ASPT.

Medical papers; Statistics; Common mistakes; Countermeasures;

This paper analyzes the common errors in the application of statistical methods in medical scientific research papers, improves the editors' ability to identify common statistical errors, ensures the scientificity, accuracy and credibility of scientific research papers, and strives to become a top-quality periodical.

DOI CNKI:ISSN: 1006- 1924 . 0 . 2004-02-0 15