PARAMETRIC STATISTICS APLICABILITY TO BIO-MEDICAL DATA ANALYSIS
Abstract
Parametric statistics calculates the statistical significance of dirences assuming normality of the random variables in question. Since all random variables used in real biomedical research are not normally distributed, it is necessary to validate parametric statistical methods, i.e. evaluate the errors in the calculation of the statistical significance of differences. Unfortunately this test is performed incorrectly. The article describes the causes of errors in parametric statistics and presents a software program which calculates the error value resulting from the abnormal empirical distribution of the mean from a set of observations.
About the Authors
A. N. Gerasimov
First Sechenov Moscow State Medical University
Russian Federation
Russian Federation
Andrey Gerasimov - Dr. Phys.-math. Sciences, Assoc. Prof., head. the Department of medical Informatics and statistics.
Р. 8, bld. 2, Trubetskaya str., Moscow, 119991
M. Yu. Serdyukova
First Sechenov Moscow State Medical University
Russian Federation
Russian Federation
Senior lecturer of the Department of medical Informatics and statistics of the state budget educational institution of higher professional education
References
1. Briko N. And., Pokrovskiy V.I. Epidemiological studies, clinical epidemiology and evidence-based medicine // sechenovskiy Gazette. 2010; 2(2): 106-111.
2. Gerasimov A.N. Parametric and non-parametric methods in medical statistics / Epidemiology and vaccinal prevention. 2015; 5(84): 6-12.
3. Gerasimov A.N. Medical statistics. M. MIA. 2007. 480 p.
4. Reed M., Simon B. Methods of modern mathematical physics. Vol. 1. Functional analysis. M. The World. 1977. 355 S.
5. Belov Yu.V., Gerasimov A.N., Komarov R.N. etc. ARCH EuroSCORE model: Prediction of perioperative risk in surgery of the aortic arch with the protection of the brain (GM) // Cardiology and cardiovascular surgery. 2015; No. 4, volume 8: 17-25.
Review
Views:
315
Email this article 