Использование метода непараметрической статистики в экспериментальной кардиологии

С. М. Чибисов, К. В. Шебзухов, А. С. Чибисов et al. · Вестник Российского университета дружбы народов. Серия: Медицина · 2002

The analysis of the experimental data submitted in the scientific literature in range of medicine and biology and, showed, that of 50 % 80 % of outcomes of biomedical researches do not submit to normal allocation. Most typically, it for bunches of the patients of the patients or experimental samples of researched plants. Traditional use in overwhelming majority of works of criteria and methods demanding normality of allocation, such as St'judenf s criterion and Pirsonvs coefficients of linear correlation, is incorrect and can result in to false outcomes. It is necessary to notice, that the spread practice of a rejection of extreme values distinguished from average more than on 3 у, does not compensate abnormality of allocation connected to an asymmetry or a bimodality. It often reduces in errors in a choice of statistical methods.

Cited by 0 works.

Authors