Ecological Systems and Devices Annotation << Back Approaches to statistical analysis of biological and medical data with confounders A.N. Varaksin, V.G. Panov, E.D. Konstantinova Analysis of approaches to assessing effects of factors of interest performed under the presence of concomitants (confounders). The outcome variable considered to be a continuous, while risk factor and confounder are categorized. The raw data thus may be processed by ANOVA. In the example it is revealed differences in the assessing of influence of the risk factor on the outcome under various account for confounder procedures those are common in biomedical studies. There are considered following methods to adjust for confounder: ordinary weighed mean, marginal quantities, equal weights, and ANCOVA. It is shown that most relevant are the marginal quantities and equal weights methods. ANCOVA may be appropriate to estimate the effect under the necessary pre-assumptions hold. Key words: ANOVA; biomedical data; confounder (covariate); bias; adjustment for confounder. Contacts: E-mail: varaksin@ecko.uran.ru Pp. 37-42.