Ecological Systems and Devices Annotation << Back SIMPLE AND MULTIPLE LINEAR REGRESSION IN MEDICAL AND ECOLOGICAL INVESTIGATIONS: APPROACHES BASED ON PRIMARY DATA AND ON DATA OF MOVING AVERAGE A.N. Varaksin, T.A. Maslakova, Yu.V. Shalaumova The applicability conditions for simple and multiple linear regressions are discussed. We propose an approach based on the transition from primary data to a moving average as one of the methods to test the linearity of the relationship between predictor X and outcome Y variables. The features of building a moving average for simple and multiple regression models are indicated. Differences in methods for testing the assumption of linearity in simple and multiple regression models are formulated. Specific examples illustrate the advantages of a moving average approach in comparison with data stratification. The importance of visual graphical representation of regression analysis results is emphasized. We paid special attention to the analysis of diagrams representing the relationship between observed and predicted values of the outcome in multiple regression models. These diagrams show the explanatory degree of multiple linear regression models graphically and provide additional information to the multiple correlation coefficient. Keywords. Linear regression; primary data; moving average data; linearity condition. DOI: 10.25791/esip.1.2023.1346 Pp. 03-16.