BIOL 373 Lecture Notes - Lecture 19: Regression Analysis, Biostatistics, Covariance
Document Summary
Why do we see deviations from the mean or regression line: other factors may be involved, multiple regression allows you to measure additional factors & prediction variables. A non-zero relationship, each factor/variable adds a new dimension. Allows to ask questions about multiple predictors y=b0+b1 x1+b2 x2+ +bmxm+ : ex: y = weight, x1 = height, x2 = weight. Multiple regression/correlation: find the combination that minimizes least squares (error, increase number of parameters estimated, increase variation explained, effect on degrees of freedom. Will not be asked on the exam: option 1: H0: 1 = 2 = 3 = = m = 0. Ha: at least on slope < > 0: option 2: : population parameter for the amount of variation explained. Identifying explanatory variables: compare calculated f with critical f (1), m, n m 1, partition variation & compare to a null distribution (random variation) Reg df residual mswith term: two ways of testing terms: