EEB225H1 Lecture Notes - Lecture 16: General Linear Model, Dependent And Independent Variables, Survival Analysis
Document Summary
Regression approach extended: to include multiple x variables. Incorporates categorical x variables as well as numerical x variables. Y = response: the value of y when x is 0. Where is the slope: rise over run, total variability in the response. + x = regression line: variability explained by the model. A 1-way anova is also a linear model: y = + a + . Where a is the group effect: unexplained variability in the error. A 2-way factorial anova: y = + a + b + a*b + : where is the error (unexplained variation, where is the error (unexplained variation) An analysis of covariance (ancova): y = + x + a + . Factorial experiment: where is the intercept or the grand mean, where is the slope estimate associated with the numerical variable x, where a is the group 1 effect. All combinations of the values of two or more categorical explanatory variables are investigated.