MKTG 3596 Lecture 7: Topic 7- Regression
Document Summary
Alternative to correlation analysis for resting bivariate relationships. Statistical approach to modeling the relationship between two variables: one variable is the predictor (i. e. explanatory variable, independent variable, the other variable is the criterion (i. e. outcome variable, dependent variable) Y = the actual value of the dependent. ( y hat ) = the predicted value of the dependent that does not account for error. Least squared error principle: goal is to have as little error as possible across the range of scores, error (residual)= difference between points and regression line. = variance in y that is not explained by x. = y : want the smallest sum of squared errors. Multiple correlation association between a dependent variable and two or more predictor variables. Multiple regression statistical procedure for predicting scores on a dependent variable from two or more predictor variables. Used to test the relationship between a large number of independent variables and a dependent variable.
