33:630:385 Lecture Notes - Lecture 7: Collinearity, Statistical Hypothesis Testing
18 Apr 2016
School
Department
Course
Professor
Document Summary
Sales a = 500 - 10 pricea + 5 priceb use numeric variables to predict sales. Dummy coding dummy variables est. indicate the impact of the category on dependent variable. 1 = presence of characterisic if have more than 2 categories (n), extend dummy coding to n 1 variables. Yi = a + b1 x1 + b2 x2. Ordinal predictors if have n categories, make n-1 dummy variables use either lowest/highest category as category w/o dummy variable. Interacion efect to examine how some independent variable moderates the relaionship between the dependent variable and another independent variable. Yi = a + b1 x1 + b2 x2 + b3 x1 x2. **if this is the case, interacion efect is not applicable to model. Mostly in retail data (example: sales, price, ad exp) use if a variable is numeric and does not have linear relaionship with y. Ln(yi) = a + b1 ln(x1) + b2 ln(x2) a = %, b = %
