Newfood Example Applied Regression Analysis Applied Regression Analysis 应用回归分析课件（英文版）.pdfVIP

Putting It All Together- “Newfood” A new instant breakfast product was the subject of a 6 month market test study. The goal of this test was to determine an optimal introductory marketing plan for use when the product is introduced on a national scale and to obtain improved sales forecasts. This dataset is the outcome of a controlled experiment in which the values of the independent variables which affect sales are chosen by the analyst. Most of the datasets we have examined so far are passively collected data. Newfood Example. Slide 1 Design of the Experiment A controlled store test was designed involving stores in four different cities. Variables to be considered: 1. Price 2. Advertising 3. Shelf-location “Design” of the experiment consists of determining the levels of each of these three variables as well as the possible combinations of these levels. Newfood Example. Slide 2 Design of the Experiment- Levels of the Variables price levels: 24, 29, and 34 cents per oz (why three?) Advertising levels: high and low (why two?) All advertising was in the form of television commercials. The high level was used in two of the four cities and was meant to simulate the effect of a $6 million nationwide advertising campaign. The low level was used in the other two cities and was designed to simulate the effect of a $3 million nationwide advertising campaign. Two shelf locations: bread or instant brkfast sections Newfood Example. Slide 3 Design of the Experiment- Design