The EMMA dataset program was previously used to build a simple regression model that predicts "Total customer value" (Y) from "Tenure" (X). . .

1.If using SAS, connect to the VCL and run the SAS code program, EMMA to output a simple regression analysis. Scroll through the output. What is the simple regression equation? (This should match your answer from last week)

If not using SAS, go directly to step 2 . . .

2.Revise your regression model to predict customer value from a simple regression model to a multiple regression model by using TENURE, AGE, ORDERS as your predictors (If using Excel, these columns will need to become adjacent to each other)

3.Note that one of the three predictors is not a statistically significant predictor (How do you determine which one?).

4.Adjust the model* so that it includes only the two variables that are statistically significant

5.What is the equation of your final multiple regression model with the two predictors?

6.What is the predicted value for a 25 year old, that has placed 2 orders?

7.What is the goodness of fit measure for your model? Is it a good model? (explanation does not require SAS coding)

*Helpful hint if using SAS:

/* The SAS code for a multiple regression model to predict Y from several variables has the following form */

proc reg data=edc.emma;

model Y = variable1 variable2 variable3 variable4 variable5 ;