Rock n Roll Heaven is an outdoor pavilion that presents musical performers throughout a 6-month season, from late spring to early fall. Rock n Roll presents a diverse venue of artists in a set of approximately 40 events each season. In order to better project its costs and expected attendance, Rock n Roll uses regression analysis to project expected ticketsales for upcoming events for each performer. The regression results shown below are derived from the three most recent seasons. The dependent variable for Rock n Roll is the number of paying tickets holders for each event, and the independent variables are:

1. whether or not this particular performer appeared at Rock n Roll previously (a dummy variable, 0 if no and 1 if yes)

2. the spending on advertising targeted to the performers appearance

3. the performers local sales of CDs in the most recent year prior to their appearance

4. the number of television appearances for the performer in the most recent year

5. the number of public appearances in the U.S. by the performer in the recent year

Independent Variables Results

Regression intercept 1,224 Attendance at prior concert

Coefficient 3,445

t-value 4.11 Spending on advertising

Coefficient 0.113

t-value 1.88 Performer's CD sales

Coefficient 0.00044

t-value 1.22 Television appearances

Coefficient 898

t-value 2.4 Other Public performances

Coefficient 1,233

t-value 3.7 R-squared 0.88 Standard error of the estimate 2,447

Required:

1. Using the above regression, what attendance would be predicted for a performer who had appeared at Rock n Roll previously, had 6 other public appearances but no

TV appearances, CD sales of $10 million, and Rock n Roll planned to spend $35,000 on advertising? (Hint: y = a + v1x1+v2x2+v3x3+v4x4+v5x5)

2. Evaluate the precision and reliability of the regression results shown above. What changes if any, do you propose for the regression? Which variables should be deleted, and which do you think should be added, and why?