Road construction contracts in the state of Florida are awarded on the basis of competitive, sealed bids; the contractor who submits the lowest bid price wins the contract. During the 1980s, the Office of the Florida Attorney General (FLAG) suspected numerous con- i tractors of practicing bid collusion, i.e., setting the winning bid price above the fair, or competitive, price in order to increase profit margin. FLAG collected data for 279 road construction contracts; the data are available in the file FLAG.DAT. For each contract, the following variables were measured:

1. Price of contract ($thousands) bid by lowest bidder

2. Department of Transportation (DOT) engineer's estimate of fair contract price ($thousands)

3. Ratio of low (winning) bid price to DOT engineer's estimate of fair price.

4. Status (fixed or competitive) of contract

5. District (1,2,3,4, or 5) in which construction project is located

6. Number of bidders on contract

7. Estimated number of days to complete work

8. Length of road project (miles)

9. Percentage of costs allocated to liquid asphalt

10. Percentage of costs allocated to base material

11. Percentage of costs allocated to excavation

12. Percentage of costs allocated to mobilization

13. Percentage of costs allocated to structures

14. Percentage of costs allocated to traffic control

15. Subcontractor utilization (yes or no)

FLAG wants to model the price (y) of the contract bid by lowest bidder in hopes of preventing price-fixing in the future.

a. Do you detect any multicollinearity in these variables? If so, do you recommend that all of these variables be used to predict low bid pricey? If not. which variables do you recommend?

b. Using the variables selected in part a, fit a first order model for E(y) to the data stored in the file.

c. Conduct a complete residual analysis on the model fit in part b. Do you detect any outliers? Are the standard regression assumptions reasonably satisfied?