The regression below uses a cars length (in inches) and weight (in pounds) to estimate its engine size

or displacement (in cubic inches).

. regress displacement length weight

Source | SS df MS Number of obs = 74

-------------+---------------------------------- F(2, 71) = ______

Model | 493832.977 2 246916.488 Prob > F = 0.0000

Residual | 121854.483 ___ 1716.26032 R-squared = ______

-------------+---------------------------------- Adj R-squared = ______

Total | __________ ___ 8434.07479 Root MSE = 41.428

------------------------------------------------------------------------------

displacement | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

length | -.4490482 .6718105 ____ 0.506 -1.7886 .890504

weight | .117916 .0192472 6.13 0.000 .0795383 .1562937

_cons | -74.35457 ________ -1.006 0.318 _________ ________

------------------------------------------------------------------------------

3 Missing Values

Fill in the remaining values that are missing from the regression output.

4 Modified Specifications

If the variable trunk space were added to the model above,

a) Would you expect R2 to increase or decrease? Why?

b) Would you expect R2 to increase or decrease? Why?

c) If length were removed from the model, would that cause the coefficient on weight to be biased in the

regression displacement \ =

0 +

1weight? Why or why not?

d) Another factor that affects how large of an engine a car needs is its gear ratio. Basically, lower gear

ratios mean that a car needs a smaller engine to propel the car. Heavier cars tend to have smaller gear

ratios. Since gear ratio is omitted from the model above, do you expect that the estimated coefficient on

weight is too large, too small, or just about right? Explain your reasoning.

e) What more information would you need in order to actually calculate the size of the OVB?