In: Statistics and Probability
An analyst wanted to predict the time required to complete a construction project using four variables - size of the contract ?1x1 (in $1000 units), number of workdays adversely affected by the weather ?2x2, number of subcontractors involved in the project ?4x4, and a variable ?3x3 that measured the presence ( ?3=1x3=1) or absence ( ?3=0x3=0) of a workers' strike during the construction. Fifteen construction projects were randomly chosen, and each of the four variables as well as the time to completion were measured. The full model shown below uses ?1x1, ?2x2, ?3x3, ?4x4, ?1?3x1x3, ?2?3x2x3, and ?4?3x4x3.
Estimate | Std. Error | t value | |
(Intercept) | 13.692 | 8.674 | 1.579 |
x1 | 0.028 | 0.016 | 1.708 |
x2 | -1.046 | 0.838 | -1.248 |
x3 | 66.781 | 86.623 | 0.771 |
x4 | 2.068 | 1.624 | 1.274 |
x1:x3 | -0.035 | 0.018 | -1.942 |
x2:x3 | 5.105 | 1.567 | 3.257 |
x4:x3 | -7.297 | 9.122 | -0.8 |
a) Find the estimaing equation for the presence of workers on
strike.
?̂ y^ = - ?1x1 + ?2x2
- ?4x4
b) Find the estimating equation for the absence of workers on
strike.
?̂ y^ = + ?1x1 - ?2x2
+ ?4x4
c) Interpret ?2(???)=0.9102R2(adj)=0.9102
% of the variation in time to completion contract
size workdays affected number subcontractors workers strike
status can be explained by time,
contract size, workdays affected, and number subcontractors
contract size, workdays affected, number subcontractors, and worker
strike status time, workdays affected, number subcontractors, and
worker strike status time, contract size, number subcontractors,
and worker strike status .
The estimated regression equation is,
y =13.692 + 0.028 x1 - 1.046 x2 + 66.781 x3 + 2.068 x4 - 0.035 x1:x3 + 5.105 x2:x3 - 7.297 x4:x3
a.
Estimated equation for the presence of workers on strike (x3 = 1)
y =13.692 + 0.028 x1 - 1.046 x2 + 66.781 x3 + 2.068 x4 - 0.035 x1:x3 + 5.105 x2:x3 - 7.297 x4:x3
y =13.692 + 0.028 x1 - 1.046 x2 + 66.781 (1) + 2.068 x4 - 0.035 x1:1 + 5.105 x2:1 - 7.297 x4:1
y = (13.692 + 66.781) + (0.028 - 0.035 ) x1 + (-1.046 + 5.105 ) x2 + (2.068 - 7.297 ) x4
y = 80.473 - 0.007 x1 + 4.059 x2 - 5.229 x4
b.
Estimated equation for the absence of workers on strike (x3 = 0)
y =13.692 + 0.028 x1 - 1.046 x2 + 66.781 x3 + 2.068 x4 - 0.035 x1:x3 + 5.105 x2:x3 - 7.297 x4:x3
y =13.692 + 0.028 x1 - 1.046 x2 + 66.781 (0) + 2.068 x4 - 0.035 x1:0 + 5.105 x2:0 - 7.297 x4:0
y = 13.692 + 0.028 x1 - 1.046 x2 + 2.068 x4
c.
91.02% of the variation in time to completion of project is explained by variables size of the contract ?1, number of workdays adversely affected by the weather ?2, number of subcontractors involved in the project ?4, and a variable ?3 that measured the presence or absence of a workers' strike during the construction.