Multiple Linear Regression Discussion Questions
Description
Questions:
Although natural gas is currently inexpensive and nuclear power currently (and perhaps deservedly) does not have a good reputation, it is possible that more nuclear power plants will be constructed in the future. Table 4 presents data concerning the construction costs of light water reactor (LWR) nuclear power plants. The dependent variable, C, construction cost, is expressed in millions of dollars, adjusted to a 1976 base. Preliminary analysis of the data and economic theory indicate that variation in cost increases as cost increases. This suggests transforming cost by taking its natural logarithm.
S Power plant capacity in MWe
N Cumulative number of power plants built by the contractor
Build a multiple regression model to predict ln(C) by taking S and N or their natural logarithms as the independent variables. Make sure to check for multicollinearity.
Use residual analysis and R2 to check your model.
- State which variables are important in predicting the cost of constructing an LWR plant, and
- State a prediction equation that can be used to predict ln(C).
- Table 4
- Data Concerning Construction of Light Water Reactors
Plant
C
S
N
1
460.05
687
14
2
452.99
1,065
1
3
443.22
1,065
1
4
652.32
1,065
12
5
642.23
1,065
12
6
345.39
514
3
7
272.37
822
5
8
317.21
457
1
9
457.12
822
5
10
690.19
792
2
11
350.63
560
3
12
402.59
790
6
13
412.18
530
2
14
495.58
1,050
7
15
394.36
850
16
16
423.32
778
3
17
712.27
845
17
18
289.66
530
2
19
881.24
1,090
1
20
490.88
1,050
8
21
567.79
913
15
22
665.99
828
20
23
621.45
786
18
24
608.8
821
3
25
473.64
538
19
26
697.14
1,130
21
27
207.51
745
8
28
288.48
821
7
29
284.88
886
11
30
280.36
886
11
31
217.38
745
8
32
270.71
886
11
