Question No 40-
Please analyse Yield and Cost versus input variables. Worksheet: [Link]
Answer:
when we carryout Analyse factorial design with Yield in the response, we get the following output.
Y- Yield
Coded Coefficients
Term Effect Coef SE Coef T-Value P-Value VIF
Constant 45.5592 0.0955 477.25 0.000
Blocks
1 -0.0484 0.0955 -0.51 0.628 1.00
Time2 2.9594 1.4797 0.0955 15.50 0.000 1.00
Temp2 2.7632 1.3816 0.0955 14.47 0.000 1.00
Catalyst2 0.1618 0.0809 0.0955 0.85 0.425 1.00
Time2*Temp2 0.8624 0.4312 0.0955 4.52 0.003 1.00
Time2*Catalyst2 0.0744 0.0372 0.0955 0.39 0.708 1.00
Temp2*Catalyst2 -0.0867 -0.0434 0.0955 -0.45 0.663 1.00
Time2*Temp2*Catalyst2 0.0230 0.0115 0.0955 0.12 0.907 1.00
P-Value : <0.05, p low, null go..
We have evidence to reject the null hypothesis.
So there is the significant difference for these terms. (Highlighted ones - A, B & AB)
The insignificant terms are Blocks, Catalyst, AC, BC & ABC (P valve >0.05). So we will remove these
terms from the analysis model.
Model Summary
S R-sq R-sq(adj) R-sq(pred)
0.318478 98.26% 97.82% 96.90%
R-sq - 98%, 98% of the variation in yield can be explained by this model. The model
is the good fit. Only significant terms are included in the model.
Coded Coefficients
Term Effect Coef SE Coef T-Value P-Value VIF
Constant 45.5592 0.0796 572.21 0.000
Time2 2.9594 1.4797 0.0796 18.58 0.000 1.00
Temp2 2.7632 1.3816 0.0796 17.35 0.000 1.00
Time2*Temp2 0.8624 0.4312 0.0796 5.42 0.000 1.00
Regression Equation in Uncoded Units
Yield2 = 39.48 - 0.1026 Time2 + 0.01502 Temp2 + 0.001150 Time2*Temp2
Y- Cost
Coded Coefficients
Term Effect Coef SE Coef T-Value P-Value VIF
Constant 31.8928 0.0909 351.03 0.000
Blocks
1 -0.0914 0.0909 -1.01 0.348 1.00
Time2 1.7815 0.8907 0.0909 9.80 0.000 1.00
Temp2 3.7339 1.8670 0.0909 20.55 0.000 1.00
Catalyst2 3.8813 1.9407 0.0909 21.36 0.000 1.00
Time2*Temp2 -0.0963 -0.0481 0.0909 -0.53 0.613 1.00
Time2*Catalyst2 0.5740 0.2870 0.0909 3.16 0.016 1.00
Temp2*Catalyst2 0.6943 0.3471 0.0909 3.82 0.007 1.00
Time2*Temp2*Catalyst2 -0.0342 -0.0171 0.0909 -0.19 0.856 1.00
P-Value : <0.05, p low, null go..
We have evidence to reject the null hypothesis.
So there is the significant difference for these terms. (Highlighted ones - A, B, C, BC & AC)
The insignificant terms are Blocks, AB & ABC (P valve >0.05). So we will remove these terms from the
analysis model.
Model Summary
S R-sq R-sq(adj) R-sq(pred)
0.331640 99.17% 98.76% 97.88%
R-sq - 99%, 99% of the variation in Cost can be explained by this model. The model
is the good fit. Only significant terms are included in the model.
Coded Coefficients
Term Effect Coef SE Coef T-Value P-Value VIF
Constant 31.8928 0.0829 384.67 0.000
Time2 1.7815 0.8907 0.0829 10.74 0.000 1.00
Temp2 3.7339 1.8670 0.0829 22.52 0.000 1.00
Catalyst2 3.8813 1.9407 0.0829 23.41 0.000 1.00
Time2*Catalyst2 0.5740 0.2870 0.0829 3.46 0.006 1.00
Temp2*Catalyst2 0.6943 0.3471 0.0829 4.19 0.002 1.00
Regression Equation in Uncoded Units
Cost2 = 16.746 + 0.05938 Time2 + 0.07468 Temp2 - 1.159 Catalyst2
+ 0.01913 Time2*Catalyst2 + 0.01389 Temp2*Catalyst2
Y - Yield
Significant Terms - A,B,AB
Y-Cost
Significant terms - A,B,C,AC & BC
For response optimization, we need to include all the terms significant to either Y measures.