Solutions from Montgomery, D. C.

(2012) Design and Analysis of Experiments, Wiley, NY

Chapter 6
k
The 2 Factorial Design
Solutions

6.1. An engineer is interested in the effects of cutting speed (A), tool geometry (B), and cutting angle on
the life (in hours) of a machine tool. Two levels of each factor are chosen, and three replicates of a 23
factorial design are run. The results are as follows:

Treatment Replicate
A B C Combination I II III
- - - (1) 22 31 25
+ - - a 32 43 29
- + - b 35 34 50
+ + - ab 55 47 46
- - + c 44 45 38
+ - + ac 40 37 36
- + + bc 60 50 54
+ + + abc 39 41 47

(a) Estimate the factor effects. Which effects appear to be large?

From the normal probability plot of effects below, factors B, C, and the AC interaction appear to be
significant.

DESIGN-EXPERT Plot
Life
N o rm a l p lo t
A: Cutting Speed
B: Tool Geometry 99
N o r m a l % p r o b a b i l i ty

C: Cutting Angle
B
95
90 C
80
70

50 A

30
20
10
5

AC
1

-8 .8 3 -3 .7 9 1 .2 5 6 .2 9 1 1 .3 3

E ffe c t

(b) Use the analysis of variance to confirm your conclusions for part (a).

The analysis of variance confirms the significance of factors B, C, and the AC interaction.

6-1
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output

Response: Life in hours
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 1612.67 7 230.38 7.64 0.0004 significant
A 0.67 1 0.67 0.022 0.8837
B 770.67 1 770.67 25.55 0.0001
C 280.17 1 280.17 9.29 0.0077
AB 16.67 1 16.67 0.55 0.4681
AC 468.17 1 468.17 15.52 0.0012
BC 48.17 1 48.17 1.60 0.2245
ABC 28.17 1 28.17 0.93 0.3483
Pure Error 482.67 16 30.17
Cor Total 2095.33 23

The Model F-value of 7.64 implies the model is significant. There is only
a 0.04% chance that a "Model F-Value" this large could occur due to noise.

The reduced model ANOVA is shown below. Factor A was included to maintain hierarchy.

Design Expert Output

Response: Life in hours
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 1519.67 4 379.92 12.54 < 0.0001 significant
A 0.67 1 0.67 0.022 0.8836
B 770.67 1 770.67 25.44 < 0.0001
C 280.17 1 280.17 9.25 0.0067
AC 468.17 1 468.17 15.45 0.0009
Residual 575.67 19 30.30
Lack of Fit 93.00 3 31.00 1.03 0.4067 not significant
Pure Error 482.67 16 30.17
Cor Total 2095.33 23

The Model F-value of 12.54 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Effects B, C and AC are significant at 1%.

(c) Write down a regression model for predicting tool life (in hours) based on the results of this
experiment.

yijk = 40.8333 + 0.1667 x A + 5.6667 xB + 3.4167 xC + 4.4167 x A xC

Design Expert Output

Coefficient Standard 95% CI 95% CI
Factor Estimate DF Error Low High VIF
Intercept 40.83 1 1.12 38.48 43.19
A-Cutting Speed 0.17 1 1.12 -2.19 2.52 1.00
B-Tool Geometry 5.67 1 1.12 3.31 8.02 1.00
C-Cutting Angle 3.42 1 1.12 1.06 5.77 1.00
AC -4.42 1 1.12 -6.77 -2.06 1.00
Final Equation in Terms of Coded Factors:

Life =
+40.83
+0.17 *A
+5.67 *B
+3.42 *C
-4.42 *A*C

Final Equation in Terms of Actual Factors:

6-2
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Life =
+40.83333
+0.16667 * Cutting Speed
+5.66667 * Tool Geometry
+3.41667 * Cutting Angle
-4.41667 * Cutting Speed * Cutting Angle

The equation in part (c) and in the given in the computer output form a “hierarchial” model, that is, if an
interaction is included in the model, then all of the main effects referenced in the interaction are also
included in the model.

(d) Analyze the residuals. Are there any obvious problems?

Normal plot of residuals Residuals vs. Predicted

11.5

99

95
6.79167
Norm al % probability

90
80

Res iduals
70

50 2.08333

30
20
10
-2.625
5

-7.33333

-7.33333 -2.625 2.08333 6.79167 11.5 27.17 33.92 40.67 47.42 54.17

Res idual Predicted

There is nothing unusual about the residual plots.

(e) Based on the analysis of main effects and interaction plots, what levels of A, B, and C would you
recommend using?

Since B has a positive effect, set B at the high level to increase life. The AC interaction plot reveals that life
would be maximized with C at the high level and A at the low level.

6-3
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

Inte ra c tio n G ra p h O ne F a c to r P lo t
Life C u ttin g A n g le Life
60 60

X = A: Cutting Speed X = B: Tool Geometry

Y = C: Cutting Angle
Actual Factors
5 0 .5 5 0 .5
C- -1.000 A: Cutting Speed = 0.00
C+ 1.000 C: Cutting Angle = 0.00
Actual Factor
B: Tool Geometry = 0.00 L i fe

L i fe
41 41

3 1 .5 3 1 .5

22 22

-1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0 -1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0

C u ttin g S p e e d T o o l G e o m e try

6.2. Reconsider part (c) of Problem 6.1. Use the regression model to generate response surface and
contour plots of the tool life response. Interpret these plots. Do they provide insight regarding the
desirable operating conditions for this process?

The response surface plot and the contour plot in terms of factors A and C with B at the high level are
shown below. They show the curvature due to the AC interaction. These plots make it easy to see the
region of greatest tool life.

1.00 3
Life 3
52 46

50

0.50 56
48 54
52
50
C: Cutting Angle

48
46
44
0.00 42
40
Life

46 38

-0.50 44

42 1.00
1.00
40 0.50
0.50
-1.00 3 3
0.00
-1.00 -0.50 0.00 0.50 1.00 0.00
C: Cutting Angle -0.50 -0.50
A: Cutting Speed
A: Cutting Speed -1.00 -1.00

6-4
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.3. Find the standard error of the factor effects and approximate 95 percent confidence limits for the
factor effects in Problem 6.1. Do the results of this analysis agree with the conclusions from the analysis of
variance?

1 1
SE( effect ) = S2 = 30.17 = 2.24
n2 k −2
(3)2 3− 2
Variable Effect
A 0.333
B 11.333 *
AB -1.667
C 6.833 *
AC -8.833 *
BC -2.833
ABC -2.167

The 95% confidence intervals for factors B, C and AC do not contain zero. This agrees with the analysis of
variance approach.

6.4. Plot the factor effects from Problem 6.1 on a graph relative to an appropriately scaled t distribution.
Does this graphical display adequately identify the important factors? Compare the conclusions from this
MS E 30.17
plot with the results from the analysis of variance. S = = = 3.17
n 3

Scaled t Distribution

AC C B

-10.0 0.0 10.0

Factor Effects

This method identifies the same factors as the analysis of variance.

6.5. A router is used to cut locating notches on a printed circuit board. The vibration level at the surface
of the board as it is cut is considered to be a major source of dimensional variation in the notches. Two
factors are thought to influence vibration: bit size (A) and cutting speed (B). Two bit sizes (1/16 and 1/8
inch) and two speeds (40 and 90 rpm) are selected, and four boards are cut at each set of conditions shown
below. The response variable is vibration measured as a resultant vector of three accelerometers (x, y, and
z) on each test circuit board.

6-5
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Treatment Replicate
A B Combination I II III IV
- - (1) 18.2 18.9 12.9 14.4
+ - a 27.2 24.0 22.4 22.5
- + b 15.9 14.5 15.1 14.2
+ + ab 41.0 43.9 36.3 39.9

(a) Analyze the data from this experiment.

Design Expert Output

Response: Vibration
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 1638.11 3 546.04 91.36 < 0.0001 significant
A 1107.23 1 1107.23 185.25 < 0.0001
B 227.26 1 227.26 38.02 < 0.0001
AB 303.63 1 303.63 50.80 < 0.0001
Residual 71.72 12 5.98
Lack of Fit 0.000 0
Pure Error 71.72 12 5.98
Cor Total 1709.83 15

The Model F-value of 91.36 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

(b) Construct a normal probability plot of the residuals, and plot the residuals versus the predicted
vibration level. Interpret these plots.

Normal plot of residuals Residuals vs. Predicted

3.625

99

95
1.725
Norm al % probability

90
80
Res iduals

70

50 -0.175

30
20
10
-2.075
5

-3.975

-3.975 -2.075 -0.175 1.725 3.625 14.92 21.26 27.60 33.94 40.27

Res idual Predicted

There is nothing unusual about the residual plots.

(c) Draw the AB interaction plot. Interpret this plot. What levels of bit size and speed would you
recommend for routine operation?

To reduce the vibration, use the smaller bit. Once the small bit is specified, either speed will work equally
well, because the slope of the curve relating vibration to speed for the small tip is approximately zero. The
process is robust to speed changes if the small bit is used.

6-6
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot
Inte ra c tio n G ra p h
Vibration C u ttin g S p e e d
4 3 .9

X = A: Bit Size
Y = B: Cutting Speed
3 6 .1 5
Design Points

V i b r a ti o n
B- -1.000
B+ 1.000
2 8 .4

2 0 .6 5

1 2 .9

-1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0

B it S ize

6.6. Reconsider the experiment described in Problem 6.1. Suppose that the experimenter only performed
the eight trials from replicate I. In addition, he ran four center points and obtained the following response
values: 36, 40, 43, 45.

(a) Estimate the factor effects. Which effects are large?

DESIGN-EXPERT Plot
Life
N o rm a l p lo t
A: Cutting Speed
B: Tool Geometry 99
N o r m a l % p r o b a b i l i ty

C: Cutting Angle
95 B
90
80 C
70

50

30
20
10
5 AC

-1 3 .7 5 -7 .1 3 -0 .5 0 6 .1 2 1 2 .7 5

E ffe c t

Effects B, C, and AC appear to be large.

(b) Perform an analysis of variance, including a check for pure quadratic curvature. What are your
n F n C (y F − y C )2 (8)(4)(40.875 − 41.000)2
conclusions? SS PureQuadratic = = = 0.0417
n F + nC 8+ 4

6-7
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output

Response: Life in hours
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 1048.88 7 149.84 9.77 0.0439 significant
A 3.13 1 3.13 0.20 0.6823
B 325.13 1 325.13 21.20 0.0193
C 190.12 1 190.12 12.40 0.0389
AB 6.13 1 6.13 0.40 0.5722
AC 378.12 1 378.12 24.66 0.0157
BC 55.12 1 55.12 3.60 0.1542
ABC 91.12 1 91.12 5.94 0.0927
Curvature 0.042 1 0.042 2.717E-003 0.9617 not significant
Pure Error 46.00 3 15.33
Cor Total 1094.92 11

The Model F-value of 9.77 implies the model is significant. There is only
a 4.39% chance that a "Model F-Value" this large could occur due to noise.

The "Curvature F-value" of 0.00 implies the curvature (as measured by difference between the
average of the center points and the average of the factorial points) in the design space is not
significant relative to the noise. There is a 96.17% chance that a "Curvature F-value"
this large could occur due to noise.

Design Expert Output

Response: Life in hours
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 896.50 4 224.13 7.91 0.0098 significant
A 3.13 1 3.13 0.11 0.7496
B 325.12 1 325.12 11.47 0.0117
C 190.12 1 190.12 6.71 0.0360
AC 378.12 1 378.12 13.34 0.0082
Residual 198.42 7 28.35
Lack of Fit 152.42 4 38.10 2.49 0.2402 not significant
Pure Error 46.00 3 15.33
Cor Total 1094.92 11

The Model F-value of 7.91 implies the model is significant. There is only
a 0.98% chance that a "Model F-Value" this large could occur due to noise.

Effects B, C and AC are significant at 5%. There is no effect of curvature.

(c) Write down an appropriate model for predicting tool life, based on the results of this experiment. Does
this model differ in any substantial way from the model in Problem 6.1, part (c)?

The model shown in the Design Expert output below does not differ substantially from the model in
Problem 6.1, part (c).

Design Expert Output

Final Equation in Terms of Coded Factors:

Life =
+40.88
+0.62 *A
+6.37 *B
+4.87 *C
-6.88 *A*C

(d) Analyze the residuals.

6-8
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal Plot of Residuals Residuals vs. Predicted

6.83333

99

95
3.39583
90
Normal % Probability

80
70

Residuals
50 -0.0416667

30
20

10
-3.47917
5

-6.91667

-6.91667 -3.47917 -0.0416667 3.39583 6.83333 22.17 31.23 40.29 49.35 58.42

Residual Predicted

(e) What conclusions would you draw about the appropriate operating conditions for this process?

To maximize life run with B at the high level, A at the low level and C at the high level

Cube Graph
Life

58.38 45.88

B+ 34.88 49.88
B: Tool Geom etry

45.63 33.13 C+

C: Cutting Angle

B- 22.13 37.13 C-
A- A+
A: Cutting Speed

6.7. An experiment was performed to improve the yield of a chemical process. Four factors were
selected, and two replicates of a completely randomized experiment were run. The results are shown in the
following table:

Treatment Replicate Replicate Treatment Replicate Replicate

Combination I II Combination I II
(1) 90 93 d 98 95
a 74 78 ad 72 76
b 81 85 bd 87 83

6-9
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

ab 83 80 abd 85 86
c 77 78 cd 99 90
ac 81 80 acd 79 75
bc 88 82 bcd 87 84
abc 73 70 abcd 80 80

(a) Estimate the factor effects.

Design Expert Output

Term Effect SumSqr % Contribtn
Model Intercept
Error A -9.0625 657.031 40.3714
Error B -1.3125 13.7812 0.84679
Error C -2.6875 57.7813 3.55038
Error D 3.9375 124.031 7.62111
Error AB 4.0625 132.031 8.11267
Error AC 0.6875 3.78125 0.232339
Error AD -2.1875 38.2813 2.3522
Error BC -0.5625 2.53125 0.155533
Error BD -0.1875 0.28125 0.0172814
Error CD 1.6875 22.7812 1.3998
Error ABC -5.1875 215.281 13.228
Error ABD 4.6875 175.781 10.8009
Error ACD -0.9375 7.03125 0.432036
Error BCD -0.9375 7.03125 0.432036
Error ABCD 2.4375 47.5313 2.92056

(b) Prepare an analysis of variance table, and determine which factors are important in explaining yield.

Design Expert Output

Response: yield
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 1504.97 15 100.33 13.10 < 0.0001 significant
A 657.03 1 657.03 85.82 < 0.0001
B 13.78 1 13.78 1.80 0.1984
C 57.78 1 57.78 7.55 0.0143
D 124.03 1 124.03 16.20 0.0010
AB 132.03 1 132.03 17.24 0.0007
AC 3.78 1 3.78 0.49 0.4923
AD 38.28 1 38.28 5.00 0.0399
BC 2.53 1 2.53 0.33 0.5733
BD 0.28 1 0.28 0.037 0.8504
CD 22.78 1 22.78 2.98 0.1038
ABC 215.28 1 215.28 28.12 < 0.0001
ABD 175.78 1 175.78 22.96 0.0002
ACD 7.03 1 7.03 0.92 0.3522
BCD 7.03 1 7.03 0.92 0.3522
ABCD 47.53 1 47.53 6.21 0.0241
Residual 122.50 16 7.66
Lack of Fit 0.000 0
Pure Error 122.50 16 7.66
Cor Total 1627.47 31

The Model F-value of 13.10 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, D, AB, AD, ABC, ABD, ABCD are significant model terms.

F0.01,1,16 = 8.53 , and F0.025,1,16 = 612

. therefore, factors A and D and interactions AB, ABC, and ABD are
significant at 1%. Factor C and interactions AD and ABCD are significant at 5%.

6-10
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(b) Write down a regression model for predicting yield, assuming that all four factors were varied over the
range from -1 to +1 (in coded units).

Model with hierarchy maintained:

Design Expert Output

Final Equation in Terms of Coded Factors:

yield =
+82.78
-4.53 *A
-0.66 *B
-1.34 *C
+1.97 *D
+2.03 *A*B
+0.34 *A*C
-1.09 *A*D
-0.28 *B*C
-0.094 *B*D
+0.84 *C*D
-2.59 *A*B*C
+2.34 *A*B*D
-0.47 *A*C*D
-0.47 *B*C*D
+1.22 *A*B*C*D

Model without hierarchy terms:

Design Expert Output

Final Equation in Terms of Coded Factors:

yield =
+82.78
-4.53 *A
-1.34 *C
+1.97 *D
+2.03 *A*B
-1.09 *A*D
-2.59 *A*B*C
+2.34 *A*B*D
+1.22 *A*B*C*D

Confirmation runs might be run to see if the simpler model without hierarchy is satisfactory.

(d) Plot the residuals versus the predicted yield and on a normal probability scale. Does the residual
analysis appear satisfactory?

There appears to be one large residual both in the normal probability plot and in the plot of residuals versus
predicted.

6-11
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal plot of residuals Residuals vs. Predicted

6.96875

99

95
3.96875
Norm al % probability

90
80

Res iduals
70

50 0.96875

30
20 2
10
-2.03125
5

-5.03125

-5.03125 -2.03125 0.96875 3.96875 6.96875 71.91 78.30 84.69 91.08 97.47

Res idual Predicted

(e) Two three-factor interactions, ABC and ABD, apparently have large effects. Draw a cube plot in the
factors A, B, and C with the average yields shown at each corner. Repeat using the factors A, B, and D.
Do these two plots aid in data interpretation? Where would you recommend that the process be run
with respect to the four variables?

Cube Graph Cube Graph

yield yield

86.53 76.34 86.00 83.50

B+ 84.03 84.22 B+ 84.56 77.06

B: B

B: B

85.41 77.47 C+ 94.75 74.75 D+

C: C D: D

B- 93.28 74.97 C- B- 83.94 77.69 D-

A- A+ A- A+
A: A A: A

Run the process at A low B low, C low and D high.

6.8. A bacteriologist is interested in the effects of two different culture media and two different times on
the growth of a particular virus. She performs six replicates of a 22 design, making the runs in random
order. Analyze the bacterial growth data that follow and draw appropriate conclusions. Analyze the
residuals and comment on the model’s adequacy.

6-12
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Culture Medium
Time 1 M di 2
21 22 25 26
12 hr 23 28 24 25
20 26 29 27
37 39 31 34
18 hr 38 38 29 33
35 36 30 35

Design Expert Output

Response: Virus growth
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 691.46 3 230.49 45.12 < 0.0001 significant
A 9.38 1 9.38 1.84 0.1906
B 590.04 1 590.04 115.51 < 0.0001
AB 92.04 1 92.04 18.02 0.0004
Residual 102.17 20 5.11
Lack of Fit 0.000 0
Pure Error 102.17 20 5.11
Cor Total 793.63 23

The Model F-value of 45.12 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case B, AB are significant model terms.

Normal plot of residuals Residuals vs. Predicted

4.66667

99

95
2.66667
Norm al % probability

90
80
Res iduals

70
2
50 0.666667

30
20
10 2
-1.33333
5

-3.33333

-3.33333 -1.33333 0.666667 2.66667 4.66667 23.33 26.79 30.25 33.71 37.17

Res idual Predicted

Growth rate is affected by factor B (Time) and the AB interaction (Culture medium and Time). There is
some very slight indication of inequality of variance shown by the small decreasing funnel shape in the plot
of residuals versus predicted.

6-13
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot
Inte ra c tio n G ra p h
Virus growth T im e
39
2

X = A: Culture Medium
Y = B: Time
3 4 .2 5
Design Points

V i r u s g r o w th
B- 12.000
B+ 18.000
2 9 .5 2

2
2 4 .7 5

20

1 2

C u ltu re M e d iu m

6.9. An industrial engineer employed by a beverage bottler is interested in the effects of two different
types of 32-ounce bottles on the time to deliver 12-bottle cases of the product. The two bottle types are
glass and plastic. Two workers are used to perform a task consisting of moving 40 cases of the product 50
feet on a standard type of hand truck and stacking the cases in a display. Four replicates of a 22 factorial
design are performed, and the times observed are listed in the following table. Analyze the data and draw
the appropriate conclusions. Analyze the residuals and comment on the model’s adequacy.

Worker
Bottle Type 1 1 2 2
Glass 5.12 4.89 6.65 6.24
4.98 5.00 5.49 5.55

Plastic 4.95 4.43 5.28 4.91

4.27 4.25 4.75 4.71

Design Expert Output

Response:Times
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 4.86 3 1.62 13.04 0.0004 significant
A 2.02 1 2.02 16.28 0.0017
B 2.54 1 2.54 20.41 0.0007
AB 0.30 1 0.30 2.41 0.1463
Residual 1.49 12 0.12
Lack of Fit 0.000 0
Pure Error 1.49 12 0.12
Cor Total 6.35 15

The Model F-value of 13.04 implies the model is significant. There is only
a 0.04% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B are significant model terms.

There is some indication of non-constant variance in this experiment.

6-14
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal plot of residuals Residuals vs. Predicted

0.6675

99

95
0.3775
Norm al % probability

90
80

Res iduals
70

50 0.0875

30
20
10
-0.2025
5

-0.4925

-0.4925 -0.2025 0.0875 0.3775 0.6675 4.47 4.85 5.23 5.61 5.98

Res idual Predicted

Residuals vs. Worker Residuals vs. Bottle Type

3
0.6675

0.3775 1.5
Res iduals

Residuals

0.0875 0

-0.2025 -1.5

-0.4925 -3

1 2 1 2

Worker Bottle Type

6.10. In problem 6.9, the engineer was also interested in potential fatigue differences resulting from the
two types of bottles. As a measure of the amount of effort required, he measured the elevation of heart rate
(pulse) induced by the task. The results follow. Analyze the data and draw conclusions. Analyze the
residuals and comment on the model’s adequacy.

Worker
Bottle Type 1 1 2 2
Glass 39 45 20 13
58 35 16 11

Plastic 44 35 13 10
42 21 16 15

6-15
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output

Response: Pulse
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 2784.19 3 928.06 16.03 0.0002 significant
A 2626.56 1 2626.56 45.37 < 0.0001
B 105.06 1 105.06 1.81 0.2028
AB 52.56 1 52.56 0.91 0.3595
Residual 694.75 12 57.90
Lack of Fit 0.000 0
Pure Error 694.75 12 57.90
Cor Total 3478.94 15

The Model F-value of 16.03 implies the model is significant. There is only
a 0.02% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A are significant model terms.

Normal Plot of Residuals Residuals vs. Predicted

13.75

99

95
6.6875
90
Normal % Probability

80
70
Residuals

50 -0.375

30
20

10
-7.4375
5

-14.5

-14.5 -7.4375 -0.375 6.6875 13.75 13.50 21.19 28.88 36.56 44.25

Residual Predicted

Residuals vs. Worker Residuals vs. Bottle Type

13.75 13.75

6.6875 6.6875
Residuals

Residuals

-0.375 -0.375 2

-7.4375 -7.4375

-14.5 -14.5

1 2 1 2

Worker Bottle T ype

There is an indication that one worker exhibits greater variability than the other.

6-16
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.11. Calculate approximate 95 percent confidence limits for the factor effects in Problem 6.10. Do the
results of this analysis agree with the analysis of variance performed in Problem 6.10?

1 1
SE( effect ) = S2 = 57.90 = 3.80
n2 k −2
(4)2 2− 2
Variable Effect C.I.
A -25.625 ±3.80(1.96)= ±7.448
B -5.125 ±3.80(1.96)= ±7.448
AB 3.625 ±3.80(1.96)= ±7.448

The 95% confidence interval for factor A does not contain zero. This agrees with the analysis of variance
approach.

6.12. An article in the AT&T Technical Journal (March/April 1986, Vol. 65, pp. 39-50) describes the
application of two-level factorial designs to integrated circuit manufacturing. A basic processing step is to
grow an epitaxial layer on polished silicon wafers. The wafers mounted on a susceptor are positioned
inside a bell jar, and chemical vapors are introduced. The susceptor is rotated and heat is applied until the
epitaxial layer is thick enough. An experiment was run using two factors: arsenic flow rate (A) and
deposition time (B). Four replicates were run, and the epitaxial layer thickness was measured (in mm).
The data are shown below:

Replicate Factor Levels

A B I II III IV Low (-) High (+)
- - 14.037 16.165 13.972 13.907 A 55% 59%
+ - 13.880 13.860 14.032 13.914
- + 14.821 14.757 14.843 14.878 B Short Long
+ + 14.888 14.921 14.415 14.932 (10 min) (15 min)

(a) Estimate the factor effects.

Design Expert Output

Term Effect SumSqr % Contribtn
Model Intercept
Error A -0.31725 0.40259 6.79865
Error B 0.586 1.37358 23.1961
Error AB 0.2815 0.316969 5.35274
Error Lack Of Fit 0 0
Error Pure Error 3.82848 64.6525

(b) Conduct an analysis of variance. Which factors are important?

From the analysis of variance shown below, no factors appear to be important. Factor B is only marginally
interesting with an F-value of 4.31.

Design Expert Output

Response: Thickness
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 2.09 3 0.70 2.19 0.1425 not significant

6-17
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

A 0.40 1 0.40 1.26 0.2833

B 1.37 1 1.37 4.31 0.0602
AB 0.32 1 0.32 0.99 0.3386
Residual 3.83 12 0.32
Lack of Fit 0.000 0
Pure Error 3.83 12 0.32
Cor Total 5.92 15

The "Model F-value" of 2.19 implies the model is not significant relative to the noise. There is a
14.25 % chance that a "Model F-value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case there are no significant model terms.

(c) Write down a regression equation that could be used to predict epitaxial layer thickness over the region
of arsenic flow rate and deposition time used in this experiment.

Design Expert Output

Final Equation in Terms of Coded Factors:

Thickness =
+14.51
-0.16 *A
+0.29 *B
+0.14 *A*B

Final Equation in Terms of Actual Factors:

Thickness =
+37.62656
-0.43119 * Flow Rate
-1.48735 * Dep Time
+0.028150 * Flow Rate * Dep Time

(d) Analyze the residuals. Are there any residuals that should cause concern? Observation #2 falls outside
the groupings in the normal probability plot and the plot of residual versus predicted.

Normal plot of residuals Residuals vs. Predicted

1.64475

99

95
1.08025
Norm al % probability

90
80
Res iduals

70

50 0.51575

30
20
10
-0.04875
5

-0.61325

-0.61325 -0.04875 0.51575 1.08025 1.64475 13.92 14.15 14.37 14.60 14.82

Res idual Predicted

(e) Discuss how you might deal with the potential outlier found in part (d).

One approach would be to replace the observation with the average of the observations from that
experimental cell. Another approach would be to identify if there was a recording issue in the original data.

6-18
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

The first analysis below replaces the data point with the average of the other three. The second analysis
assumes that the reading was incorrectly recorded and should have been 14.165.

The analysis with the run associated with standard order 2 replaced with the average of the remaining three
runs in the cell, 13.972, is shown below.

Design Expert Output

Response: Thickness
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 2.97 3 0.99 53.57 < 0.0001 significant
A 7.439E-003 1 7.439E-003 0.40 0.5375
B 2.96 1 2.96 160.29 < 0.0001
AB 2.176E-004 1 2.176E-004 0.012 0.9153
Pure Error 0.22 12 0.018
Cor Total 3.19 15

The Model F-value of 53.57 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case B are significant model terms.

Final Equation in Terms of Coded Factors:

Thickness =
+14.38
-0.022 *A
+0.43 *B
+3.688E-003 *A*B

Final Equation in Terms of Actual Factors:

Thickness =
+13.36650
-0.020000 * Flow Rate
+0.12999 * Dep Time
+7.37500E-004 * Flow Rate * Dep Time

Normal plot of residuals Residuals vs. Predicted

0.143

99

95
0.01375
Norm al % probability

90 2

80
Res iduals

70

50 -0.1155

30
20
10
-0.24475
5

-0.374

-0.374 -0.24475 -0.1155 0.01375 0.143 13.92 14.15 14.37 14.60 14.82

Res idual Predicted

A new outlier is present and should be investigated.

6-19
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Analysis with the run associated with standard order 2 replaced with the value 14.165:

Design Expert Output

Response: Thickness
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 2.82 3 0.94 45.18 < 0.0001 significant
A 0.018 1 0.018 0.87 0.3693
B 2.80 1 2.80 134.47 < 0.0001
AB 3.969E-003 1 3.969E-003 0.19 0.6699
Pure Error 0.25 12 0.021
Cor Total 3.07 15

The Model F-value of 45.18 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case B are significant model terms.

Final Equation in Terms of Coded Factors:

Thickness =
+14.39
-0.034 *A
+0.42 *B
+0.016 *A*B

Final Equation in Terms of Actual Factors:

Thickness =
+15.50156
-0.056188 * Flow Rate
-0.012350 * Dep Time
+3.15000E-003 * Flow Rate * Dep Time

Normal plot of residuals Residuals vs. Predicted

3.00

99

95
Studentized Res iduals

1.50
Norm al % probability

90
80
70

50 0.00

30
20
10
-1.50
5

-3.00

-3.00 -1.96 -0.92 0.12 1.16 13.92 14.15 14.37 14.60 14.82

Studentized Res iduals Predicted

Another outlier is present and should be investigated.

6-20
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.13. Continuation of Problem 6.12. Use the regression model in part (c) of Problem 6.12 to generate a
response surface contour plot for epitaxial layer thickness. Suppose it is critically important to obtain layer
thickness of 14.5 mm. What settings of arsenic flow rate and deposition time would you recommend?

Arsenic flow rate may be set at any of the experimental levels, while the deposition time should be set at
12.4 minutes.

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

1 5 .0 0
4 Thic k ne s s 4 Inte ra c tio n G ra p h
Thickness Thickness D e p T im e
1 6 .1 6 5
X = A: Flow Rate
Y = B: Dep Time
X = A: Flow Rate
Design Points 1 4 .6 7 4 2 Y = B: Dep Time
1 3 .7 5
1 5 .4 9 5 3
Design Points

T h ic k n e s s
D e p T im e

B- 10.000
B+ 15.000
1 2 .5 0 1 4 .8 2 5 7
1 4 .5 2 3 7

1 4 .3 7 3 1
1 4 .1 5 6
1 1 .2 5
1 4 .2 2 2 6

1 4 .0 7 2
1 3 .4 8 6 4
4 4
1 0 .0 0
5 5 .0 0 5 6 .0 0 5 7 .0 0 5 8 .0 0 5 9 .0 0 5 5 .0 0 5 6 .0 0 5 7 .0 0 5 8 .0 0 5 9 .0 0

F lo w R a te F lo w R a te

6.14. Continuation of Problem 6.13. How would your answer to Problem 6.13 change if arsenic flow
rate was more difficult to control in the process than the deposition time?

Running the process at a high level of Deposition Time there is no change in thickness as flow rate
changes.

6.15. A nickel-titanium alloy is used to make components for jet turbine aircraft engines. Cracking is a
potentially serious problem in the final part, as it can lead to non-recoverable failure. A test is run at the
parts producer to determine the effects of four factors on cracks. The four factors are pouring temperature
(A), titanium content (B), heat treatment method (C), and the amount of grain refiner used (D). Two
replicated of a 24 design are run, and the length of crack (in µm) induced in a sample coupon subjected to a
standard test is measured. The data are shown below:

Treatment Replicate Replicate

A B C D Combination I II
- - - - (1) 7.037 6.376
+ - - - a 14.707 15.219
- + - - b 11.635 12.089
+ + - - ab 17.273 17.815
- - + - c 10.403 10.151
+ - + - ac 4.368 4.098
- + + - bc 9.360 9.253
+ + + - abc 13.440 12.923
- - - + d 8.561 8.951

6-21
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

+ - - + ad 16.867 17.052
- + - + bd 13.876 13.658
+ + - + abd 19.824 19.639
- - + + cd 11.846 12.337
+ - + + acd 6.125 5.904
- + + + bcd 11.190 10.935
+ + + + abcd 15.653 15.053

(a) Estimate the factor effects. Which factors appear to be large?

From the half normal plot of effects shown below, factors A, B, C, D, AB, AC, and ABC appear to be large.

Design Expert Output

Term Effect SumSqr % Contribtn
Model Intercept
Model A 3.01888 72.9089 12.7408
Model B 3.97588 126.461 22.099
Model C -3.59625 103.464 18.0804
Model D 1.95775 30.6623 5.35823
Model AB 1.93412 29.9267 5.22969
Model AC -4.00775 128.496 22.4548
Error AD 0.0765 0.046818 0.00818145
Error BC 0.096 0.073728 0.012884
Error BD 0.04725 0.0178605 0.00312112
Error CD -0.076875 0.0472781 0.00826185
Model ABC 3.1375 78.7512 13.7618
Error ABD 0.098 0.076832 0.0134264
Error ACD 0.019125 0.00292613 0.00051134
Error BCD 0.035625 0.0101531 0.00177426
Error ABCD 0.014125 0.00159613 0.000278923

DESIGN-EXPERT Plot Ha lf No r m a l p lo t
Crack Length

A: Pour Temp
99
B: Titanium Content
C: Heat Treat Method AC
D: Grain Ref iner 97

95 B
Half Normal % probability

90
C
ABC
85 A
80 D
AB
70

60 BC

40

20

0 .0 0 1 .0 0 2 .0 0 3 .0 1 4 .0 1

| E ffec t|

6-22
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(b) Conduct an analysis of variance. Do any of the factors affect cracking? Use α=0.05.

The Design Expert output below identifies factors A, B, C, D, AB, AC, and ABC as significant.

Design Expert Output

Response: Crack Lengthin mm x 10^-2
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 570.95 15 38.06 468.99 < 0.0001 significant
A 72.91 1 72.91 898.34 < 0.0001
B 126.46 1 126.46 1558.17 < 0.0001
C 103.46 1 103.46 1274.82 < 0.0001
D 30.66 1 30.66 377.80 < 0.0001
AB 29.93 1 29.93 368.74 < 0.0001
AC 128.50 1 128.50 1583.26 < 0.0001
AD 0.047 1 0.047 0.58 0.4586
BC 0.074 1 0.074 0.91 0.3547
BD 0.018 1 0.018 0.22 0.6453
CD 0.047 1 0.047 0.58 0.4564
ABC 78.75 1 78.75 970.33 < 0.0001
ABD 0.077 1 0.077 0.95 0.3450
ACD 2.926E-003 1 2.926E-003 0.036 0.8518
BCD 0.010 1 0.010 0.13 0.7282
ABCD 1.596E-003 1 1.596E-003 0.020 0.8902
Residual 1.30 16 0.081
Lack of Fit 0.000 0
Pure Error 1.30 16 0.081
Cor Total 572.25 31

The Model F-value of 468.99 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, C, D, AB, AC, ABC are significant model terms.

(c) Write down a regression model that can be used to predict crack length as a function of the significant
main effects and interactions you have identified in part (b).

Design Expert Output

Final Equation in Terms of Coded Factors:

Crack Length=
+11.99
+1.51 *A
+1.99 *B
-1.80 *C
+0.98 *D
+0.97 *A*B
-2.00 *A*C
+1.57 *A*B*C

(d) Analyze the residuals from this experiment.

6-23
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal plot of residuals Residuals vs. Predicted

0.454875

99

95
0.232688
Norm al % probability

90
80

Res iduals
70

50 0.0105

30
20
10
-0.211687
5

-0.433875

-0.433875 -0.211687 0.0105 0.232688 0.454875 4.19 8.06 11.93 15.80 19.66

Res idual Predicted

There is nothing unusual about the residuals.

(e) Is there an indication that any of the factors affect the variability in cracking?

By calculating the range of the two readings in each cell, we can also evaluate the effects of the factors on
variation. The following is the normal probability plot of effects:

DESIGN-EXPERT Plot
Range
N o rm a l p lo t
A: Pour Temp
B: Titanium Content 99
N o r m a l % p r o b a b i l i ty

C: Heat Treat Method

D: Grain Refiner CD
95
90 AB
80
70

50

30
20
10
5

-0 .1 0 -0 .0 2 0 .0 5 0 .1 3 0 .2 0

E ffe c t

It appears that the AB and CD interactions could be significant. The following is the ANOVA for the range
data:

6-24
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output

Response: Range
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 0.29 2 0.14 11.46 0.0014 significant
AB 0.13 1 0.13 9.98 0.0075
CD 0.16 1 0.16 12.94 0.0032
Residual 0.16 13 0.013
Cor Total 0.45 15

The Model F-value of 11.46 implies the model is significant. There is only
a 0.14% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case AB, CD are significant model terms.

Final Equation in Terms of Coded Factors:

Range =
+0.37
+0.089 * A * B
+0.10 * C * D

(f) What recommendations would you make regarding process operations? Use interaction and/or main
effect plots to assist in drawing conclusions.

From the interaction plots, choose A at the high level and B at the low level. In each of these plots, D can
be at either level. From the main effects plot of C, choose C at the high level. Based on the range analysis,
with C at the high level, D should be set at the low level.

From the analysis of the crack length data:

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

Inte ra c tio n G ra p h Inte ra c tio n G ra p h
Crack Length B : T ita n iu m C o n te n t Crack Length C : H e a t T re a t M e th o d
1 9 .8 2 4 1 9 .8 2 4

X = A: Pour Temp X = A: Pour Temp

Y = B: Titanium Content Y = C: Heat Treat Method
1 5 .8 9 2 5 1 5 .8 9 2 5
C r a c k L e n g th

C r a c k L e n g th

B- -1.000 C1 -1
B+ 1.000 C2 1
Actual Factors Actual Factors
C: Heat Treat Method = 1 B: Titanium Content = 0.00
D: Grain Refiner = 0.00 D: Grain Refiner = 0.00
1 1 .9 6 1 1 1 .9 6 1

8 .0 2 9 5 8 .0 2 9 5

4 .0 9 8 4 .0 9 8

-1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0 -1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0

A: P o u r T e m p A: P o u r T e m p

6-25
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

O ne F a c to r P lo t C ub e G ra p h
Crack Length Crack Length C ra c k L e n g th
1 9 .8 2 4
X = A: Pour Temp
Y = B: Titanium Content 1 0 .1 8 1 4 .2 7
X = D: Grain Refiner Z = C: Heat Treat Method

Actual Factors Actual Factor

1 5 .8 9 2 5

C r a c k L e n g th
A: Pour Temp = 0.00 D: Grain Refiner = 0.00
B: Titanium Content = 0.00

C o n te n t
C: Heat Treat Method = 1
B+ 1 2 .8 1 1 8 .6 4
1 1 .9 6 1

B : T i ta n i u m
8 .0 2 9 5 1 1 .1 8 5 .1 2 C+

C : H e a t T re a t M e
4 .0 9 8

B- 7 .7 3 1 5 .9 6 C -
-1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0 A- A+
A: P o u r T e m p
D : G ra in R e fin e r

From the analysis of the ranges:

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

Inte ra c tio n G ra p h Inte ra c tio n G ra p h
Range B : T ita n iu m C o n te n t Range D : G ra in R e fin e r
0 .6 6 1 0 .6 6 1

X = A: Pour Temp X = C: Heat Treat Method

Y = B: Titanium Content Y = D: Grain Refiner
0 .5 2 2 5 0 .5 2 2 5
B- -1.000 D- -1.000
B+ 1.000 D+ 1.000
Actual Factors Actual Factors
R ange

R ange

C: Heat Treat Method = 0.00 A: Pour Temp = 0.00

D: Grain Refiner = 0.00 B: Titanium Content = 0.00
0 .3 8 4 0 .3 8 4

0 .2 4 5 5 0 .2 4 5 5

0 .1 0 7 0 .1 0 7

-1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0 -1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0

A: P o u r T e m p C : H e a t T re a t M e th o d

6.16. Continuation of Problem 6.15. One of the variables in the experiment described in Problem 6.15,
heat treatment method (c), is a categorical variable. Assume that the remaining factors are continuous.

(a) Write two regression models for predicting crack length, one for each level of the heat treatment
method variable. What differences, if any, do you notice in these two equations?

Design Expert Output

Final Equation in Terms of Actual Factors

Heat Treat Method -1

Crack Length =
+13.78619
+3.51331 * Pour Temp
+1.93994 * Titanium Content
+0.97888 * Grain Refiner

6-26
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

-0.60169 * Pour Temp * Titanium Content

Heat Treat Method 1

Crack Length =
+10.18994
-0.49444 * Pour Temp
+2.03594 * Titanium Content
+0.97888 * Grain Refiner
+2.53581 * Pour Temp * Titanium Content

(b) Generate appropriate response surface contour plots for the two regression models in part (a).
DESIGN-EXPERT Plot DESIGN-EXPERT Plot
1 .0 0
C ra c k L e ng th 1 .0 0
C ra c k L e ng th
Crack Length Crack Length
X = A: Pour Temp 18 X = A: Pour Temp
Y = B: Titanium Content Y = B: Titanium Content 12
C o n te n t

C o n te n t
Actual Factors Actual Factors
C: Heat Treat Method = -1 0 .5 0 C: Heat Treat Method = 1 0 .5 0
D: Grain Refiner = 0.00 D: Grain Refiner = 0.00
B : T i ta n i u m

B : T i ta n i u m
14 16
0 .0 0 0 .0 0
12 10

-0 .5 0 10 -0 .5 0
8

6
-1 .0 0 -1 .0 0
-1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0 -1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0

A: P o u r T e m p A: P o u r T e m p

(c) What set of conditions would you recommend for the factors A, B and D if you use heat treatment
method C=+?

High level of A, low level of B, and low level of D.

(d) Repeat part (c) assuming that you wish to use heat treatment method C=-.

Low level of A, low level of B, and low level of D.

6.17. An experimenter has run a single replicate of a 24 design. The following effect estimates have been
calculated:

A = 76.95 AB = -51.32 ABC = -2.82

B = -67.52 AC = 11.69 ABD = -6.50
C = -7.84 AD = 9.78 ACD = 10.20
D = -18.73 BC = 20.78 BCD = -7.98
BD = 14.74 ABCD = -6.25
CD = 1.27

(a) Construct a normal probability plot of these effects.

The plot from Minitab follows.

6-27
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal Probability Plot

.999
.99
A
.95

Probability
.80
.50
.20 AB
.05 B
.01
.001

-50 0 50
Value
Average: -1.57 Anderson-Darling Normality Test
StDev: 32.3257 A-Squared: 0.781
N: 15 P-Value: 0.033

(b) Identify a tentative model, based on the plot of the effects in part (a).

öy = Intercept + 38.475 x A − 33.76 x B − 25.66 x A x B

6.18. The effect estimates from a 24 factorial design are as follows: ABCD = -1.5138, ABC = -1.2661,
ABD = -0.9852, ACD = -0.7566, BCD = -0.4842, CD = -0.0795, BD = -0.0793, AD = 0.5988, BC =
0.9216, AC = 1.1616, AB = 1.3266, D = 4.6744, C = 5.1458, B = 8.2469, and A = 12.7151. Are you
comfortable with the conclusions that all main effects are active?

Probability Plot of Effect Estimate

Normal - 95% CI
99
Mean 1.975
StDev 4.049
95 N 15
AD 1.280
90
P-Value <0.005
80
70
Percent

60
50
40
30
20

10

1
-10 -5 0 5 10 15
Effect Estimate

The upper right 4 dots are the four main effects. Since they do not follow the rest of the data, the normal
probability plot shows that they are active.

6.19. The effect estimates from a 24 factorial experiment are listed here. Are any of the effects
significant?

6-28
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

ABCD = -2.5251 AD = -1.6564

BCD = 4.4054 AC = 1.1109
ACD = -0.4932 AB = -10.5229
ABD = -5.0842 D= -6.0275
ABC = -5.7696 C= -8.2045
CD = 4.6707 B= -6.5304
BD = -4.6620 A= -0.7914
BC = -0.7982

Probability Plot of Effect Est

Normal - 95% CI
99
Mean -2.859
StDev 4.379
95 N 15
AD 0.221
90
P-Value 0.794
80
70
Percent

60
50
40
30
20

10

1
-20 -15 -10 -5 0 5 10
Effect Est

No effects appear to be significant.

6.20. Consider a variation of the bottle filling experiment from Example 5.3. Suppose that only two levels
of carbonation are used so that the experiment is a 23 factorial design with two replicates. The data are
shown below.

Coded Factors Fill Height Deviation

Run A B C Replicate 1 Replicate 2
1 - - - -3 -1
2 + - - 0 1
3 - + - -1 0
4 + + - 2 3
5 - - + -1 0
6 + - + 2 1
7 - + + 1 1
8 + + + 6 5

Factor Levels
Low (-1) High (+1)
A (%) 10 12

6-29
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

B (psi) 25 30
C (b/m) 200 250

(a) Analyze the data from this experiment. Which factors significantly affect fill height deviation?

The half normal probability plot of effects shown below identifies the factors A, B, and C as being
significant and the AB interaction as being marginally significant. The analysis of variance in the Design
Expert output below confirms that factors A, B, and C are significant and the AB interaction is marginally
significant.

DESIGN-EXPERT Plot
Fill Deviation
H a lf N o rm a l p lo t

H a l f N o r m a l % p r o b a b i l i ty
A: Carbonation 99
B: Pressure
C: Speed
97
A
95

90
B
85
80 C
AB
70

60

40

20

0 .0 0 0 .7 5 1 .5 0 2 .2 5 3 .0 0

|E ffe c t|

Design Expert Output

Response: Fill Deviation
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 70.75 4 17.69 26.84 < 0.0001 significant
A 36.00 1 36.00 54.62 < 0.0001
B 20.25 1 20.25 30.72 0.0002
C 12.25 1 12.25 18.59 0.0012
AB 2.25 1 2.25 3.41 0.0917
Residual 7.25 11 0.66
Lack of Fit 2.25 3 0.75 1.20 0.3700 not significant
Pure Error 5.00 8 0.63
Cor Total 78.00 15

The Model F-value of 26.84 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, C are significant model terms.

Std. Dev. 0.81 R-Squared 0.9071

Mean 1.00 Adj R-Squared 0.8733
C.V. 81.18 Pred R-Squared 0.8033
PRESS 15.34 Adeq Precision 15.424

(b) Analyze the residuals from this experiment. Are there any indications of model inadequacy?

The residual plots below do not identify any violations to the assumptions.

6-30
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal plot of residuals Residuals vs. Predicted

1.125

99

95
0.5625
Norm al % probability

90

80

Res iduals
70

50 0

2
30
20

10
-0.5625
5

-1.125

-1.125 -0.5625 0 0.5625 1.125 -2.13 -0.38 1.38 3.13 4.88

Res idual Predicted

Residuals vs. Run Residuals vs. Carbonation

1.125 1.125

0.5625 0.5625
Res iduals

Res iduals

2
0 0

2 2

-0.5625 -0.5625

-1.125 -1.125

1 4 7 10 13 16 10 11 12

Run Num ber Carbonation

Residuals vs. Pressure Residuals vs. Speed

1.125 1.125

0.5625 0.5625
Res iduals

Res iduals

2
0 0

3 2 2

-0.5625 -0.5625

-1.125 -1.125

25 26 27 28 29 30 200 208 217 225 233 242 250

Pres s ure Speed

6-31
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(c) Obtain a model for predicting fill height deviation in terms of the important process variables. Use this
model to construct contour plots to assist in interpreting the results of the experiment.

The model in both coded and actual factors are shown below.

Design Expert Output

Final Equation in Terms of Coded Factors

Fill Deviation =
+1.00
+1.50 *A
+1.13 *B
+0.88 *C
+0.38 *A*B

Final Equation in Terms of Actual Factors

Fill Deviation =
+9.62500
-2.62500 * Carbonation
-1.20000 * Pressure
+0.035000 * Speed
+0.15000 * Carbonation * Pressure

The following contour plots identify the fill deviation with respect to carbonation and pressure. The plot on
the left sets the speed at 200 b/m while the plot on the right sets the speed at 250 b/m. Assuming a faster
bottle speed is better, settings in pressure and carbonation that produce a fill deviation near zero can be
found in the lower left hand corner of the contour plot on the right.

30.00
2 Fill Deviation 2
30.00
2 Fill Deviation 2
3
4.5
2.5
4
2
3.5
28.75 28.75
1.5
3
1
B: Pres s ure

B: Pres s ure

2.5
0.5
27.50 27.50
2
-0.5 0 1.5
1

-1

0.5
26.25 26.25
-1.5

-2
2 2 2 2
25.00 25.00

10.00 10.50 11.00 11.50 12.00 10.00 10.50 11.00 11.50 12.00

A: Carbonation A: Carbonation

Speed set at 200 b/m Speed set at 250 b/m

(d) In part (a), you probably noticed that there was an interaction term that was borderline significant. If
you did not include the interaction term in your model, include it now and repeat the analysis. What
difference did this make? If you elected to include the interaction term in part (a), remove it and repeat
the analysis. What difference does this make?

The following analysis of variance, residual plots, and contour plots represent the model without the
interaction. As in the original analysis, the residual plots do not identify any concerns with the
assumptions. The contour plots did not change significantly either. The interaction effect is small relative
to the main effects.

Design Expert Output

6-32
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Response: Fill Deviation

ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 68.50 3 22.83 28.84 < 0.0001 significant
A 36.00 1 36.00 45.47 < 0.0001
B 20.25 1 20.25 25.58 0.0003
C 12.25 1 12.25 15.47 0.0020
Residual 9.50 12 0.79
Lack of Fit 4.50 4 1.13 1.80 0.2221 not significant
Pure Error 5.00 8 0.63
Cor Total 78.00 15

The Model F-value of 28.84 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, C are significant model terms.

Std. Dev. 0.89 R-Squared 0.8782

Mean 1.00 Adj R-Squared 0.8478
C.V. 88.98 Pred R-Squared 0.7835
PRESS 16.89 Adeq Precision 15.735

Final Equation in Terms of Coded Factors

Fill Deviation =
+1.00
+1.50 *A
+1.13 *B
+0.88 *C

Final Equation in Terms of Actual Factors

Fill Deviation =
-35.75000
+1.50000 * Carbonation
+0.45000 * Pressure
+0.035000 * Speed

Normal plot of residuals Residuals vs. Predicted

1.5

99

95
0.8125
Norm al % probability

90

80
Res iduals

70

50 0.125

30
20

10 2
-0.5625
5

-1.25

-1.25 -0.5625 0.125 0.8125 1.5 -2.50 -0.75 1.00 2.75 4.50

Res idual Predicted

6-33
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Residuals vs. Run Residuals vs. Carbonation

1.5 1.5

0.8125 0.8125

2
Res iduals

Res iduals
0.125 0.125

3
-0.5625 -0.5625

-1.25 -1.25

1 4 7 10 13 16 10 11 12

Run Num ber Carbonation

Residuals vs. Pressure Residuals vs. Speed

1.5 1.5

0.8125 0.8125
Res iduals

Res iduals

2 2
0.125 0.125

2 2

2 2 2 2
-0.5625 -0.5625

2 2

-1.25 -1.25

25 26 27 28 29 30 200 208 217 225 233 242 250

Pres s ure Speed

30.00 2
Fill Deviation 2 30.00 2
Fill Deviation 2
2.5
4
2
3.5
28.75 28.75
1.5
3
1
B: Pressure

B: Pressure

2.5

0 0.5 1.5 2
27.50 27.50
-0.5
1

-1
0.5
26.25 -1.5 26.25

0
-2
-0.5
25.00 2 2 25.00 2 2
10.00 10.50 11.00 11.50 12.00 10.00 10.50 11.00 11.50 12.00

A: Carbonation A: Carbonation

Speed set at 200 b/m Speed set at 250 b/m

6-34
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.21. I am always interested in improving my golf scores. Since a typical golfer uses the putter for about
35-45% of his or her strokes, it seems logical that in improving one’s putting score is a logical and perhaps
simple way to improve a golf score (“The man who can putt is a match for any man.” – Willie Parks, 1864-
1925, two-time winner of the British Open). An experiment was conducted to study the effects of four
factors on putting accuracy. The design factors are length of putt, type of putter, breaking putt vs. straight
putt, and level versus downhill putt. The response variable is distance from the ball to the center of the cup
after the ball comes to rest. One golfer performs the experiment, a 24 factorial design with seven replicates
was used, and all putts were made in random order. The results are as follows.

Design Factors Distance from cup (replicates)

Length of Type of Break Slope
putt (ft) putter of putt of putt 1 2 3 4 5 6 7
10 Mallet Straight Level 10.0 18.0 14.0 12.5 19.0 16.0 18.5
30 Mallet Straight Level 0.0 16.5 4.5 17.5 20.5 17.5 33.0
10 Cavity-back Straight Level 4.0 6.0 1.0 14.5 12.0 14.0 5.0
30 Cavity-back Straight Level 0.0 10.0 34.0 11.0 25.5 21.5 0.0
10 Mallet Breaking Level 0.0 0.0 18.5 19.5 16.0 15.0 11.0
30 Mallet Breaking Level 5.0 20.5 18.0 20.0 29.5 19.0 10.0
10 Cavity-back Breaking Level 6.5 18.5 7.5 6.0 0.0 10.0 0.0
30 Cavity-back Breaking Level 16.5 4.5 0.0 23.5 8.0 8.0 8.0
10 Mallet Straight Downhill 4.5 18.0 14.5 10.0 0.0 17.5 6.0
30 Mallet Straight Downhill 19.5 18.0 16.0 5.5 10.0 7.0 36.0
10 Cavity-back Straight Downhill 15.0 16.0 8.5 0.0 0.5 9.0 3.0
30 Cavity-back Straight Downhill 41.5 39.0 6.5 3.5 7.0 8.5 36.0
10 Mallet Breaking Downhill 8.0 4.5 6.5 10.0 13.0 41.0 14.0
30 Mallet Breaking Downhill 21.5 10.5 6.5 0.0 15.5 24.0 16.0
10 Cavity-back Breaking Downhill 0.0 0.0 0.0 4.5 1.0 4.0 6.5
30 Cavity-back Breaking Downhill 18.0 5.0 7.0 10.0 32.5 18.5 8.0

(a) Analyze the data from this experiment. Which factors significantly affect putting performance?

The half normal probability plot of effects identifies only factors A and B, length of putt and type of putter,
as having a potentially significant affect on putting performance. The analysis of variance with only these
significant factors is presented as well and confirms significance.

6-35
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot Ha lf No r m a l p lo t
Distance f rom cup

A: Length of putt
B: Ty pe of putter A
C: Break of putt
D: Slope of putt 99

97

Half Normal % probability

95 B

90

85
80

70

60

40

20

0 .0 0 1 .4 3 2 .8 6 4 .2 9 5 .7 2

| E ffec t|

Design Expert Output with Only Factors A and B

Response: Distance from cup
ANOVA for Selected Factorial Model
Analysis of variance table [Terms added sequentially (first to last)]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 1305.29 2 652.65 7.69 0.0008 significant
A 917.15 1 917.15 10.81 0.0014
B 388.15 1 388.15 4.57 0.0347
Residual 9248.94 109 84.85
Lack of Fit 933.15 13 71.78 0.83 0.6290 not significant
Pure Error 8315.79 96 86.62
Cor Total 10554.23 111

The Model F-value of 7.69 implies the model is significant. There is only
a 0.08% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, are significant model terms.

Std. Dev. 9.21 R-Squared 0.1237

Mean 12.30 Adj R-Squared 0.1076
C.V. 74.90 Pred R-Squared 0.0748
PRESS 9765.06 Adeq Precision 6.266

(b) Analyze the residuals from this experiment. Are there any indications of model inadequacy?

The residual plots for the model containing only the significant factors A and B are shown below. The
normality assumption appears to be violated. Also, as a golfer might expect, there is a slight inequality of
variance with regards to the length of putt. A square root transformation is applied which corrects the
violations. The analysis of variance and corrected residual plots are also presented. Finally, an effects plot
identifies a 10 foot putt and the cavity-back putter reduce the mean distance from the cup.

6-36
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal Plot of Residuals Residuals vs. Predicted

29.7009

99

95
18.0201
90
Normal % Probability

80
70

Residuals
2
2
50 6.33929
2
2 2
30
2
2
20
2
2 3 2
10 2 2
-5.34152 4
5 2 2 2 2
6

1 3
3

-17.0223 2

-17.0223 -5.34152 6.33929 18.0201 29.7009 7.58 9.94 12.30 14.66 17.02

Residual Predicted

Residuals vs. Length of putt Residuals vs. Type of putter

29.7009 29.7009

18.0201 18.0201
Residuals

Residuals

2
2 2
2
6.33929 6.33929
2 2
2 2 2
2
2
2 2
2
2
3
2 2 2
3 2
2
2 2 2
-5.34152 4 -5.34152 4
2 2
2 2 2
6 6

3 3
3 3

-17.0223 2 -17.0223 2

10 13 17 20 23 27 30 1 2

Length of putt T ype of putter

Residuals vs. Break of putt Residuals vs. Slope of putt

29.7009 29.7009

18.0201 18.0201
Residuals

Residuals

2
2 2
6.33929 6.33929
2 2
2
2 2
2 2 2 2
2 2
-5.34152 4 -5.34152 3 2
2
2 2
3
5 2 4

2 2
2 3

-17.0223 -17.0223

1 2 1 2

Break of putt Slope of putt

6-37
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output with Only Factors A and B and a Square Rot Transformation
Response: Distance from cupTransform:Square root Constant: 0
ANOVA for Selected Factorial Model
Analysis of variance table [Terms added sequentially (first to last)]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 37.26 2 18.63 7.85 0.0007 significant
A 21.61 1 21.61 9.11 0.0032
B 15.64 1 15.64 6.59 0.0116
Residual 258.63 109 2.37
Lack of Fit 30.19 13 2.32 0.98 0.4807 not significant
Pure Error 228.45 96 2.38
Cor Total 295.89 111

The Model F-value of 7.85 implies the model is significant. There is only
a 0.07% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, are significant model terms.

Std. Dev. 1.54 R-Squared 0.1259

Mean 3.11 Adj R-Squared 0.1099
C.V. 49.57 Pred R-Squared 0.0771
PRESS 273.06 Adeq Precision 6.450

Normal Plot of Residuals Residuals vs. Predicted

3.36106

99

95
1.54064
90 2
2
Normal % Probability

2
80 2 2
70 2
Residuals

2
2 2
3 2 2
50 -0.279783 2 4
2
2
30 2
20 2
10
-2.10021
5 6

1 3 3

-3.92063 2

-3.92063 -2.10021 -0.279783 1.54064 3.36106 2.29 2.70 3.11 3.51 3.92

Residual Predicted

6-38
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Residuals vs. Length of putt Residuals vs. Type of putter

3.36106 3.36106

1.54064 1.54064
2
2 2
2
2 2
2 2 2
2
2 2
Residuals

Residuals
2
2 2 2 2
2
3 2
2 3
2 2
-0.279783 2 4 -0.279783 2
4
2 2
2 2
2 2
2 2

-2.10021 -2.10021
6 6

3 3 3 3

-3.92063 2 -3.92063 2

10 13 17 20 23 27 30 1 2

Length of putt T ype of putter

Residuals vs. Break of putt Residuals vs. Slope of putt

3.36106 3.36106

1.54064 1.54064
2 2
2
2
Residuals

Residuals

2 2 2
2
2 2
-0.279783 4 -0.279783 3
2
2

-2.10021 -2.10021
5 2 4

2 2 2
3

-3.92063 -3.92063

1 2 1 2

Break of putt Slope of putt

DESIGN-EXPERT Plot
In te r a c tio n G r a p h
Sqrt(Distance f rom cup) B : T y p e o f p u tt e r
20

X = A: Length of putt
Y = B: Ty pe of putter
16
B1 Mallet
B2 Cav ity -back
Actual Factors
Distance from cup

C: Break of putt = Straight

12
D: Slope of putt = Lev el

1 0 .0 0 1 5 .0 0 2 0 .0 0 2 5 .0 0 3 0 .0 0

A : L e n g th o f p u t t

6-39
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.22. Semiconductor manufacturing processes have long and complex assembly flows, so matrix marks
and automated 2d-matrix readers are used at several process steps throughout factories. Unreadable matrix
marks negatively effect factory run rates, because manual entry of part data is required before
manufacturing can resume. A 24 factorial experiment was conducted to develop a 2d-matrix laser mark on
a metal cover that protects a substrate mounted die. The design factors are A = laser power (9W, 13W), B =
laser pulse frequency (4000 Hz, 12000 Hz), C = matrix cell size (0.07 in, 0.12 in), and D = writing speed
(10 in/sec, 20 in/sec), and the response variable is the unused error correction (UEC). This is a measure of
the unused portion of the redundant information embedded in the 2d matrix. A UEC of 0 represents the
lowest reading that still results in a decodable matrix while a value of 1 is the highest reading. A DMX
Verifier was used to measure UEC. The data from this experiment are shown below.

Standard Run Laser Pulse Writing

Order Order Power Frequency Cell Size Speed UEC
8 1 1 1 1 -1 0.80
10 2 1 -1 -1 1 0.81
12 3 1 1 -1 1 0.79
9 4 -1 -1 -1 1 0.60
7 5 -1 1 1 -1 0.65
15 6 -1 1 1 1 0.55
2 7 1 -1 -1 -1 0.98
6 8 1 -1 1 -1 0.67
16 9 1 1 1 1 0.69
13 10 -1 -1 1 1 0.56
5 11 -1 -1 1 -1 0.63
14 12 1 -1 1 1 0.65
1 13 -1 -1 -1 -1 0.75
3 14 -1 1 -1 -1 0.72
4 15 1 1 -1 -1 0.98
11 16 -1 1 -1 1 0.63

(a) Analyze the data from this experiment. Which factors significantly affect UEC?

The normal probability plot of effects identifies A, C, D, and the AC interaction as significant. The Design
Expert output including the analysis of variance confirms the significance and identifies the corresponding
model. Contour plots identify factors A and C with B held constant at zero and D toggled from -1 to +1.

6-40
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot
UEC
N o r m a l p lo t
A: Laser Power
B: Pulse Frequency 99
C: Cell Size
D: Writing Speed A
95
90

Normal % probability
80
70

50

30
20
AC
10 D
5
C
1

-0 .1 3 -0 .0 6 0 .0 1 0 .0 9 0 .1 6

E f fe c t

Design Expert Output

Response: UEC
ANOVA for Selected Factorial Model
Analysis of variance table [Terms added sequentially (first to last)]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 0.24 4 0.059 35.51 < 0.0001 significant
A 0.10 1 0.10 61.81 < 0.0001
C 0.070 1 0.070 42.39 < 0.0001
D 0.051 1 0.051 30.56 0.0002
AC 0.012 1 0.012 7.30 0.0206
Residual 0.018 11 1.657E-003
Cor Total 0.25 15

The Model F-value of 35.51 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, D, AC are significant model terms.

Std. Dev. 0.041 R-Squared 0.9281

Mean 0.72 Adj R-Squared 0.9020
C.V. 5.68 Pred R-Squared 0.8479
PRESS 0.039 Adeq Precision 17.799

Final Equation in Terms of Coded Factors

UEC =
+0.72
+0.080 *A
-0.066 *C
-0.056 *D
-0.027 *A*C

Final Equation in Terms of Actual Factors

UEC =
+0.71625
+0.080000 * Laser Power
-0.066250 * Cell Size
-0.056250 * Writing Speed
-0.027500 * Laser Power * Cell Size

6-41
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

1 .0 0
UE C 1 .0 0
UE C
UEC UEC 0 .5 5
X = A: Laser Power X = A: Laser Power
Y = C: Cell Size Y = C: Cell Size

Actual Factors Actual Factors

0 .5 0 0 .7 0 .5 0
B: Pulse Frequency = 0.00 B: Pulse Frequency = 0.00
D: Writing Speed = -1.00 D: Writing Speed = 1.00 0 .6

0 .7 5
C: Cell Size

C: Cell Size
0 .6 5

0 .0 0 0 .0 0
0 .8
0 .7

0 .8 5
-0 .5 0 -0 .5 0 0 .7 5

0 .9
0 .8

-1 .0 0 -1 .0 0
-1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0 -1 .0 0 -0 .5 0 0 .0 0 0 .5 0 1 .0 0

A : L a se r P o w e r A : L a se r P o w e r

(b) Analyze the residuals from this experiment. Are there any indications of model inadequacy?

The residual plots appear acceptable with the exception of run 8, standard order 6. This value should be
verified by the engineer.

Normal plot of residuals Residuals vs. Predicted

0.04375

2
99

95
0.010625
Norm al % probability

90

80
Res iduals

70

50 -0.0225

30
20

10
-0.055625
5

-0.08875

-0.08875 -0.055625 -0.0225 0.010625 0.04375 0.54 0.64 0.74 0.84 0.95

Res idual Predicted

6-42
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Residuals vs. Run Residuals vs. Laser Power

0.04375 0.04375

0.010625 0.010625
Res iduals

Res iduals
-0.0225 -0.0225

-0.055625 -0.055625

-0.08875 -0.08875

1 4 7 10 13 16 -1 0 1

Run Num ber Las er Power

Residuals vs. Pulse Frequency Residuals vs. Cell Size

0.04375 0.04375

0.010625 0.010625
Res iduals

Res iduals

-0.0225 2 -0.0225

-0.055625 -0.055625

-0.08875 -0.08875

-1 0 1 -1 0 1

Puls e Frequency Cell Size

Residuals vs. Writing Speed

0.04375

0.010625
Res iduals

-0.0225

-0.055625

-0.08875

-1 0 1

Writing Speed

6-43
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.23. Reconsider the experiment in Problem 6.22. Suppose that four center points are available, and the
UEC response at these four runs is 0.98, 0.95, 0.93 and 0.96, respectively. Reanalyze the experiment
incorporating a test for curvature into the analysis. What conclusions can you draw? What
recommendations would you make to the experimenters?

As with the results of problem 6.20, factors A, C, D, and the AC interaction remain significant. However,
the CD interaction and curvature are significant as well. The curvature is the strongest effect;
unfortunately, we are not able to determine which factor(s) have a quadratic term. We recommend that the
engineer augment the experiment with additional experimental runs, such as axial points and a couple of
extra center points for blocking purposes. These extra runs will determine the pure quadratic effects and
allow us to fit a second order model.

Design Expert Output

Response: UEC
ANOVA for Selected Factorial Model
Analysis of variance table [Terms added sequentially (first to last)]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 0.24 5 0.048 45.07 < 0.0001 significant
A 0.10 1 0.10 95.77 < 0.0001
C 0.070 1 0.070 65.68 < 0.0001
D 0.051 1 0.051 47.35 < 0.0001
AC 0.012 1 0.012 11.32 0.0051
CD 5.625E-003 1 5.625E-003 5.26 0.0391
Curvature 0.18 1 0.18 170.59 < 0.0001 significant
Residual 0.014 13 1.069E-003
Lack of Fit 0.013 10 1.260E-003 2.91 0.2057 not significant
Pure Error 1.300E-003 3 4.333E-004
Cor Total 0.44 19

The Model F-value of 45.07 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, D, AC, CD are significant model terms.

Std. Dev. 0.033 R-Squared 0.9455

Mean 0.76 Adj R-Squared 0.9245
C.V. 4.28 Pred R-Squared 0.9209
PRESS 0.035 Adeq Precision 20.936

Coefficient Standard 95% CI 95% CI

Factor Estimate DF Error Low High VIF
Intercept 0.72 1 8.175E-003 0.70 0.73
A-Laser Power 0.080 1 8.175E-003 0.062 0.098 1.00
C-Cell Size -0.066 1 8.175E-003 -0.084 -0.049 1.00
D-Writing Speed -0.056 1 8.175E-003 -0.074 -0.039 1.00
AC -0.028 1 8.175E-003 -0.045 -9.839E-003 1.00
CD 0.019 1 8.175E-003 1.089E-003 0.036 1.00
Center Point 0.24 1 0.018 0.20 0.28 1.00

6.24. A company markets its products by direct mail. An experiment was conducted to study the effects
of three factors on the customer response rate for a particular product. The three factors are A = type of
mail used (3rd class, 1st class), B = type of descriptive brochure (color, black-and-white), and C = offer price
($19.95, $24.95). The mailings are made to two groups of 8,000 randomly selected customers, with 1,000
customers in each group receiving each treatment combination. Each group of customers is considered as a
replicate. The response variable is the number of orders placed. The experimental data is shown below.

6-44
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Coded Factors Number of Orders

Run A B C Replicate 1 Replicate 2
1 - - - 50 54
2 + - - 44 42
3 - + - 46 48
4 + + - 42 43
5 - - + 49 46
6 + - + 48 45
7 - + + 47 48
8 + + + 56 54

Factor Levels
Low (-1) High (+1)
A (class) 3rd 1st
B (type) BW Color
C ($) $19.95 $24.95

(a) Analyze the data from this experiment. Which factors significantly affect the customer response rate?

The half normal probability plot of effects identifies the two factor interactions, AB, AC, BC, and factors A
and C as significant. Factor B is not significant; however, remains in the model to satisfy the hierarchal
principle. The analysis of variance confirms the significance of two factor interactions and factor C.
Factor A is marginally significant.
DESIGN-EXPERT Plot
Number of orders
H a lf N o r m a l p lo t
A: Class
99
B: Ty pe
C: Price
97
AC
Half Normal % probability

95

90 BC
85
80
AB
C
70 A
60
B
40

20

0 .0 0 1 .2 5 2 .5 0 3 .7 5 5 .0 0

| E ffe c t |

6-45
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output

Response: Number of orders
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 241.75 6 40.29 12.95 0.0006 significant
A 12.25 1 12.25 3.94 0.0785
B 2.25 1 2.25 0.72 0.4171
C 36.00 1 36.00 11.57 0.0079
AB 42.25 1 42.25 13.58 0.0050
AC 100.00 1 100.00 32.14 0.0003
BC 49.00 1 49.00 15.75 0.0033
Residual 28.00 9 3.11
Lack of Fit 4.00 1 4.00 1.33 0.2815 not significant
Pure Error 24.00 8 3.00
Cor Total 269.75 15

The Model F-value of 12.95 implies the model is significant. There is only
a 0.06% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case C, AB, AC, BC are significant model terms.

Std. Dev. 1.76 R-Squared 0.8962

Mean 47.63 Adj R-Squared 0.8270
C.V. 3.70 Pred R-Squared 0.6719
PRESS 88.49 Adeq Precision 10.286

Coefficient Standard 95% CI 95% CI

Factor Estimate DF Error Low High VIF
Intercept 47.63 1 0.44 46.63 48.62
A-Class -0.88 1 0.44 -1.87 0.12 1.00
B-Type 0.37 1 0.44 -0.62 1.37 1.00
C-Price 1.50 1 0.44 0.50 2.50 1.00
AB 1.63 1 0.44 0.63 2.62 1.00
AC 2.50 1 0.44 1.50 3.50 1.00
BC 1.75 1 0.44 0.75 2.75 1.00

(b) Analyze the residuals from this experiment. Are there any indications of model inadequacy?

The residual plots below do not identify model inadequacy.

Normal Plot of Residuals Residuals vs. Predicted

3

99

95
1.5
90
Normal % Probability

80
70
Residuals

50 0

30
20

10
-1.5
5

-3

-2.5 -1.375 -0.25 0.875 2 42.50 45.50 48.50 51.50 54.50

Residual Predicted

6-46
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Residuals vs. Run Residuals vs. Class

3 3

1.5 1.5 2 2
Residuals

Residuals
0 0

-1.5 -1.5

-3 -3

1 4 7 10 13 16 1 2

Run Number Class

Residuals vs. Type Residuals vs. Price

3 3

1.5 2 2 1.5 3
Residuals

Residuals

0 2 0

2 2

2 2

-1.5 -1.5

-3 -3

1 2 19.95 21.20 22.45 23.70 24.95

T ype Price

(c) What would you recommend to the company?

Based on the interaction plots below, we recommend 3rd class mail, black-and-white brochures, and an
offered price of $19.95 would achieve the greatest number of orders. If the offered price must be $24.95,
then the 1st class mail with color brochures is recommended.

6-47
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

In te r a c tio n G r a p h In te r a c tio n G r a p h
Number of orders B : T yp e Number of orders C : P ri c e
56 56

X = A: Class X = A: Class
Y = B: Ty pe Y = C: Price

B1 BW 5 2 .5
Design Points 5 2 .1 5 8 5
B2 Color
Actual Factor C- 19.950
Number of orders

Number of orders
C: Price = 22.45 C+ 24.950
Actual Factor
B: Ty pe = BW
49 4 8 .3 1 6 9

4 5 .5 4 4 .4 7 5 4

42 4 0 .6 3 3 8

3 rd 1st 3 rd 1st

A : C l a ss A : C l a ss

DESIGN-EXPERT Plot
In te r a c tio n G r a p h
Number of orders C : P ri c e
56

X = B: Ty pe
Y = C: Price

Design Points 5 2 .5

C- 19.950
Number of orders

C+ 24.950
Actual Factor
A: Class = 3rd
49

4 5 .5

42

BW C o lo r

B : T yp e

6.25. Consider the single replicate of the 24 design in Example 6.2. Suppose we had arbitrarily decided to
analyze the data assuming that all three- and four-factor interactions were negligible. Conduct this analysis
and compare your results with those obtained in the example. Do you think that it is a good idea to
arbitrarily assume interactions to be negligible even if they are relatively high-order ones?

Design Expert Output

Response: Filtration Ratein A/min
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 5603.13 10 560.31 21.92 0.0016 significant
A 1870.56 1 1870.56 73.18 0.0004
B 39.06 1 39.06 1.53 0.2713
C 390.06 1 390.06 15.26 0.0113
D 855.56 1 855.56 33.47 0.0022
AB 0.063 1 0.063 2.445E-003 0.9625
AC 1314.06 1 1314.06 51.41 0.0008
AD 1105.56 1 1105.56 43.25 0.0012

6-48
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

BC 22.56 1 22.56 0.88 0.3906

BD 0.56 1 0.56 0.022 0.8879
CD 5.06 1 5.06 0.20 0.6749
Residual 127.81 5 25.56
Cor Total 5730.94 15

The Model F-value of 21.92 implies the model is significant. There is only
a 0.16% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, D, AC, AD are significant model terms.

This analysis of variance identifies the same effects as the normal probability plot of effects approach used
in Example 6.2. In general, it is not a good idea to arbitrarily pool interactions. Use the normal probability
plot of effect estimates as a guide in the choice of which effects to tentatively include in the model.

6.26. An experiment was run in a semiconductor fabrication plant in an effort to increase yield. Five
factors, each at two levels, were studied. The factors (and levels) were A = aperture setting (small, large), B
= exposure time (20% below nominal, 20% above nominal), C = development time (30 s, 45 s), D = mask
dimension (small, large), and E = etch time (14.5 min, 15.5 min). The unreplicated 25 design shown below
was run.

(1) = 7 d= 8 e= 8 de = 6
a= 9 ad = 10 ae = 12 ade = 10
b= 34 bd = 32 be = 35 bde = 30
ab = 55 abd = 50 abe = 52 abde = 53
c= 16 cd = 18 ce = 15 cde = 15
ac = 20 acd = 21 ace = 22 acde = 20
bc = 40 bcd = 44 bce = 45 bcde = 41
abc = 60 abcd = 61 abce = 65 abcde = 63

(a) Construct a normal probability plot of the effect estimates. Which effects appear to be large?

From the normal probability plot of effects shown below, effects A, B, C, and the AB interaction appear to
be large.
DESIGN-EXPERT Plot
Yield
N o rm a l p lo t
A: Aperture
B: Exposure Time
99 B
N o r m a l % p r o b a b i l i ty

C: Develop Time
D: Mask Dimension A
E: Etch Time 95 C
90
AB

80
70

50

30
20
10
5

-1 .1 9 7 .5 9 1 6 .3 8 2 5 .1 6 3 3 .9 4

E ffe c t

(b) Conduct an analysis of variance to confirm your findings for part (a).

6-49
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output

Response: Yield
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 11585.13 4 2896.28 991.83 < 0.0001 significant
A 1116.28 1 1116.28 382.27 < 0.0001
B 9214.03 1 9214.03 3155.34 < 0.0001
C 750.78 1 750.78 257.10 < 0.0001
AB 504.03 1 504.03 172.61 < 0.0001
Residual 78.84 27 2.92
Cor Total 11663.97 31

The Model F-value of 991.83 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, C, AB are significant model terms.

(c) Write down the regression model relating yield to the significant process variables.

Design Expert Output

Final Equation in Terms of Actual Factors:
Aperture small
Yield =
+0.40625
+0.65000 * Exposure Time
+0.64583 * Develop Time

Aperture large
Yield =
+12.21875
+1.04688 * Exposure Time
+0.64583 * Develop Time

(d) Plot the residuals on normal probability paper. Is the plot satisfactory?

Normal plot of residuals

99

95
Norm al % probability

90
80
70

50

30
20
10
5

-2.78125 -1.39063 -3.55271E-015 1.39062 2.78125

Res idual

There is nothing unusual about this plot.

(e) Plot the residuals versus the predicted yields and versus each of the five factors. Comment on the plots.

6-50
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Residuals vs. Aperture Residuals vs. Exposure Time

2.78125 2.78125

2
1.39062 1.39062
3 2
Res iduals

Res iduals
-3.55271E-015 -3.55271E-015
2 2
2 2
2

-1.39063 3 -1.39063 2

-2.78125 -2.78125

1 2 -20 -13 -7 0 7 13 20

Aperture Expos ure Tim e

Residuals vs. Develop Time Residuals vs. Mask Dimension

2.78125 2.78125

2 2
1.39062 1.39062
3 2
Res iduals

Res iduals

-3.55271E-015 -3.55271E-015
2
2 2
2 2

-1.39063 3 -1.39063 2

-2.78125 -2.78125

30 33 35 38 40 43 45 1 2

Develop Tim e Mas k Dim ens ion

Residuals vs. Etch Time

2.78125

2
1.39062
2
Res iduals

-3.55271E-015

-1.39063 3

-2.78125

14.50 14.75 15.00 15.25 15.50

Etch Tim e

6-51
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

The plot of residual versus exposure time shows some very slight inequality of variance. There is no strong
evidence of a potential problem.

(f) Interpret any significant interactions.

DESIGN-EXPERT Plot
Inte ra c tio n G ra p h
Yield A p e rtu re
65

X = B: Exposure Time
Y = A: Aperture
5 0 .2 5
A1 small
A2 large
Actual Factors
C: Develop Time = 37.50

Yie ld
D: Mask Dimension = Small
3 5 .5
E: Etch Time = 15.00

2 0 .7 5

-2 0 .0 0 -1 0 .0 0 0 .0 0 1 0 .0 0 2 0 .0 0

E xp o s u re T im e

Factor A does not have as large an effect when B is at its low level as it does when B is at its high level.

(g) What are your recommendations regarding process operating conditions?

To achieve the highest yield, run B at the high level, A at the high level, and C at the high level.

(h) Project the 25 design in this problem into a 2k design in the important factors. Sketch the design and
show the average and range of yields at each run. Does this sketch aid in interpreting the results of this
experiment?

DESIGN-EASE Analysis
Actual Yield

42.5000 62.2500
R=5 R=5

B+ 32.7500 52.5000
E R=5 R=5
x
p
o
s
u
r 16.0000 20.7500 C+
e e
R=3 R=2 m
i
T T
i p
m o
e l
e
B- 7.2500 10.2500 C- v
e
A- R=2 R=3 A+ D
Aperture

This cube plot aids in interpretation. The strong AB interaction and the large positive effect of C are clearly
evident.

6-52
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.27. Continuation of Problem 6.26. Suppose that the experimenter had run four runs at the center
points in addition to the 32 trials in the original experiment. The yields obtained at the center point runs
were 68, 74, 76, and 70.

(a) Reanalyze the experiment, including a test for pure quadratic curvature.

Because aperture and mask dimension are not continuous variables, the four center points were split
amongst these two factors as follows.

Mask
Aperture Dimension Yield
Small Small 68
Large Small 74
Small Large 76
Large Large 70

The sum of squares for the curvature can be estimated with the following equation and is confirmed with
the analysis of variance shown in the Design Expert output.

n F n C (y F − y C )2 (32)(4 )(30.53125 − 72)2

SS PureQuadratic = = = 6114.337
n F + nC 32 + 4

Design Expert Output

Response: Yield
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 11461.09 4 2865.27 353.92 < 0.0001 significant
A 992.25 1 992.25 122.56 < 0.0001
B 9214.03 1 9214.03 1138.12 < 0.0001
C 750.78 1 750.78 92.74 < 0.0001
AB 504.03 1 504.03 62.26 < 0.0001
Curvature 6114.34 1 6114.34 755.24 < 0.0001 significant
Residual 242.88 30 8.10
Cor Total 17818.31 35

The Model F-value of 353.92 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, C, AB are significant model terms.

Discuss what your next step would be.

Add axial points for factors B and C along with four more center points to fit a second-order model and
satisfy blocking concerns.

6.28. In a process development study on yield, four factors were studied, each at two levels: time (A),
concentration (B), pressure (C), and temperature (D). A single replicate of a 24 design was run, and the
resulting data are shown in the following table:

6-53
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Actual
Run Run Yield Factor Levels
Number Order A B C D (lbs) Low (-) High (+)
1 5 - - - - 12 A (h) 2.5 3.0
2 9 + - - - 18 B (%) 14 18
3 8 - + - - 13 C (psi) 60 80
4 13 + + - - 16 D (ºC) 225 250
5 3 - - + - 17
6 7 + - + - 15
7 14 - + + - 20
8 1 + + + - 15
9 6 - - - + 10
10 11 + - - + 25
11 2 - + - + 13
12 15 + + - + 24
13 4 - - + + 19
14 16 + - + + 21
15 10 - + + + 17
16 12 + + + + 23

(a) Construct a normal probability plot of the effect estimates. Which factors appear to have large effects?

DESIGN-EXPERT Plot
Yield
N o rm a l p lo t
A: Time
B: Concentration
99
N o r m a l % p r o b a b i l i ty

C: Pressure
A
D: Temperature
95
AD
90 D
80 C
70

50

30
20
10
5 AC

-4 .2 5 -2 .0 6 0 .1 3 2 .3 1 4 .5 0

E ffe c t

A, C, D and the AC and AD interactions appear to have large effects.

(b) Conduct an analysis of variance using the normal probability plot in part (a) for guidance in forming an
error term. What are your conclusions?

Design Expert Output

Response: Yield
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F

6-54
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Model 275.50 5 55.10 33.91 < 0.0001 significant

A 81.00 1 81.00 49.85 < 0.0001
C 16.00 1 16.00 9.85 0.0105
D 42.25 1 42.25 26.00 0.0005
AC 72.25 1 72.25 44.46 < 0.0001
AD 64.00 1 64.00 39.38 < 0.0001
Residual 16.25 10 1.62
Cor Total 291.75 15

The Model F-value of 33.91 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, D, AC, AD are significant model terms.

(c) Write down a regression model relating yield to the important process variables.

Design Expert Output

Final Equation in Terms of Coded Factors:

Yield =
+17.38
+2.25 *A
+1.00 *C
+1.63 *D
-2.13 *A*C
+2.00 *A*D

Final Equation in Terms of Actual Factors:

Yield =
+209.12500
-83.50000 * Time
+2.43750 * Pressure
-1.63000 * Temperature
-0.85000 * Time * Pressure
+0.64000 * Time * Temperature

(d) Analyze the residuals from this experiment. Does your analysis indicate any potential problems?

Normal plot of residuals Residuals vs. Predicted

1.375

99

95
0.625
Norm al % probability

90
80
Res iduals

70 2

50 -0.125

30
20
10
-0.875
5

-1.625

-1.625 -0.875 -0.125 0.625 1.375 11.63 14.81 18.00 21.19 24.38

Res idual Predicted

6-55
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Residuals vs. Run

1.375

0.625
Res iduals

-0.125

-0.875

-1.625

1 4 7 10 13 16

Run Num ber

There is nothing unusual about the residual plots.

(e) Can this design be collapsed into a 23 design with two replicates? If so, sketch the design with the
average and range of yield shown at each point in the cube. Interpret the results.

DESIGN-EASE Analysis
Actual yield

18.0 22.0
R=2 R=2

D+ 11.5 24.5
R=3 R=1
t
e
m
p
e
r 18.5 15.0 C+
a R=3 R=0
t e
u r
u
r s
e s
e
r
D- 12.5 17.0 C- p
A- R=1 R=2 A+
time

6.29. Continuation of Problem 6.28. Use the regression model in part (c) of Problem 6.23 to generate a
response surface contour plot of yield. Discuss the practical purpose of this response surface plot.

The response surface contour plot shows the adjustments in the process variables that lead to an increasing
or decreasing response. It also displays the curvature of the response in the design region, possibly

6-56
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

indicating where robust operating conditions can be found. Two response surface contour plots for this
process are shown below. These were formed from the model written in terms of the original design
variables.

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

8 0 .0 0
Yie ld 2 5 0 .0 0
Yie ld
Yield Yield
X = A: Time X = A: Time
Y = C: Pressure Y = D: Temperature

Actual Factors Actual Factors

B: Concentration = 16.00 7 5 .0 0 B: Concentration = 16.002 4 3 .7 5
1 7 .8 3 3 3 1 9 .2 9 1 7

T e m p e r a tu r e
D: Temperature = 237.50 C: Pressure = 70.00
P re s s u re

1 7 .8 3 3 3
7 0 .0 0 2 3 7 .5 0
1 6 .3 7 5

1 9 .2 9 1 7
1 6 .3 7 5
1 4 .9 1 6 7
6 5 .0 0 2 3 1 .2 5

1 3 .4 5 8 3

6 0 .0 0 2 2 5 .0 0
2 .5 0 2 .6 3 2 .7 5 2 .8 8 3 .0 0 2 .5 0 2 .6 3 2 .7 5 2 .8 8 3 .0 0

T im e T im e

6.30. The scrumptious brownie experiment. The author is an engineer by training and a firm believer in
learning by doing. I have taught experimental design for many years to a wide variety of audiences and
have always assigned the planning, conduct, and analysis of an actual experiment to the class participants.
The participants seem to enjoy this practical experience and always learn a great deal from it. This problem
uses the results of an experiment performed by Gretchen Krueger at Arizona State University.

There are many different ways to bake brownies. The purpose of this experiment was to determine how the
pan material, the brand of brownie mix, and the stirring method affect the scrumptiousness of brownies.
The factor levels were
Factor Low (-) High (+)
A = pan material Glass Aluminum
B = stirring method Spoon Mixer
C = brand of mix Expensive Cheap
The response variable was scrumptiousness, a subjective measure derived from a questionnaire given to the
subjects who sampled each batch of brownies. (The questionnaire dealt with such issues as taste,
appearance, consistency, aroma, and so forth.) An eight-person test panel sampled each batch and filled
out the questionnaire. The design matrix and the response data are shown below:

Brownie Test Panel Results

Batch A B C 1 2 3 4 5 6 7 8
1 - - - 11 9 10 10 11 10 8 9
2 + - - 15 10 16 14 12 9 6 15
3 - + - 9 12 11 11 11 11 11 12
4 + + - 16 17 15 12 13 13 11 11
5 - - + 10 11 15 8 6 8 9 14
6 + - + 12 13 14 13 9 13 14 9
7 - + + 10 12 13 10 7 7 17 13
8 + + + 15 12 15 13 12 12 9 14

6-57
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(a) Analyze the data from this experiment as if there were eight replicates of a 23 design. Comment on the
results.
Design Expert Output
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 93.25 7 13.32 2.20 0.0475 significant
A 72.25 1 72.25 11.95 0.0010
B 18.06 1 18.06 2.99 0.0894
C 0.063 1 0.063 0.010 0.9194
AB 0.062 1 0.062 0.010 0.9194
AC 1.56 1 1.56 0.26 0.6132
BC 1.00 1 1.00 0.17 0.6858
ABC 0.25 1 0.25 0.041 0.8396
Residual 338.50 56 6.04
Lack of Fit 0.000 0
Pure Error 338.50 56 6.04
Cor Total 431.75 63

The Model F-value of 2.20 implies the model is significant. There is only
a 4.75% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A are significant model terms.

In this analysis, A, the pan material and B, the stirring method, appear to be significant. There are 56
degrees of freedom for the error, yet only eight batches of brownies were cooked, one for each recipe.

(b) Is the analysis in part (a) the correct approach? There are only eight batches; do we really have eight
replicates of a 23 factorial design?

The different rankings by the taste-test panel are not replicates, but repeat observations by different testers
on the same batch of brownies. It is not a good idea to use the analysis in part (a) because the estimate of
error may not reflect the batch-to-batch variation.

(c) Analyze the average and standard deviation of the scrumptiousness ratings. Comment on the results.
Is this analysis more appropriate than the one in part (a)? Why or why not?
DESIGN-EXPERT Plot DESIGN-EXPERT Plot
Average
N o rm a l p lo t Stdev
N o rm a l p lo t
A: Pan Material 99 A: Pan Material 99
B: Stirring Method B: Stirring Method
N o r m a l % p r o b a b i l i ty

N o r m a l % p r o b a b i l i ty

C: Mix Brand C: Mix Brand

95 95
90 A 90 C
80 80
B
70 70
A
50 50

30 30
20 20
10 10
5 5 AC

1 1

-0 .3 2 0 .2 9 0 .9 1 1 .5 2 2 .1 3 -1 .5 7 -1 .0 1 -0 .4 5 0 .1 1 0 .6 8

E ffe c t E ffe c t

6-58
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output

Response: Average
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 11.28 2 5.64 76.13 0.0002 significant
A 9.03 1 9.03 121.93 0.0001
B 2.25 1 2.25 30.34 0.0027
Residual 0.37 5 0.074
Cor Total 11.65 7

The Model F-value of 76.13 implies the model is significant. There is only
a 0.02% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B are significant model terms.

Design Expert Output

Response: Stdev
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 6.05 3 2.02 9.77 0.0259 significant
A 0.24 1 0.24 1.15 0.3432
C 0.91 1 0.91 4.42 0.1034
AC 4.90 1 4.90 23.75 0.0082
Residual 0.82 4 0.21
Cor Total 6.87 7

The Model F-value of 9.77 implies the model is significant. There is only
a 2.59% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case AC are significant model terms.

Variables A and B affect the mean rank of the brownies. Note that the AC interaction affects the standard
deviation of the ranks. This is an indication that both factors A and C have some effect on the variability in
the ranks. It may also indicate that there is some inconsistency in the taste test panel members. For the
analysis of both the average of the ranks and the standard deviation of the ranks, the mean square error is
now determined by pooling apparently unimportant effects. This is a more accurate estimate of error than
obtained assuming that all observations were replicates.

6.31. An experiment was conducted on a chemical process that produces a polymer. The four factors
studied were temperature (A), catalyst concentration (B), time (C), and pressure (D). Two responses,
molecular weight and viscosity, were observed. The design matrix and response data are shown below:

Actual
Run Run Molecular Factor Levels
Number Order A B C D Weight Viscosity Low (-) High (+)
1 18 - - - - 2400 1400 A (ºC) 100 120
2 9 + - - - 2410 1500 B (%) 4 8
3 13 - + - - 2315 1520 C (min) 20 30
4 8 + + - - 2510 1630 D (psi) 60 75
5 3 - - + - 2615 1380
6 11 + - + - 2625 1525
7 14 - + + - 2400 1500

6-59
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

8 17 + + + - 2750 1620
9 6 - - - + 2400 1400
10 7 + - - + 2390 1525
11 2 - + - + 2300 1500
12 10 + + - + 2520 1500
13 4 - - + + 2625 1420
14 19 + - + + 2630 1490
15 15 - + + + 2500 1500
16 20 + + + + 2710 1600
17 1 0 0 0 0 2515 1500
18 5 0 0 0 0 2500 1460
19 16 0 0 0 0 2400 1525
20 12 0 0 0 0 2475 1500

(a) Consider only the molecular weight response. Plot the effect estimates on a normal probability scale.
What effects appear important?

DESIGN-EXPERT Plot
Molecular Wt
H a lf N o rm a l p lo t
H a l f N o r m a l % p r o b a b i l i ty

A: Temperature
B: Catalyst Con. 99
C: Time
D: Pressure C
97
95
A
90
AB
85
80

70
60

40

20

0 .0 0 5 0 .3 1 1 0 0 .6 3 1 5 0 .9 4 2 0 1 .2 5

E ffe c t

A, C and the AB interaction appear to be important.

(b) Use an analysis of variance to confirm the results from part (a). Is there an indication of curvature?
A,C and the AB interaction are significant. While the main effect of B is not significant, it could be
included to preserve hierarchy in the model. There is no indication of quadratic curvature.

Design Expert Output

Response: Molecular Wt
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 2.809E+005 3 93620.83 73.00 < 0.0001 significant
A 61256.25 1 61256.25 47.76 < 0.0001
C 1.620E+005 1 1.620E+005 126.32 < 0.0001
AB 57600.00 1 57600.00 44.91 < 0.0001
Curvature 3645.00 1 3645.00 2.84 0.1125 not significant
Residual 19237.50 15 1282.50
Lack of Fit 11412.50 12 951.04 0.36 0.9106 not significant

6-60
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Pure Error 7825.00 3 2608.33

Cor Total 3.037E+005 19

The Model F-value of 73.00 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, AB are significant model terms.

(c) Write down a regression model to predict molecular weight as a function of the important variables.

Design Expert Output

Final Equation in Terms of Coded Factors:

Molecular Wt =
+2506.25
+61.87 *A
+100.63 *C
+60.00 *A*B

(d) Analyze the residuals and comment on model adequacy.

Normal plot of residuals Residuals vs. Predicted

42.5

99

95
10.625
Norm al % probability

90
80 2
Res iduals

70

50 -21.25

30
20
10
-53.125
5

-85

-85 -53.125 -21.25 10.625 42.5 2283.75 2395.00 2506.25 2617.50 2728.75

Res idual Predicted

There are two residuals that appear to be large and should be investigated.

(e) Repeat parts (a) - (d) using the viscosity response.

6-61
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot
Viscosity
N o rm a l p lo t
A: Temperature
B: Catalyst Con.
99

N o r m a l % p r o b a b i l i ty
C: Time A
D: Pressure
95 B
90
80
70

50

30
20
10
5

-2 5 .0 0 5 .3 1 3 5 .6 2 6 5 .9 4 9 6 .2 5

E ffe c t

Design Expert Output

Response: Viscosity
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 70362.50 2 35181.25 35.97 < 0.0001 significant
A 37056.25 1 37056.25 37.88 < 0.0001
B 33306.25 1 33306.25 34.05 < 0.0001
Curvature 61.25 1 61.25 0.063 0.8056 not significant
Residual 15650.00 16 978.13
Lack of Fit 13481.25 13 1037.02 1.43 0.4298 not significant
Pure Error 2168.75 3 722.92
Cor Total 86073.75 19

The Model F-value of 35.97 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B are significant model terms.

Final Equation in Terms of Coded Factors:

Viscosity =
+1500.62
+48.13 *A
+45.63 *B

6-62
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal plot of residuals Residuals vs. Predicted

35.625

99 2

95
3.125 23
Norm al % probability

90
2
80

Res iduals
70

50 -29.375

30
20
10
-61.875
5

-94.375

-94.375 -61.875 -29.375 3.125 35.625 1406.87 1453.75 1500.62 1547.50 1594.37

Res idual Predicted

There is one large residual that should be investigated.

6.32. Continuation of Problem 6.31. Use the regression models for molecular weight and viscosity to
answer the following questions.

(a) Construct a response surface contour plot for molecular weight. In what direction would you adjust
the process variables to increase molecular weight? Increase temperature, catalyst and time.

DESIGN-EXPERT Plot
8 .0 0 2 4 0 0
M o le c ula r W t
Molecular Wt
X = A: Temperature 2600
Y = B: Catalyst Con.
2425
2575
Design Points
B : C a ta l y s t C o n .

7 .0 0
2450
Actual Factors
C: Time = 25.00
D: Pressure = 67.50 2550
2475
4
6 .0 0
2525

2500

5 .0 0

4 .0 0
1 0 0 .0 0 1 0 5 .0 0 1 1 0 .0 0 1 1 5 .0 0 1 2 0 .0 0

A : T e m p e ra tu re

(b) Construct a response surface contour plot for viscosity. In what direction would you adjust the process
variables to decrease viscosity?

6-63
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot
8 .0 0
V is c o s ity
Viscosity
X = A: Temperature 1575
Y = B: Catalyst Con.

Design Points 1550

B : C a ta l y s t C o n .
7 .0 0
Actual Factors
C: Time = 25.00
D: Pressure = 67.50
1525

6 .0 0
1 540 0

1475

5 .0 0
1450

1425

4 .0 0
1 0 0 .0 0 1 0 5 .0 0 1 1 0 .0 0 1 1 5 .0 0 1 2 0 .0 0

A : T e m p e ra tu re

Decrease temperature and catalyst.

(c) What operating conditions would you recommend if it was necessary to produce a product with a
molecular weight between 2400 and 2500, and the lowest possible viscosity?

DESIGN-EXPERT Plot
8 .0 0
O ve rla y P lo t
Overlay Plot
X = A: Temperature
Y = B: Catalyst Con.

Actual Factors
B : C a ta l y s t C o n .

C: Time = 24.50 7 .0 0
D: Pressure = 67.50

6 .0 0

Mo le c u la r W t: 2 5 0 0
5 .0 0
V is c o s ity : 1 4 5 0

4 .0 0
1 0 0 .0 0 1 0 5 .0 0 1 1 0 .0 0 1 1 5 .0 0 1 2 0 .0 0

A : T e m p e ra tu re

Set the temperature between 100 and 105, the catalyst between 4 and 5%, and the time at 24.5 minutes.
The pressure was not significant and can be set at conditions that may improve other results of the process
such as cost.

6.33. Consider the single replicate of the 24 design in Example 6.2. Suppose that we ran five points at the
center (0, 0, 0, 0) and observed the following responses: 73, 75, 71, 69, and 76. Test for curvature in this
experiment. Interpret the results.

6-64
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design Expert Output

Response: Filtration Rate
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 5535.81 5 1107.16 68.01 < 0.0001 significant
A 1870.56 1 1870.56 114.90 < 0.0001
C 390.06 1 390.06 23.96 0.0002
D 855.56 1 855.56 52.55 < 0.0001
AC 1314.06 1 1314.06 80.71 < 0.0001
AD 1105.56 1 1105.56 67.91 < 0.0001
Curvature 28.55 1 28.55 1.75 0.2066 not significant
Residual 227.93 14 16.28
Lack of Fit 195.13 10 19.51 2.38 0.2093 not significant
Pure Error 32.80 4 8.20
Cor Total 5792.29 20

The Model F-value of 68.01 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, D, AC, AD are significant model terms.

The "Curvature F-value" of 1.75 implies the curvature (as measured by difference between the
average of the center points and the average of the factorial points) in the design space is not
significant relative to the noise. There is a 20.66% chance that a "Curvature F-value"
this large could occur due to noise.

There is no indication of curvature.

6.34. A missing value in a 2k factorial. It is not unusual to find that one of the observations in a 2k
design is missing due to faulty measuring equipment, a spoiled test, or some other reason. If the design is
replicated n times (n>1) some of the techniques discussed in Chapter 5 can be employed. However, for an
unreplicated factorial (n-1) some other method must be used. One logical approach is to estimate the
missing value with a number that makes the highest-order interaction contrast zero. Apply this technique to
the experiment in Example 6.2 assuming that run ab is missing. Compare the results with the results of
Example 6.2.

Treatment Response *
Combination Response ABCD ABCD A B C D
(1) 45 45 1 -1 -1 -1 -1
a 71 -71 -1 1 -1 -1 -1
b 48 -48 -1 -1 1 -1 -1
ab missing missing * 1 1 1 1 -1 -1
c 68 -68 -1 -1 -1 1 -1
ac 60 60 1 1 -1 1 -1
bc 80 80 1 -1 1 1 -1
abc 65 -65 -1 1 1 1 -1
d 43 -43 -1 -1 -1 -1 1
ad 100 100 1 1 -1 -1 1
bd 45 45 1 -1 1 -1 1
abd 104 -104 -1 1 1 -1 1
cd 75 75 1 -1 -1 1 1
acd 86 -86 -1 1 -1 1 1
bcd 70 -70 -1 -1 1 1 1
abcd 96 96 1 1 1 1 1

6-65
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Contrast (ABCD) = Missing – 54 = 0

Missing = 54

Substitute the value 54 for the missing run at ab. From the effects list and half normal plot shown below,
factors A, C, D, AC, and AD appear to be large; the same result as found in Example 6.2.

Design Expert Output

Term Effect SumSqr % Contribtn
Model Intercept
Model A 20.25 1640.25 27.5406
Model B 1.75 12.25 0.205684
Model C 11.25 506.25 8.50019
Model D 16 1024 17.1935
Model AB -1.25 6.25 0.104941
Model AC -16.75 1122.25 18.8431
Model AD 18 1296 21.7605
Model BC 3.75 56.25 0.944465
Model BD 1 4 0.067162
Model CD -2.5 25 0.419762
Model ABC 3.25 42.25 0.709398
Model ABD 5.5 121 2.03165
Model ACD -3 36 0.604458
Model BCD -4 64 1.07459
Model ABCD 0 0 0
Lenth's ME 11.5676
Lenth's SME 23.4839

DESIGN-EXPERT Plot
Filtration Rate
N o rm a l p lo t
A: Temperature
B: Pressure 99
N o r m a l % p r o b a b i l i ty

C: Concentration
D: Stirring Rate A
95
90 AD
80 D
70 C

50

30
20
10
5
AC
1

-1 6 .7 5 -7 .5 0 1 .7 5 1 1 .0 0 2 0 .2 5

E ffe c t

6.35. An engineer has performed an experiment to study the effect of four factors on the surface roughness
of a machined part. The factors (and their levels) are A = tool angle (12 degrees, 15 degrees), B = cutting
fluid viscosity (300, 400), C = feed rate (10 in/min, 15 in/min), and D = cutting fluid cooler used (no, yes).
The data from this experiment (with the factors coded to the usual -1, +1 levels) are shown below.

Run A B C D Surface Roughness

1 - - - - 0.00340
2 + - - - 0.00362
3 - + - - 0.00301
4 + + - - 0.00182

6-66
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

5 - - + - 0.00280
6 + - + - 0.00290
7 - + + - 0.00252
8 + + + - 0.00160
9 - - - + 0.00336
10 + - - + 0.00344
11 - + - + 0.00308
12 + + - + 0.00184
13 - - + + 0.00269
14 + - + + 0.00284
15 - + + + 0.00253
16 + + + + 0.00163

(a) Estimate the factor effects. Plot the effect estimates on a normal probability plot and select a tentative
model.

DESIGN-EXPERT Plot
Surface Roughness
N o rm a l p lo t
A: Tool Angle
B: Viscosity
99
N o r m a l % p r o b a b i l i ty

C: Feed Rate
D: Cutting Fluid
95
90
80
70

50

30 A
20 C
AB
10
5 B

-0 .0 0 0 9 -0 .0 0 0 6 -0 .0 0 0 4 -0 .0 0 0 1 0 .0 0 0 1

E ffe c t

(b) Fit the model identified in part (a) and analyze the residuals. Is there any indication of model
inadequacy?

Design Expert Output

Response: Surface Roughness
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 6.406E-006 4 1.601E-006 114.97 < 0.0001 significant
A 8.556E-007 1 8.556E-007 61.43 < 0.0001
B 3.080E-006 1 3.080E-006 221.11 < 0.0001
C 1.030E-006 1 1.030E-006 73.96 < 0.0001
AB 1.440E-006 1 1.440E-006 103.38 < 0.0001
Residual 1.532E-007 11 1.393E-008
Cor Total 6.559E-006 15

The Model F-value of 114.97 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.

6-67
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

In this case A, B, C, AB are significant model terms.

Normal plot of residuals Residuals vs. Predicted

0.00016625

99

95
8.5625E-005
Norm al % probability

90
80

Res iduals
70

50 5E-006

30
20
10
-7.5625E-005
5

-0.00015625

-0.00015625-7.5625E-005 5E-006 8.5625E-005 0.00016625 0.0015 0.0020 0.0025 0.0030 0.0035

Res idual Predicted

Residuals vs. Tool Angle

0.00016625

8.5625E-005
Res iduals

5E-006

-7.5625E-005

-0.00015625

12 13 14 15

Tool Angle

The plot of residuals versus predicted shows a slight “u-shaped” appearance in the residuals, and the plot of
residuals versus tool angle shows an outward-opening funnel.

(c) Repeat the analysis from parts (a) and (b) using 1/y as the response variable. Is there and indication
that the transformation has been useful?

The plots of the residuals are more representative of a model that does not violate the constant variance
assumption.

6-68
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

DESIGN-EXPERT Plot
1.0/(Surface Roughness)
N o rm a l p lo t
A: Tool Angle
B: Viscosity
99

N o r m a l % p r o b a b i l i ty
C: Feed Rate
B
D: Cutting Fluid
95
AB
90 A
80 C
70

50

30
20
10
5

-8 .3 0 3 1 .1 4 7 0 .5 9 1 1 0 .0 4 1 4 9 .4 9

E ffe c t

Design Expert Output

Response: Surface RoughnessTransform:Inverse
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 2.059E+005 4 51472.28 1455.72 < 0.0001 significant
A 42610.92 1 42610.92 1205.11 < 0.0001
B 89386.27 1 89386.27 2527.99 < 0.0001
C 18762.29 1 18762.29 530.63 < 0.0001
AB 55129.62 1 55129.62 1559.16 < 0.0001
Residual 388.94 11 35.36
Cor Total 2.063E+005 15

The Model F-value of 1455.72 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, C, AB are significant model terms.

DESIGN-EXPERT Plot DESIGN-EXPERT Plot

1.0/(Surface Roughness)
R e s id ua ls vs . P re d ic te d 1.0/(Surface Roughness)
R e s id ua ls vs . To o l A ng le
8 .9 7 1 3 8 .9 7 1 3

4 .9 1 4 0 4 4 .9 1 4 0 4
R e s id u a ls

R e s id u a ls

0 .8 5 6 7 9 1 0 .8 5 6 7 9 1

-3 .2 0 0 4 6 -3 .2 0 0 4 6

-7 .2 5 7 7 1 -7 .2 5 7 7 1

2 8 1 .7 3 3 6 5 .5 7 4 4 9 .4 1 5 3 3 .2 6 6 1 7 .1 0 12 13 14 15

P re d ic te d T o o l A n g le

6-69
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(d) Fit a model in terms of the coded variables that can be used to predict the surface roughness. Convert
this prediction equation into a model in the natural variables.

Design Expert Output

Final Equation in Terms of Coded Factors:

1.0/(Surface Roughness) =
+397.81
+51.61 *A
+74.74 *B
+34.24 *C
+58.70 *A*B

6.36. Resistivity on a silicon wafer is influenced by several factors. The results of a 24 factorial
experiment performed during a critical process step is shown below.

Run A B C D Resistivity
1 - - - - 1.92
2 + - - - 11.28
3 - + - - 1.09
4 + + - - 5.75
5 - - + - 2.13
6 + - + - 9.53
7 - + + - 1.03
8 + + + - 5.35
9 - - - + 1.60
10 + - - + 11.73
11 - + - + 1.16
12 + + - + 4.68
13 - - + + 2.16
14 + - + + 9.11
15 - + + + 1.07
16 + + + + 5.30

6-70
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(a) Estimate the factor effects. Plot the effect estimates on a normal probability plot and select a tentative
model.

DESIGN-EXPERT Plot
Resistivity
N o rm a l p lo t
A: A
B: B
99

N o r m a l % p r o b a b i l i ty
C: C
A
D: D
95
90
80
70

50

30
20
AB
10
5 B

-3 .0 0 -0 .6 7 1 .6 6 3 .9 9 6 .3 2

E ffe c t

(b) Fit the model identified in part (a) and analyze the residuals. Is there any indication of model
inadequacy?

The normal probability plot of residuals is not satisfactory. The plots of residual versus predicted, residual
versus factor A, and the residual versus factor B are funnel shaped indicating non-constant variance.

Design Expert Output

Response: Resistivity
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 214.22 3 71.41 148.81 < 0.0001 significant
A 159.83 1 159.83 333.09 < 0.0001
B 36.09 1 36.09 75.21 < 0.0001
AB 18.30 1 18.30 38.13 < 0.0001
Residual 5.76 12 0.48
Cor Total 219.98 15

The Model F-value of 148.81 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, AB are significant model terms.

6-71
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal plot of residuals Residuals vs. Predicted

1.3175

99

95
0.6625
Norm al % probability

90
80

Res iduals
70

50 0.0075

30
20
10
-0.6475
5

-1.3025

-1.3025 -0.6475 0.0075 0.6625 1.3175 1.09 3.42 5.75 8.08 10.41

Res idual Predicted

Residuals vs. A Residuals vs. B

1.3175 1.3175

0.6625 0.6625
Res iduals

Res iduals

0.0075 0.0075

-0.6475 -0.6475

-1.3025 -1.3025

-1 0 1 -1 0 1

A B

6-72
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(c) Repeat the analysis from parts (a) and (b) using ln(y) as the response variable. Is there any indication
that the transformation has been useful?

DESIGN-EXPERT Plot
Ln(Resistivity)
N o rm a l p lo t
A: A
B: B
99

N o r m a l % p r o b a b i l i ty
C: C
A
D: D
95
90
80
70

50

30
20
10
5 B

-0 .6 3 -0 .0 6 0 .5 0 1 .0 6 1 .6 3

E ffe c t

Design Expert Output

Response: Resistivity Transform: Natural log Constant: 0.000
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 12.15 2 6.08 553.44 < 0.0001 significant
A 10.57 1 10.57 962.95 < 0.0001
B 1.58 1 1.58 143.94 < 0.0001
Residual 0.14 13 0.011
Cor Total 12.30 15

The Model F-value of 553.44 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B are significant model terms.

The transformed data no longer indicates that the AB interaction is significant. A simpler model has
resulted from the log transformation.

6-73
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal plot of residuals Residuals vs. Predicted

0.149585

99

95
0.0579833
Norm al % probability

90
80

Res iduals
70

50 -0.033618

30
20
10
-0.125219
5

-0.216821

-0.216821 -0.125219 -0.033618 0.0579833 0.149585 0.06 0.62 1.19 1.75 2.31

Res idual Predicted

Residuals vs. A Residuals vs. B

0.149585 0.149585

0.0579833 0.0579833
Res iduals

Res iduals

-0.033618 -0.033618

-0.125219 -0.125219

-0.216821 -0.216821

-1 0 1 -1 0 1

A B

The residual plots are much improved.

(d) Fit a model in terms of the coded variables that can be used to predict the resistivity.

Design Expert Output

Final Equation in Terms of Coded Factors:

Ln(Resistivity) =
+1.19
+0.81 *A
-0.31 *B

6-74
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.37. Continuation of Problem 6.36. Suppose that the experiment had also run four center points along
with the 16 runs in Problem 6.36. The resistivity measurements at the center points are: 8.15, 7.63, 8.95,
6.48. Analyze the experiment again incorporating the center points. What conclusions can you draw now?

DESIGN-EXPERT Plot
Resistivity
N o rm a l p lo t
A: A
B: B
99

N o r m a l % p r o b a b i l i ty
C: C A
D: D
95
90
80
70

50

30
20
10 AB
5 B

-3 .0 0 -0 .6 7 1 .6 6 3 .9 9 6 .3 2

E ffe c t

Design Expert Output

Response: Resistivity
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 214.22 3 71.41 119.35 < 0.0001 significant
A 159.83 1 159.83 267.14 < 0.0001
B 36.09 1 36.09 60.32 < 0.0001
AB 18.30 1 18.30 30.58 < 0.0001
Curvature 31.19 1 31.19 52.13 < 0.0001 significant
Residual 8.97 15 0.60
Lack of Fit 5.76 12 0.48 0.45 0.8632 not significant
Pure Error 3.22 3 1.07
Cor Total 254.38 19

The Model F-value of 119.35 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B, AB are significant model terms.

6-75
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Normal plot of residuals Residuals vs. Predicted

1.3175

99

95
0.6575
Norm al % probability

90
80

Res iduals
70

50 -0.0025

30
20
10
-0.6625
5

-1.3225

-1.3225 -0.6625 -0.0025 0.6575 1.3175 1.09 3.42 5.75 8.08 10.41

Res idual Predicted

Because of the funnel shaped residual versus predicted plot, the analysis was repeated with the natural log
transformation.

DESIGN-EXPERT Plot
Ln(Resistivity)
N o rm a l p lo t
A: A
B: B
99
N o r m a l % p r o b a b i l i ty

C: C A
D: D
95
90
80
70

50

30
20
10
5 B

-0 .6 3 -0 .0 6 0 .5 0 1 .0 6 1 .6 3

E ffe c t

Design Expert Output

Response: Resistivity Transform: Natural log Constant: 0.000
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 12.15 2 6.08 490.37 < 0.0001 significant
A 10.57 1 10.57 853.20 < 0.0001
B 1.58 1 1.58 127.54 < 0.0001
Curvature 2.38 1 2.38 191.98 < 0.0001 significant
Residual 0.20 16 0.012
Lack of Fit 0.14 13 0.011 0.59 0.7811 not significant
Pure Error 0.056 3 0.019
Cor Total 14.73 19

The Model F-value of 490.37 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

6-76
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, B are significant model terms.

The "Curvature F-value" of 191.98 implies there is significant curvature (as measured by
difference between the average of the center points and the average of the factorial points) in
the design space. There is only a 0.01% chance that a "Curvature F-value" this large
could occur due to noise.

The curvature test indicates that the model has significant pure quadratic curvature.

6.38. The book by Davies (Design and Analysis of Industrial Experiments) describes an experiment to
study the yield of isatin. The factors studies and their levels are as follows:

Factor Low (-) High (+)

A: Acid strength (%) 87 93
B: Reaction time (min) 15 30
C: Amount of acid (ml) 35 45
D: Reaction temperature (°C) 60 70

The data from the 24 factorial is shown in Table P6.11.

Table P6.11
A B C D Yield
-1 -1 -1 -1 6.08
1 -1 -1 -1 6.04
-1 1 -1 -1 6.53
1 1 -1 -1 6.43
-1 -1 1 -1 6.31
1 -1 1 -1 6.09
-1 1 1 -1 6.12
1 1 1 -1 6.36
-1 -1 -1 1 6.79
1 -1 -1 1 6.68
-1 1 -1 1 6.73
1 1 -1 1 6.08
-1 -1 1 1 6.77
1 -1 1 1 6.38
-1 1 1 1 6.49
1 1 1 1 6.23

(a) Fit a main-effects-only model to the data from this experiment. Are any of the main effects
significant?

Temperature appears to be significant.

Design Expert Output

Response 1 Yield
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 0.47 4 0.12 2.26 0.1287 not significant
A-Acid strength 0.15 1 0.15 2.81 0.1221
B-Reaction time 1.806E-003 1 1.806E-003 0.035 0.8558

6-77
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

C-Amount of acid 0.023 1 0.023 0.45 0.5181

D-Temperature 0.30 1 0.30 5.75 0.0354
Residual 0.57 11 0.052
Cor Total 1.04 15

The "Model F-value" of 2.26 implies the model is not significant relative to the noise. There is a
12.87 % chance that a "Model F-value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case D are significant model terms.

(b) Analyze the residuals. Are there any indications of model inadequacy or violation of the
assumptions?

The residual plots appear to be acceptable.

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-0
.4 -0
.3 -0
.2 -0
.1 0 0.1 0.2 0.3

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
0.3

0.2

0.1
Residuals

-0
.1

-0
.2

-0
.3

-0
.4

6.10 6.20 6.30 6.40 6.50 6.60 6.70

Pre
dicted

6-78
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Acidstre
ngth
0.3

0.2

0.1
Residuals

-0
.1

-0
.2

-0
.3

-0
.4

87
.00 88
.00 89
.00 90
.00 91
.00 92
.00 93
.00

Acidstre
ngth

Re
sidua
lsvs. Re
actio
ntime
0.3

0.2

0.1
Residuals

-0
.1

-0
.2

-0
.3

-0
.4

15
.00 18
.00 21
.00 24
.00 27
.00 30
.00

B:Re
a ctiontime

6-79
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Amo
unto
facid
0.3

0.2

0.1
Residuals

-0
.1

-0
.2

-0
.3

-0
.4

35
.00 37
.00 39
.00 41
.00 43
.00 45
.00

C:Amoun
tof a
cid

Re
sidua
lsvs. Re
actio
nte
m pe
rature
0.3

0.2

0.1
Residuals

-0
.1

-0
.2

-0
.3

-0
.4

60
.00 62
.00 64
.00 66
.00 68
.00 70
.00

D:Re
a ctiontemp
era
ture

(c) Find an equation for predicting the yield of isatin over the design space. Express the equation in
both coded and engineering units.

Design Expert Output

Final Equation in Terms of Coded Factors:

Yield =
+6.38
-0.096 *A
-0.011 *B
-0.038 *C
+0.14 *D

Design Expert Output

Final Equation in Terms of Coded Factors:

Yield =
+7.80813

6-80
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

-0.031875 * Acid strength

-1.41667E-003 * Reaction time
-7.62500E-003 * Amount of acid
+0.027375 * Reaction temperature

(d) Is there any indication that adding interactions to the model would improve the results that you
have obtained?

The normal probability of effect shown below suggests that the BD and other interactions may
improve the results. The BD interaction plot shows why the interaction is strong, yet the main
effect B is not significant.

No
rma
lPlo
t

99

95 D
90
ABC
Normal % Probability

80 BCD
70

50

30
20
AD
10
A
5
BD
1

-0
.25 -0
.12 0.01 0.14 0.27

Sta
nda
rdizedEffe
ct

Inte
ractio
n
6.8 D:Re
actionte
m pera
ture
Warn
ing!Termin
volv
e dinm
ultiplein
teraction
s.

D+

6.6
Yield

6.4

6.2

D-

15
.00 18
.00 21
.00 24
.00 27
.00 30
.00

B:Re
actiontime

6-81
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

6.39. An article in Quality and Reliability Engineering International (2010, Vol. 26, pp. 223-233) presents
a 25 factorial design. The experiment is shown in the following table:

A B C D E y
-1 -1 -1 -1 -1 8.11
1 -1 -1 -1 -1 5.56
-1 1 -1 -1 -1 5.77
1 1 -1 -1 -1 5.82
-1 -1 1 -1 -1 9.17
1 -1 1 -1 -1 7.8
-1 1 1 -1 -1 3.23
1 1 1 -1 -1 5.69
-1 -1 -1 1 -1 8.82
1 -1 -1 1 -1 14.23
-1 1 -1 1 -1 9.2
1 1 -1 1 -1 8.94
-1 -1 1 1 -1 8.68
1 -1 1 1 -1 11.49
-1 1 1 1 -1 6.25
1 1 1 1 -1 9.12
-1 -1 -1 -1 1 7.93
1 -1 -1 -1 1 5
-1 1 -1 -1 1 7.47
1 1 -1 -1 1 12
-1 -1 1 -1 1 9.86
1 -1 1 -1 1 3.65
-1 1 1 -1 1 6.4
1 1 1 -1 1 11.61
-1 -1 -1 1 1 12.43
1 -1 -1 1 1 17.55
-1 1 -1 1 1 8.87
1 1 -1 1 1 25.38
-1 -1 1 1 1 13.06
1 -1 1 1 1 18.85
-1 1 1 1 1 11.78
1 1 1 1 1 26.05

(a) Analyze the data from this experiment. Identify the significant factors and interactions.

The half normal plot of effects below identifies the significant factors and interactions.

6-82
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Ha
lf-No
rma
lPlo
t

99

D
Half-Normal % Probability
95
E
90 AD
A
DE
80 BE
AB
ABE
70
ADA
EE

50

30
20
10
0

0.00 1.00 1.99 2.99 3.98 4.98 5.98

|Sta
ndard
ize
dEffe
ct|

Design Expert Output

Response 1 y
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 879.62 11 79.97 49.15 < 0.0001 significant
A-A 83.56 1 83.56 51.36 < 0.0001
B-B 0.060 1 0.060 0.037 0.8492
D-D 285.78 1 285.78 175.66 < 0.0001
E-E 153.17 1 153.17 94.15 < 0.0001
AB 48.93 1 48.93 30.08 < 0.0001
AD 88.88 1 88.88 54.63 < 0.0001
AE 33.76 1 33.76 20.75 0.0002
BE 52.71 1 52.71 32.40 < 0.0001
DE 61.80 1 61.80 37.99 < 0.0001
ABE 44.96 1 44.96 27.64 < 0.0001
ADE 26.01 1 26.01 15.99 0.0007
Residual 32.54 20 1.63
Cor Total 912.16 31

The Model F-value of 49.15 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, D, E, AB, AD, AE, BE, DE, ABE, ADE are significant model terms.

(b) Analyze the residuals from this experiment. Are there any indications of model inadequacy or
violations of the assumptions?

The residual plots below do not identify any concerns with model adequacy or the violations of
the assumptions.

6-83
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-3 -2 -1 0 1 2 3

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
3

1
Residuals

-1

-2

-3

0.00 5.00 10
.00 15
.00 20
.00 25
.00 30
.00

Pre
dicted

6-84
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. A
3

1
Residuals

-1

-2

-3

-1
.00 -0
.50 0.00 0.50 1.00

Re
sidua
lsvs. B
3

1
Residuals

-1

-2

-3

-1
.00 -0
.50 0.00 0.50 1.00

B:B

6-85
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. C
3

1
Residuals

-1

-2

-3

-1
.00 -0
.50 0.00 0.50 1.00

C:C

Re
sidua
lsvs. D
3

1
Residuals

-1

-2

-3

-1
.00 -0
.50 0.00 0.50 1.00

D:D

6-86
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. E
3

1
Residuals

-1

-2

-3

-1
.00 -0
.50 0.00 0.50 1.00

E:E

(c) One of the factors from this experiment does not seem to be important. If you drop this factor,
what type of design remains? Analyze the data using the full factorial model for only the four
active factors. Compare your results with those obtained in part (a).

The resulting experimental design is a replicated 24 full factorial design. The ANOVA is shown
below. The factor names in the output below were modified to match the factor names in the
original problem. The same factors are significant below as were significant in the original
analysis.

Design Expert Output

Response 1 y
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 888.80 15 59.25 40.58 < 0.0001 significant
A-A 83.56 1 83.56 57.23 < 0.0001
B-B 0.060 1 0.060 0.041 0.8414
D-D 285.78 1 285.78 195.74 < 0.0001
E-E 153.17 1 153.17 104.91 < 0.0001
AB 48.93 1 48.93 33.51 < 0.0001
AD 88.88 1 88.88 60.88 < 0.0001
AE 33.76 1 33.76 23.13 0.0002
BD 5.778E-003 1 5.778E-003 3.958E-003 0.9506
BE 52.71 1 52.71 36.10 < 0.0001
DE 61.80 1 61.80 42.33 < 0.0001
ABD 3.82 1 3.82 2.61 0.1255
ABE 44.96 1 44.96 30.79 < 0.0001
ADE 26.01 1 26.01 17.82 0.0006
BDE 0.050 1 0.050 0.035 0.8549
ABDE 5.31 1 5.31 3.63 0.0747
Pure Error 23.36 16 1.46
Cor Total 912.16 31

The Model F-value of 40.58 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, D, AB, AC, AD, BD, CD, ABD, ACD are significant model terms.

6-87
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(d) Find the settings of the active factors that maximize the predicted response.

The cube plot below, with factors E set at +1 and C set at 0, identifies the maximum predicted
response with the remaining factors, A, B, and D all set at +1.
Design-Expert® Software
Factor Coding: Actual Cube
y y
X1 = A: A
X2 = B: B
X3 = D: D 10.44 25.7063
Actual Factors
C: C = 0.00
E: E = 1.00

B+: 1.00 6.82 11.8138

B: B

12.63 18.2088 D+: 1.00

D: D

B-: -1.00 9.01 4.31625 D-: -1.00

A-: -1.00 A+: 1.00
A: A

6.40. A paper in the Journal of Chemical Technology and Biotechnology (“Response Surface
Optimization of the Critical Media Components for the Production of Surfactin,” 1997, Vol. 68, pp. 263-
270) describes the use of a designed experiment to maximize surfactin production. A portion of the data
from this experiment is shown in the table below. Surfactin was assayed by an indirect method, which
involves measurement of surface tension of the diluted broth samples. Relative surfactin concentrations
were determined by serially diluting the broth until the critical micelle concentration (CMC) was reached.
The dilution at which the surface tension starts rising abruptly was denoted by CMC-1 and was considered
proportional to the amount of surfactant present in the original sample.

Glucose NH 4 NO 3 FeSO 4 MnSO 4 y

Run (g dm-3) (g dm-3) (g dm-3 x 10-4) (g dm-3 x 10-2) (CMC)-1
1 20.00 2.00 6.00 4.00 23
2 60.00 2.00 6.00 4.00 15
3 20.00 6.00 6.00 4.00 16
4 60.00 6.00 6.00 4.00 18
5 20.00 2.00 30.00 4.00 25
6 60.00 2.00 30.00 4.00 16
7 20.00 6.00 30.00 4.00 17
8 60.00 6.00 30.00 4.00 26
9 20.00 2.00 6.00 20.00 28
10 60.00 2.00 6.00 20.00 16
11 20.00 6.00 6.00 20.00 18
12 60.00 6.00 6.00 20.00 21
13 20.00 2.00 30.00 20.00 36
14 60.00 2.00 30.00 20.00 24
15 20.00 6.00 30.00 20.00 33
16 60.00 6.00 30.00 20.00 34

6-88
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

(a) Analyze the data from this experiment. Identify the significant factors and interactions.

The half normal probability plot of effects, followed by the ANOVA, identify the significant
factors and interactions. Although factor B is not significant, the AB interaction is.

Ha
lf-No
rma
lPlo
t

99
Half-Normal % Probability

AB
95

90
C

80
D
CD
70
A

50

30
20
10
0

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00

|Sta
ndard
ize
dEffe
ct|

Design Expert Output

Response 1 y
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 680.50 6 113.42 20.73 < 0.0001 significant
A-Glucose 42.25 1 42.25 7.72 0.0214
B-NH4NO3 0.000 1 0.000 0.000 1.0000
C-FeSO4 196.00 1 196.00 35.82 0.0002
D-MnSO4 182.25 1 182.25 33.30 0.0003
AB 196.00 1 196.00 35.82 0.0002
CD 64.00 1 64.00 11.70 0.0076
Residual 49.25 9 5.47
Cor Total 729.75 15

The Model F-value of 20.73 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, C, D, AB, CD are significant model terms.

(b) Analyze the residuals from this experiment. Are there any indications of model inadequacy or
violations of the assumptions?

The residual plots below do not identify any concerns with model adequacy or the violations of
the assumptions.

6-89
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-3 -2 -1 0 1 2 3 4

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
4

2
Residuals

-1

-2

-3

10
.00 15
.00 20
.00 25
.00 30
.00 35
.00 40
.00

Pre
dicted

6-90
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Gluco
se
4

2
Residuals

2
0
2

2
-1

-2

-3

20
.00 28
.00 36
.00 44
.00 52
.00 60
.00

Glu
cose

Re
sidua
lsvs. NH4NO3
4

2
3

2
2
Residuals

-1

-2

-3

2.00 3.00 4.00 5.00 6.00

B:NH4
N O3

6-91
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Fe
S O4
4

2
3

2
2
Residuals

0
2

-1

-2

-3

6.00 12
.00 18
.00 24
.00 30
.00

C:F
eSO4

Re
sidua
lsvs. M
nSO4
4

2
Residuals

2
0
2

2
-1

-2

-3

4.00 8.00 12
.00 16
.00 20
.00

D:Mn
S O4

(c) What conditions would optimize the surfactin production?

The response, y, is maximized when factor A is 20, B is 2, C is 30, and D is 20.

6-92
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Design-Expert® Software
Factor Coding: Actual Cube
y y
X1 = A: Glucose
X2 = B: NH4NO3
X3 = C: FeSO4 29.875 33.625
Actual Factor
D: MnSO4 = 20.00

B+: 6.00 18.875 22.625

B: NH4NO3
36.875 26.625 C+: 30.00

C: FeSO4

B-: 2.00 25.875 15.625 C-: 6.00

A-: 20.00 A+: 60.00
A: Glucose

6.41. Continuation of Problem 6.40. The experiment in Problem 6.40 actually included six center
points. The responses at these conditions were 35, 35, 35, 36, 36, and 34. Is there any indication of
curvature in the response function? Are additional experiments necessary? What would you recommend
doing now?

Curvature appears to be very significant with a p value less than 0.0001. Axial runs, along with additional
center point runs to identify blocking effects, should be run.

Design Expert Output

Response 1 y
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 713.00 8 89.12 54.61 < 0.0001 significant
A-Glucose 42.25 1 42.25 25.89 0.0003
B-NH4NO3 0.000 1 0.000 0.000 1.0000
C-FeSO4 196.00 1 196.00 120.10 < 0.0001
D-MnSO4 182.25 1 182.25 111.68 < 0.0001
AB 196.00 1 196.00 120.10 < 0.0001
AD 12.25 1 12.25 7.51 0.0179
BC 20.25 1 20.25 12.41 0.0042
CD 64.00 1 64.00 39.22 < 0.0001
Curvature 659.28 1 659.28 403.98 < 0.0001 significant
Residual 19.58 12 1.63
Lack of Fit 16.75 7 2.39 4.22 0.0660 not significant
Pure Error 2.83 5 0.57
Cor Total 1391.86 21

6.42. An article in the Journal of Hazardous Materials (“Feasibility of Using Natural Fishbone Apatite as
a Substitute for Hydroxyapatite in Remediating Aqueous Heavy Metals,” Vol. 69, Issue 2, 1999, pp. 187-
197) describes an experiment to study the suitability of fishbone, a natural, apatite rich substance, as a
substitute for hydroxyapatite in the sequestering of aqueous divalent heavy metal ions. Direct comparison
of hydroxyapatite and fishbone apatite was performed using a three-factor two-level full factorial design.
Apatite (30 or 60 mg) was added to 100mL deionized water and gently agitated overnight in a shaker. The
pH was then adjusted to 5 or 7 using nitric acid. Sufficient concentration of lead nitrate solution was added

6-93
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

to each flask to result in a final volume of 200 mL and a lead concentration of 0.483 or 2.41 mM,
respectively. The experiment was a 23 replicated twice and it was performed for both fishbone and
synthetic apatite. Results are shown below:

Fishbone Hydroxyapatite
Apatite pH Pb Pb,mM pH Pb,mM pH
+ + + 1.82 5.22 0.11 3.49
+ + + 1.81 5.12 0.12 3.46
+ + - 0.01 6.84 0.00 5.84
+ + - 0.00 6.61 0.00 5.90
+ - + 1.11 3.35 0.80 2.70
+ - + 1.04 3.34 0.76 2.74
+ - - 0.00 5.77 0.03 3.36
+ - - 0.01 6.25 0.05 3.24
- + + 2.11 5.29 1.03 3.22
- + + 2.18 5.06 1.05 3.22
- + - 0.03 5.93 0.00 5.53
- + - 0.05 6.02 0.00 5.43
- - + 1.70 3.39 1.34 2.82
- - + 1.69 3.34 1.26 2.79
- - - 0.05 4.50 0.06 3.28
- - - 0.05 4.74 0.07 3.28

(a) Analyze the lead response for fishbone apatite. What factors are important?

As shown below, all main effects and interactions are significant.

Design Expert Output

Response 1 Fishbone Pb
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 12.19 7 1.74 2629.41 < 0.0001 significant
A-Apatite 0.27 1 0.27 400.34 < 0.0001
B-pH 0.35 1 0.35 525.43 < 0.0001
C-Pb 10.99 1 10.99 16587.51 < 0.0001
AB 0.023 1 0.023 33.96 0.0004
AC 0.19 1 0.19 285.62 < 0.0001
BC 0.36 1 0.36 543.40 < 0.0001
ABC 0.020 1 0.020 29.58 0.0006
Pure Error 5.300E-003 8 6.625E-004
Cor Total 12.20 15

Coefficient Standard 95% CI 95% CI

Factor Estimate df Error Low High VIF
Intercept 0.85 1 6.435E-003 0.84 0.87
A-Apatite -0.13 1 6.435E-003 -0.14 -0.11 1.00
B-pH 0.15 1 6.435E-003 0.13 0.16 1.00
C-Pb 0.83 1 6.435E-003 0.81 0.84 1.00
AB 0.038 1 6.435E-003 0.023 0.052 1.00
AC -0.11 1 6.435E-003 -0.12 -0.094 1.00
BC 0.15 1 6.435E-003 0.14 0.16 1.00
ABC 0.035 1 6.435E-003 0.020 0.050 1.00

(b) Analyze the residuals from this response and comment on model adequacy.

6-94
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

The normal plot identifies slightly thicker tails in the distribution of the residuals. The plots of
residuals vs. predicted and residuals vs. the Pb effect identify nonconstant variance.

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-0
.04 -0
.02 0 0.02 0.04

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
0.04

0.02
Residuals

2
0 2
2

-0
.02

-0
.04

0.00 0.50 1.00 1.50 2.00 2.50

Pre
dicted

6-95
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Apa
tite
0.04

0.02
Residuals

3
0 2
3

-0
.02

-0
.04

30
.00 36
.00 42
.00 48
.00 54
.00 60
.00

Apatite

Re
sidua
lsvs. pH
0.04

0.02
Residuals

2 2
0 2
2 2

-0
.02

-0
.04

5.00 5.50 6.00 6.50 7.00

B:p
H

6-96
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Pb
0.04

0.02
Residuals

2 2
0 2
2 2

-0
.02

2
-0
.04

0.48 0.96 1.45 1.93 2.41

C:Pb

(c) Analyze the pH response for fishbone apatite. What factors are important?

The AB and ABC interactions are only moderately significant; all other main effects and
interactions are significant.

Design Expert Output

Response 1 Fishbone pH
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 21.09 7 3.01 116.29 < 0.0001 significant
A-Apatite 1.12 1 1.12 43.17 0.0002
B-pH 8.14 1 8.14 314.08 < 0.0001
C-Pb 9.84 1 9.84 379.98 < 0.0001
AB 0.098 1 0.098 3.77 0.0881
AC 1.17 1 1.17 45.23 0.0001
BC 0.61 1 0.61 23.64 0.0013
ABC 0.11 1 0.11 4.14 0.0763
Pure Error 0.21 8 0.026
Cor Total 21.30 15

Coefficient Standard 95% CI 95% CI

Factor Estimate df Error Low High VIF
Intercept 5.05 1 0.040 4.96 5.14
A-Apatite 0.26 1 0.040 0.17 0.36 1.00
B-pH 0.71 1 0.040 0.62 0.81 1.00
C-Pb -0.78 1 0.040 -0.88 -0.69 1.00
AB -0.078 1 0.040 -0.17 0.015 1.00
AC -0.27 1 0.040 -0.36 -0.18 1.00
BC 0.20 1 0.040 0.10 0.29 1.00
ABC 0.082 1 0.040 -0.011 0.17 1.00

(d) Analyze the residuals from this response and comment on model adequacy.

Although the normal probability plot is acceptable, the plots of residuals vs. predicted identifies
nonconstant variance.

6-97
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-0
.3 -0
.2 -0
.1 0 0.1 0.2 0.3

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
0.3

0.2

0.1
Residuals

-0
.1

-0
.2

-0
.3

3.00 4.00 5.00 6.00 7.00

Pre
dicted

6-98
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Apa
tite
0.3

0.2

0.1
Residuals

-0
.1

-0
.2

-0
.3

30
.00 36
.00 42
.00 48
.00 54
.00 60
.00

Apatite

Re
sidua
lsvs. pH
0.3

0.2

2
0.1
Residuals

-0
.1
2

-0
.2

-0
.3

5.00 5.50 6.00 6.50 7.00

B:p
H

6-99
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Pb
0.3

0.2

0.1
Residuals

-0
.1

-0
.2

-0
.3

0.48 0.96 1.45 1.93 2.41

C:Pb

(e) Analyze the lead response for hydroxyapatite apatite. What factors are important?

As shown below, all main effects and interactions are significant.

Design Expert Output

Response 1 Hydroxyapatite Pb
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 4.01 7 0.57 1018.21 < 0.0001 significant
A-Apatite 0.54 1 0.54 960.40 < 0.0001
B-pH 0.27 1 0.27 471.51 < 0.0001
C-Pb 2.45 1 2.45 4354.18 < 0.0001
AB 0.036 1 0.036 64.18 < 0.0001
AC 0.50 1 0.50 896.18 < 0.0001
BC 0.17 1 0.17 298.84 < 0.0001
ABC 0.046 1 0.046 82.18 < 0.0001
Pure Error 4.500E-003 8 5.625E-004
Cor Total 4.01 15

Coefficient Standard 95% CI 95% CI

Factor Estimate df Error Low High VIF
Intercept 0.42 1 5.929E-003 0.40 0.43
A-Apatite -0.18 1 5.929E-003 -0.20 -0.17 1.00
B-pH -0.13 1 5.929E-003 -0.14 -0.12 1.00
C-Pb 0.39 1 5.929E-003 0.38 0.40 1.00
AB -0.048 1 5.929E-003 -0.061 -0.034 1.00
AC -0.18 1 5.929E-003 -0.19 -0.16 1.00
BC -0.10 1 5.929E-003 -0.12 -0.089 1.00
ABC -0.054 1 5.929E-003 -0.067 -0.040 1.00

(f) Analyze the residuals from this response and comment on model adequacy.

The normal plot identifies slightly thicker tails in the distribution of the residuals. The plots of
residuals vs. predicted and residuals vs. the effects identifies nonconstant variance.

6-100
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-0
.04 -0
.02 0 0.02 0.04 0.06

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
0.06

0.04

0.02
Residuals

0 4

-0
.02

-0
.04

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

Pre
dicted

6-101
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Apa
tite
0.06

0.04

0.02
Residuals

0 2 2

-0
.02

-0
.04

30
.00 36
.00 42
.00 48
.00 54
.00 60
.00

Apatite

Re
sidua
lsvs. pH
0.06

0.04

0.02
Residuals

0 4

-0
.02

-0
.04

5.00 5.50 6.00 6.50 7.00

B:p
H

6-102
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Pb
0.06

0.04

0.02
Residuals

0 4

-0
.02

-0
.04

0.48 0.96 1.45 1.93 2.41

C:Pb

(g) Analyze the pH response for hydroxyapatite apatite. What factors are important?

The ABC interaction is not significant; all of the main effects and two factor interactions are significant.

Design Expert Output

Response 1 Hydroxyapatite pH
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 20.44 7 2.92 1487.66 < 0.0001 significant
A-Apatite 0.084 1 0.084 42.85 0.0002
B-pH 8.82 1 8.82 4494.73 < 0.0001
C-Pb 8.15 1 8.15 4153.39 < 0.0001
AB 0.13 1 0.13 64.22 < 0.0001
AC 0.014 1 0.014 7.34 0.0267
BC 3.24 1 3.24 1650.96 < 0.0001
ABC 2.250E-004 1 2.250E-004 0.11 0.7436
Pure Error 0.016 8 1.963E-003
Cor Total 20.45 15

Coefficient Standard 95% CI 95% CI

Factor Estimate df Error Low High VIF
Intercept 3.77 1 0.011 3.74 3.79
A-Apatite 0.072 1 0.011 0.047 0.098 1.00
B-pH 0.74 1 0.011 0.72 0.77 1.00
C-Pb -0.71 1 0.011 -0.74 -0.69 1.00
AB 0.089 1 0.011 0.063 0.11 1.00
AC -0.030 1 0.011 -0.056 -4.461E-003 1.00
BC -0.45 1 0.011 -0.48 -0.42 1.00
ABC -3.750E-003 1 0.011 -0.029 0.022 1.00

(h) Analyze the residuals from this response and comment on model adequacy.

The only potential concern with the residual plots is the nonconstant variance shown in the plot of
residuals vs. pH.

6-103
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-0
.06 -0
.04 -0
.02 0 0.02 0.04 0.06 0.08

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
0.08

0.06

0.04
Residuals

0.02

0 22

-0
.02

-0
.04

-0
.06

2.00 3.00 4.00 5.00 6.00

Pre
dicted

6-104
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Apa
tite
0.08

0.06

0.04
Residuals

0.02

0 4

-0
.02

-0
.04

-0
.06

30
.00 36
.00 42
.00 48
.00 54
.00 60
.00

Apatite

Re
sidua
lsvs. pH
0.08

0.06

0.04
Residuals

0.02

0 2 2

-0
.02

-0
.04

-0
.06

5.00 5.50 6.00 6.50 7.00

B:p
H

6-105
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Pb
0.08

0.06

0.04
Residuals

0.02
2

0 2 2

2
-0
.02

-0
.04

-0
.06

0.48 0.96 1.45 1.93 2.41

C:Pb

(i) What differences do you see between fishbone and hydroxyapatite apatite? The authors of this
paper concluded that fishbone apatite was comparable to hydroxyapatite apatite. Because the
fishbone apatite is cheaper, it was recommended for adoption. Do you agree with these
conclusions?

The authors of the journal article did not show their analysis for this experiment. When
comparing the Fishbone and Hydroxyapatite models main effects and interactions for the Pb and
pH responses, we might disagree with the authors.

A more effective approach to understand the differences between Fishbone and Hydroxyapatite
would be to include this as a factor in the experimental design. The modified table is shown
below followed by the analysis.

Apatite pH Pb Type Pb,mM pH

+ + + Fishbone 1.82 5.22
+ + + Fishbone 1.81 5.12
+ + - Fishbone 0.01 6.84
+ + - Fishbone 0.00 6.61
+ - + Fishbone 1.11 3.35
+ - + Fishbone 1.04 3.34
+ - - Fishbone 0.00 5.77
+ - - Fishbone 0.01 6.25
- + + Fishbone 2.11 5.29
- + + Fishbone 2.18 5.06
- + - Fishbone 0.03 5.93
- + - Fishbone 0.05 6.02
- - + Fishbone 1.70 3.39
- - + Fishbone 1.69 3.34
- - - Fishbone 0.05 4.50
- - - Fishbone 0.05 4.74
+ + + Hydroxyapatite 0.11 3.49
+ + + Hydroxyapatite 0.12 3.46

6-106
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

+ + - Hydroxyapatite 0.00 5.84

+ + - Hydroxyapatite 0.00 5.90
+ - + Hydroxyapatite 0.80 2.70
+ - + Hydroxyapatite 0.76 2.74
+ - - Hydroxyapatite 0.03 3.36
+ - - Hydroxyapatite 0.05 3.24
- + + Hydroxyapatite 1.03 3.22
- + + Hydroxyapatite 1.05 3.22
- + - Hydroxyapatite 0.00 5.53
- + - Hydroxyapatite 0.00 5.43
- - + Hydroxyapatite 1.34 2.82
- - + Hydroxyapatite 1.26 2.79
- - - Hydroxyapatite 0.06 3.28
- - - Hydroxyapatite 0.07 3.28

The ANOVA below identifies factor D, the type of apatite, as being very significant.

Design Expert Output

Response 1 Pb Response
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 17.73 15 1.18 1929.32 < 0.0001 significant
A-Apatite 0.78 1 0.78 1275.51 < 0.0001
B-pH 2.813E-003 1 2.813E-003 4.59 0.0478
C-Pb 11.91 1 11.91 19440.33 < 0.0001
D-Type 1.52 1 1.52 2485.73 < 0.0001
AB 8.000E-004 1 8.000E-004 1.31 0.2699
AC 0.66 1 0.66 1070.22 < 0.0001
AD 0.024 1 0.024 39.51 < 0.0001
BC 0.018 1 0.018 29.47 < 0.0001
BD 0.61 1 0.61 996.76 < 0.0001
CD 1.53 1 1.53 2500.00 < 0.0001
ABC 2.813E-003 1 2.813E-003 4.59 0.0478
ABD 0.058 1 0.058 94.37 < 0.0001
ACD 0.038 1 0.038 61.73 < 0.0001
BCD 0.51 1 0.51 832.73 < 0.0001
ABCD 0.063 1 0.063 102.88 < 0.0001
Pure Error 9.800E-003 16 6.125E-004
Cor Total 17.74 31

The residual plots below identify concerns, so a power transformation with lambda of 0.7 was applied to
the Pb response.

6-107
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-0
.06 -0
.04 -0
.02 0 0.02 0.04

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
0.04

0.02

2
2
0 42
Residuals

2
2

-0
.02

-0
.04

-0
.06

0.00 0.50 1.00 1.50 2.00 2.50

Pre
dicted

6-108
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Apa
tite
0.04

0.02

2
2 4
0 4 2
Residuals

2 4
2

-0
.02

-0
.04

-0
.06

30
.00 36
.00 42
.00 48
.00 54
.00 60
.00

Apatite

Re
sidua
lsvs. pH
0.04

0.02

2
3 3
0 2 4
Residuals

3 3
2

-0
.02

-0
.04

-0
.06

5.00 5.50 6.00 6.50 7.00

B:p
H

6-109
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Pb
0.04
2

0.02

2
3 3
0 6
Residuals

3 3
2

-0
.02

2
-0
.04

-0
.06

0.48 0.96 1.45 1.93 2.41

C:Pb

Re
sidua
lsvs. Type
0.04
2

0.02

2
4 2
0 2 4
Residuals

4 2
2

-0
.02

2
-0
.04

-0
.06

1.00 1.20 1.40 1.60 1.80 2.00

D:Type

Design Expert Output

Response 1 Pb Response
Transform: Power Lambda: 0.7 Constant: 0
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 12.10 15 0.81 1844.15 < 0.0001 significant
A-Apatite 0.49 1 0.49 1123.40 < 0.0001
B-pH 0.014 1 0.014 32.86 < 0.0001
C-Pb 9.02 1 9.02 20612.51 < 0.0001
D-Type 0.75 1 0.75 1712.43 < 0.0001
AB 3.973E-003 1 3.973E-003 9.08 0.0082
AC 0.29 1 0.29 661.73 < 0.0001
AD 0.024 1 0.024 55.70 < 0.0001
BC 5.077E-003 1 5.077E-003 11.61 0.0036
BD 0.38 1 0.38 877.65 < 0.0001
CD 0.73 1 0.73 1670.33 < 0.0001
ABC 0.011 1 0.011 25.97 0.0001

6-110
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

ABD 0.049 1 0.049 112.97 < 0.0001

ACD 0.067 1 0.067 152.27 < 0.0001
BCD 0.21 1 0.21 472.40 < 0.0001
ABCD 0.057 1 0.057 131.40 < 0.0001
Pure Error 6.999E-003 16 4.374E-004
Cor Total 12.11 31

There are no concerns with the residual plots shown below.

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-0
.03 -0
.02 -0
.01 0 0.01 0.02 0.03

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
0.03

0.02 2
2

0.01
Residuals

0 4 2

-0
.01

2
-0
.02 2

-0
.03

0.00 0.50 1.00 1.50 2.00

Pre
dicted

6-111
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Apa
tite
0.03

0.02 2

0.01
Residuals

0 4 2

-0
.01

-0
.02 2

-0
.03

30
.00 36
.00 42
.00 48
.00 54
.00 60
.00

Apatite

Re
sidua
lsvs. pH
0.03

0.02

0.01
Residuals

0 2 4

-0
.01

-0
.02

-0
.03

5.00 5.50 6.00 6.50 7.00

B:p
H

6-112
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Pb
0.03

0.02 2
2

0.01
Residuals

0 6

-0
.01

2
-0
.02 2

-0
.03

0.48 0.96 1.45 1.93 2.41

C:Pb

Re
sidua
lsvs. Type
0.03

0.02 2

0.01
Residuals

0 2 4

-0
.01

-0
.02 2

-0
.03

1.00 1.20 1.40 1.60 1.80 2.00

D:Type

The analysis for the pH response is shown below.

Design Expert Output

Response 2 pH Response
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 54.62 15 3.64 261.32 < 0.0001 significant
A-Apatite 0.91 1 0.91 65.15 < 0.0001
B-pH 16.95 1 16.95 1216.47 < 0.0001
C-Pb 17.96 1 17.96 1288.54 < 0.0001
D-Type 13.09 1 13.09 939.72 < 0.0001
AB 9.031E-004 1 9.031E-004 0.065 0.8023
AC 0.72 1 0.72 51.89 < 0.0001
AD 0.29 1 0.29 21.14 0.0003
BC 0.52 1 0.52 37.15 < 0.0001
BD 6.903E-003 1 6.903E-003 0.50 0.4916
CD 0.040 1 0.040 2.86 0.1100

6-113
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

ABC 0.049 1 0.049 3.50 0.0796

ABD 0.22 1 0.22 15.99 0.0010
ACD 0.46 1 0.46 33.24 < 0.0001
BCD 3.33 1 3.33 239.31 < 0.0001
ABCD 0.059 1 0.059 4.21 0.0569
Pure Error 0.22 16 0.014
Cor Total 54.84 31

Nonconstant variance is identified in the residual plots below, so an inverse square root transformation
was applied.

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-0
.3 -0
.2 -0
.1 0 0.1 0.2 0.3

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
0.3

0.2

0.1
Residuals

0 22

-0
.1

-0
.2

-0
.3

2.00 3.00 4.00 5.00 6.00 7.00

Pre
dicted

6-114
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Apa
tite
0.3

0.2

0.1
Residuals

0 4

-0
.1

-0
.2

-0
.3

30
.00 36
.00 42
.00 48
.00 54
.00 60
.00

Apatite

Re
sidua
lsvs. pH
0.3

0.2

2
0.1
Residuals

0 2 2

-0
.1
2

-0
.2

-0
.3

5.00 5.50 6.00 6.50 7.00

B:p
H

6-115
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Pb
0.3

0.2

0.1
Residuals

2
0 2 2
2

-0
.1

-0
.2

-0
.3

0.48 0.96 1.45 1.93 2.41

C:Pb

Re
sidua
lsvs. Pb
0.3

0.2

0.1
Residuals

2
0 2 2
2

-0
.1

-0
.2

-0
.3

0.48 0.96 1.45 1.93 2.41

C:Pb

Design Expert Output

Response 2 pH Response
Transform: Inverse Sqrt Constant: 0
ANOVA for selected factorial model
Analysis of variance table [Partial sum of squares - Type III]
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 0.17 15 0.011 483.43 < 0.0001 significant
A-Apatite 1.418E-003 1 1.418E-003 60.44 < 0.0001
B-pH 0.055 1 0.055 2346.06 < 0.0001
C-Pb 0.054 1 0.054 2303.65 < 0.0001
D-Type 0.045 1 0.045 1918.39 < 0.0001
AB 1.417E-005 1 1.417E-005 0.60 0.4484
AC 9.441E-004 1 9.441E-004 40.23 < 0.0001
AD 3.287E-004 1 3.287E-004 14.00 0.0018
BC 4.687E-005 1 4.687E-005 2.00 0.1768
BD 8.263E-004 1 8.263E-004 35.21 < 0.0001
CD 3.302E-004 1 3.302E-004 14.07 0.0017
ABC 3.443E-004 1 3.443E-004 14.67 0.0015

6-116
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

ABD 7.072E-004 1 7.072E-004 30.14 < 0.0001

ACD 7.593E-004 1 7.593E-004 32.36 < 0.0001
BCD 0.010 1 0.010 437.95 < 0.0001
ABCD 4.034E-005 1 4.034E-005 1.72 0.2083
Pure Error 3.755E-004 16 2.347E-005
Cor Total 0.17 31

There are no concerns with the residual plots shown below.

No
rma
lPlo
tofRe
sidua
ls

99

95

90
Normal % Probability

80
70

50

30
20

10

-1
.00
E -0
2 -5
.00
E -0
3 0.00E+00 5.00E-03 1.00E-02

Re
sid
uals

Re
sidua
lsvs. Pre
dicte
d
1.00E-02

5.00E-03
Residuals

0.00E+00 22

-5
.00
E -0
3

-1
.00
E -0
2

0.35 0.40 0.45 0.50 0.55 0.60 0.65

Pre
dicted

6-117
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Apa
tite
1.00E-02

5.00E-03
Residuals

0.00E+00 4

-5
.00
E -0
3

-1
.00
E -0
2

30
.00 36
.00 42
.00 48
.00 54
.00 60
.00

Apatite

Re
sidua
lsvs. pH
1.00E-02

5.00E-03
Residuals

0.00E+00 2 2

-5
.00
E -0
3

-1
.00
E -0
2

5.00 5.50 6.00 6.50 7.00

B:p
H

6-118
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Re
sidua
lsvs. Pb
1.00E-02

5.00E-03
Residuals

0.00E+00 2 2

-5
.00
E -0
3

-1
.00
E -0
2

0.48 0.96 1.45 1.93 2.41

C:Pb

Re
sidua
lsvs. Type
1.00E-02

5.00E-03
Residuals

0.00E+00 4

-5
.00
E -0
3

-1
.00
E -0
2

1.00 1.20 1.40 1.60 1.80 2.00

D:Type

In summary, there is a difference between the Fishbone Apatite and the synthetic Hydroxyapatite.

6.43. Often the fitted regression model from a 2k factorial design is used to make predictions at points of
interest in the design space.

(a) Find the variance of the predicted response öy at the point x1 , x 2 ,…, x k in the design space. Hint:
Remember that the x’s are coded variables, and assume a 2k design with an equal number of replicates
σ2
n at each design point so that the variance of a regression coefficient βö is and that the covariance
n2 k
between any pair of regression coefficients is zero.

Let’s assume that the model can be written as follows:

yˆ (x)=βˆ0 + βˆ1 x1 + βˆ2 x2 + ... + βˆ p x p

6-119
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

where x′ = [ x1 , x2 ,..., xk ] are the values of the original variables in the design at the point of interest
where a prediction is required, and the variables in the model x1 , x2 ,..., x p potentially include interaction
terms among the original k variables. Now the variance of the predicted response is

)] V ( βˆ0 + βˆ1 x1 + βˆ2 x2 + ... + βˆ p x p )

V [ yˆ (x=
= V ( βˆ0 ) + V ( βˆ1 x1 ) + V ( βˆ2 x2 ) + ... + V ( βˆ p x p )
σ2  p

= k 
1 + ∑
n 2  i =1 
xi2 

This result follows because the design is orthogonal and all model parameter estimates have the same
variance. Remember that some of the x’s involved in this equation are potentially interaction terms.

(b) Use the result of part (a) to find an equation for a 100(1-α)% confidence interval on the true mean
response at the point x1 , x 2 ,…, x k in the design space.

The confidence interval is

yˆ (x) − tα /2,df E V [ yˆ (x)] ≤ y (x) ≤ yˆ (x) + tα /2,df E V [ yˆ (x)]

where df E is the number of degrees of freedom used to estimate σ and the estimate of σ 2 has been used
2

in computing the variance of the predicted value of the response at the point of interest.

6.44. Hierarchical Models. Several times we have utilized the hierarchy principal in selecting a model;
that is, we have included non-significant terms in a model because they were factors involved in significant
higher-order terms. Hierarchy is certainly not an absolute principle that must be followed in all cases. To
illustrate, consider the model resulting from Problem 6.1, which required that a non-significant main effect
be included to achieve hierarchy. Using the data from Problem 6.1:

(a) Fit both the hierarchical model and the non-hierarchical model.

Design Expert Output for Hierarchial Model

Response: Life in hours
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 1519.67 4 379.92 12.54 < 0.0001 significant
A 0.67 1 0.67 0.022 0.8836
B 770.67 1 770.67 25.44 < 0.0001
C 280.17 1 280.17 9.25 0.0067
AC 468.17 1 468.17 15.45 0.0009
Residual 575.67 19 30.30
Lack of Fit 93.00 3 31.00 1.03 0.4067 not significant
Pure Error 482.67 16 30.17
Cor Total 2095.33 23

The Model F-value of 12.54 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case B, C, AC are significant model terms.

Std. Dev. 5.50 R-Squared 0.7253

Mean 40.83 Adj R-Squared 0.6674

6-120
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

C.V. 13.48 Pred R-Squared 0.5616

PRESS 918.52 Adeq Precision 10.747

The "Pred R-Squared" of 0.5616 is in reasonable agreement with the "Adj R-Squared" of
0.6674. A difference greater than 0.20 between the "Pred R-Squared" and the "Adj R-Squared"
indicates a possible problem with your model and/or data.

Design Expert Output for Non-Hierarchical Model

Response: Life in hours
ANOVA for Selected Factorial Model
Analysis of variance table [Partial sum of squares]
Sum of Mean F
Source Squares DF Square Value Prob > F
Model 1519.00 3 506.33 17.57 < 0.0001 significant
B 770.67 1 770.67 26.74 < 0.0001
C 280.17 1 280.17 9.72 0.0054
AC 468.17 1 468.17 16.25 0.0007
Residual 576.33 20 28.82
Lack of Fit 93.67 4 23.42 0.78 0.5566 not significant
Pure Error 482.67 16 30.17
Cor Total 2095.33 23

The Model F-value of 17.57 implies the model is significant. There is only
a 0.01% chance that a "Model F-Value" this large could occur due to noise.

Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case B, C, AC are significant model terms.

The "Lack of Fit F-value" of 0.78 implies the Lack of Fit is not significant relative to the pure
error. There is a 55.66% chance that a "Lack of Fit F-value" this large could occur due
to noise. Non-significant lack of fit is good -- we want the model to fit.

Std. Dev. 5.37 R-Squared 0.7249

Mean 40.83 Adj R-Squared 0.6837
C.V. 13.15 Pred R-Squared 0.6039
PRESS 829.92 Adeq Precision 12.320

The "Pred R-Squared" of 0.6039 is in reasonable agreement with the "Adj R-Squared" of
0.6837. A difference greater than 0.20 between the "Pred R-Squared" and the "Adj R-Squared"
indicates a possible problem with your model and/or data.

(b) Calculate the PRESS statistic, the adjusted R2 and the mean square error for both models.

The PRESS and R2 are in the Design Expert Output above. The PRESS is smaller for the non-
hierarchical model than the hierarchical model suggesting that the non-hierarchical model is a better
predictor.

(c) Find a 95 percent confidence interval on the estimate of the mean response at a cube corner ( x1 = x 2 =
x3 = ±1 ). Hint: Use the result of Problem 6.36.

Design Expert Output

Prediction SE Mean 95% CI low 95% CI high SE Pred 95% PI low 95% PI high
Life 27.45 2.18 22.91 31.99 5.79 15.37 39.54
Life 36.17 2.19 31.60 40.74 5.80 24.07 48.26
Life 38.67 2.19 34.10 43.24 5.80 26.57 50.76
Life 47.50 2.19 42.93 52.07 5.80 35.41 59.59
Life 43.00 2.19 38.43 47.57 5.80 30.91 55.09
Life 34.17 2.19 29.60 38.74 5.80 22.07 46.26
Life 54.33 2.19 49.76 58.90 5.80 42.24 66.43
Life 45.50 2.19 40.93 50.07 5.80 33.41 57.59

(d) Based on the analyses you have conducted, which model would you prefer?

6-121
 Solutions from Montgomery, D. C. (2012) Design and Analysis of Experiments, Wiley, NY

Notice that PRESS is smaller and the adjusted R2 is larger for the non-hierarchical model. This is an
indication that strict adherence to the hierarchy principle isn’t always necessary. Note also that the
confidence interval is shorter for the non-hierarchical model.

6.45. Suppose that you want to run a 23 factorial design. The variance of an individual observation is
expected to be about 4. Suppose that you want the length of a 95% confidence interval on any effect to be
less than or equal to 1.5. How many replicates of the design do you need to run?

With the equations for the se(Effect) and 100(1-α) percent confidence interval on the effects shown below,
we can iteratively estimate the number of replicates. From the table of iterations, 14 replicates are required.

se(Effect) = 2S
n2 k

Effect ± t (α/2,N-p) se(Effect)

n se(Effect) t(0.025,N-p) 95% CI Length

12 0.408 1.987 1.623
13 0.392 1.985 1.557
14 0.378 1.983 1.499
15 0.365 1.981 1.447

6-122