Parcial #2

Members: Abner J. Suárez Díaz & Andrea I. Galarcio Meléndez.

1. Three brands of batteries are under study. It is suspected that the lives (in weeks) of the
three brands are different. Five randomly selected batteries of each brand are tested with
the following results:
Figure 1. Data brands of batteries are under study

Weeks of life
Brand 1 Brand 2 Brand 3
100 76 108
96 80 100
92 75 96
96 84 98
92 82 100
µ
Own elaboration
a) Are the lives of these brands of batteries different? item Analyze the residuals from this
experiment.
Figure 2. Analysis of Variance for Weeks of life

Analysis of Variance
Source DF Adj SS Adj MS F-Value P-Value
Brands 3 119.1 598.07 38.34 0.000
Error 12 187.2 15.60    
Total 14 1383.3      
Own elaboration
According to the ANOVA analysis shown in the figure 2, it is observed that the P-
Value=0.000<0.005 which implies that the H 0 is rejected since the value of this is lower, having
a conclusion that the brands that are under study are different between them, as true the
hypothesis given in the statement of the exercise.

H 0 :μ 1=μ2=μ3 =μ

H A : μi ≠ μ j For some i≠ j

H 0 :σ 21=σ 22=σ 23=σ 2

H A :σ i2 ≠ σ 2j For some i≠ j
 Graphic 1. Analysis of the validation assumptions

Versus Fits
Own (responses are BRAND 1. BRAND 2. BRAND 3) elaboration
7,5
According to the graph 1 it
5,0
can be analyzed that
2,5
Residual

the data handle a

0,0
constant variance.
-2,5

-5,0
80 85 90 95 100
Fitted Value

H 0 : Datais distributed normally

H A : Data not is distributed normally

Graphic 2. Analysis of the validation assumptions

Normal Probability Plot

(responses are BRAND 1. BRAND 2. BRAND 3)
99

95
90

80
70
Percent

60
50
40
30
20

10
5

1
-10 -5 0 5 10
Residual

Own elaboration
Taking into account the graph 2 it is indicated that the data is distributed normally.
Therefore, it is deduced according to the residual analysis, the compliance is indicated that this
model was carried out completely at random, and that there is no other external factor that
influences significantly and that the type of brand is the one that affects the variable response
which is the life (in weeks) of each brand of battery and must be taken into account if you want
to improve its performance.

b) Construct a 95 percent confidence interval estimate on the mean life of battery Brand.
Construct a 99 percent confidence interval estimate on the mean difference between
the lives of battery brands 2 and 3.
c) Which brand would you select for use? If the manufacturer will replace without charge
any battery that fails in less than 85 weeks, what percentage would the company expect
to replace?

Construct a 95 percent interval estimate on the mean life of battery brand 2. Construct a 99
percent interval estimate on the mean difference between the lives of battery brands 2 and 3.

Performing an analysis of the data obtained using the previous equation, we can say that with
95% confidence the average life of battery number 2 is under the limits of 75,551 and 83,249.

In other words, with a 95% confidence level we indicate that if you took 100 random samples
from the population, you could expect that approximately 95 of the samples would contain the
population difference. So we can say that, based on these results, it is 95% certain that the mean
of the populations is between -28,631 and -13,369.

c) Which brand would you select for use? If the manufacturer will replace without charge any
battery that fails in less than 85 weeks, what percentage would the company expect to replace?

Chose brand 3 for longest life. Mean life of this brand in 100.4 weeks, and the variance of life is

estimated by 15.60 (MSE). Assuming normality, then the probability of failure before 85 weeks
is:

85−100.4
Ф( ) = Ф (-3.90) = 0.0005
√ 15.60

That is, about 5 out of 100,000 batteries will fail before 85 weeks.

2. An industrial engineer is investigating the effect of four assembly methods (A, B, C, D)

on the assembly time for a color television component. Four operators are selected for the
study. Furthermore, the engineer knows that each assembly method produces such fatigue
that the time required for the last assembly may be greater than the time required for the
first, regardless of the method. That is, a trend develops in the required assembly time. To
account for this source of variability, the engineer uses the Latin square design shown
below. Analyze the data from this experiment (α = 0,05) and draw appropriate
conclusions.

Figure 3. Data of effect of four assembly methods.

Operator
Order of assembly 1 2 3 4
1 C = 10 D = 14 A=7 B=8
2 B=7 C = 18 D = 11 A=8
3 A=5 B = 10 C = 11 D=9
4 D = 10 A = 10 B = 12 C = 14
Own elaboration
H 0 :μ A=μ B=μ C =μ D=μ

H A : μi ≠ μ j For some i≠ j con i , j= A , B , C , D

Figure 4. Analysis of Variance for Weeks of life

Analysis of Variance
Source DF Adj SS Adj MS F-Value P-Value
Operario 4 51.50 17.167 9.81 0.010
Método de ensamble 4 72.50 24.167 13.81 0.004
Order Assembly 4 18.50 6.167 3.52 0.089
Error 6 10.50 1.750    
Total 15 153      

Own elaboration

According to the analysis of variance, the results of the p-Values for the operator and the
assembly order are 0.010> 0.05 and 0.089> 0.05 respectively, which means that the H o is not

rejected, that is, these variables they do not significantly affect the assembly time, while for the
assembly method with a P- Value of 0.004 <0.05 it means that the H ois rejected, which
indicates that this factor is the one who relates the changes or differences in the response
variable.
H 0 : Datais distributed normally

H A : Data not is distributed normally

H 0 :σ 2A =σ 2B=σ 2C =σ 2D =σ 2

H A :σ i2 ≠ σ 2j For some i≠ j

Graphic 3. Analysis of the validation assumptions residuals

 Residual Plots for Tiempo de montaje
Normal Probability Plot Versus Fits
99

90 1

Residual
Percent
50 0

10 -1

1
-2 -1 0 1 2 5 10 15
Residual Fitted Value

Histogram Versus Order

4,8
1
3,6
Frequency

Residual
0
2,4

1,2 -1

0,0
-1,5 -1,0 -0,5 0,0 0,5 1,0 1,5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Residual Observation Order

Own elaboration
According to the assumptions through the residuals it can be concluded that the data are
normally distributed, we can also affirm that they have an equal variance as can be seen in the
graph located in the upper right part and finally that the data are independent about from
another, which means that the data was taken randomly.

Graph 4. Comparisons Operator Fisher Individual

Fisher Individual 95% CIs
Differences of Means for Tiempo de montaje

2-1

3-1

4-1
Operario

3-2

4-2

4-3

-5,0 -2,5 0,0 2,5 5,0 7,5

If an interval does not contain zero, the corresponding means are significantly different.

Own elaboration
On the other hand, by analyzing Fisher's graph which compares two by two more exactly each
of the operators with a confidence level of 95% shows us that the operators who differ in
assembly times are (3 -2, 4-2 and 2-1) having operator # 2 as a common factor, if you want to
improve performance, you should further evaluate how this person operates.

Graph 5. Comparisons Assembly Method Fisher Individual

Fisher Individual 95% CIs

Differences of Means for Tiempo de montaje

B-A

C-A
Método de ensamble

D-A

C-B

D-B

D-C

-5,0 -2,5 0,0 2,5 5,0 7,5

If an interval does not contain zero, the corresponding means are significantly different.

Own elaboration
Following this order of idea, the assembly method is compared with the same Fisher method,
resulting in the methods that differ the most at the time of assembly are the following (C-A, D-
A, C-B and D-C).

Graph 6. Comparisons Order Assembly Fisher Individual

 Fisher Individual 95% CIs
Differences of Means for Tiempo de montaje

2-1

3-1
Order Assembly

4-1

3-2

4-2

4-3

-5,0 -2,5 0,0 2,5 5,0

If an interval does not contain zero, the corresponding means are significantly different.

Own elaboration
To finish with Fisher's analysis, we have this graph which summarizes the differences between
each of the assembly orders and the assembly time, where it is shown that the only pair that
differs is (4-3).
Graph 7. Analysis Tukey for Operator
Tukey Simultaneous 95% CIs
Differences of Means for Tiempo de montaje

2-1

3-1

4-1
Operario

3-2

4-2

4-3

-5,0 -2,5 0,0 2,5 5,0 7,5 10,0

If an interval does not contain zero, the corresponding means are significantly different.

Own elaboration
To make a better decision, not only the analysis of the Fisher graphs is taken into account but
also the comparison with the Tukey graphs is made where the differences lie in the following
pairs (2-1 and 4- 2), here it is specified with both analyzes that these pairs are of greater
significant difference.
Graph 8. Analysis Tukey for Assembly Method

Tukey Simultaneous 95% CIs

Differences of Means for Tiempo de montaje

B-A

C-A
Método de ensamble

D-A

C-B

D-B

D-C

-5,0 -2,5 0,0 2,5 5,0 7,5 10,0

If an interval does not contain zero, the corresponding means are significantly different.

Own elaboration
In the same way, the Tukey analysis is carried out for the assembly method, where again it is
obtained that the different ones are (C-A, D-A and C-B,).
Graph 9. Analysis Tukey for Order Assembly

Tukey Simultaneous 95% CIs

Differences of Means for Tiempo de montaje

2-1

3-1
Order Assembly

4-1

3-2

4-2

4-3

-5,0 -2,5 0,0 2,5 5,0 7,5

If an interval does not contain zero, the corresponding means are significantly different.
 Own elaboration
In this case, the analysis shows for Tukey's test that none of the pairs for the assembly order
are different, unlike Fisher's analysis.
Graph 10. Analysis Bonferroni for Order Assembly

Bonferroni Simultaneous 95% CIs

Differences of Means for Tiempo de montaje

2-1

3-1
Order Assembly

4-1

3-2

4-2

4-3

-5,0 -2,5 0,0 2,5 5,0 7,5

If an interval does not contain zero, the corresponding means are significantly different.

Own elaboration
For the tiebreaker, a final Bonferroni analysis is performed, which detects that for the assembly
order all the pairs are equal.

Graph 11. Analysis Bonferroni for Operator

 Bonferroni Simultaneous 95% CIs
Differences of Means for Tiempo de montaje

2-1

3-1

4-1
Operario

3-2

4-2

4-3

-5 0 5 10

If an interval does not contain zero, the corresponding means are significantly different.

Own elaboration
In an analogous way, it is carried out in analysis using the Bonferroni graph which denotes that
for the operator the only different pair is (2-1).

Graph 12. Analysis Bonferroni for Operator

Bonferroni Simultaneous 95% CIs

Differences of Means for Tiempo de montaje

B-A

C-A
Método de ensamble

D-A

C-B

D-B

D-C

-5,0 -2,5 0,0 2,5 5,0 7,5 10,0

If an interval does not contain zero, the corresponding means are significantly different.

Own elaboration
For the last analysis, the assembly method has to be, which differ are the method (C-A), which
is affirmed based on the analyzes carried out previously.
 3. Suppose that in Problem 2 the engineer suspects that the workplaces used by the four operators
may represent an additional source of variation. A fourth factor, workplace (α, β, γ, δ) may be
introduced and another experiment conducted, yielding the Graeco-Latin square that follows.
Analyze the data from this experiment (α = 0,01) and draw conclusions.

4. Figure 5. Data of effect of four assembly methods with four operators.

Operator

Order of assembly 1 2 3 4
1 Cβ =11 Bγ = 10 Dδ = 14 Aα = 8
2 Bα = 8 Cδ = 12 Aγ = 10 Dβ = 12
3 Aδ = 9 Dα = 11 Bβ = 7 Cγ = 15
4 Dγ = 9 Aβ = 8 Cα = 18 Bδ = 6
Own elaboration

Performing the ordering of the data and subsequent analysis in the Minitab software, we proceed
to demarcate the main data as a reason for answering the question mentioned above.

Therefore, taking the observations and data from the Greek - Latin table, Minitab ANOVA gives
us the following.

Factor Type Levels Values

Method fixed 4 ABCD
Order random 4 1234
Operator random 4 1234
Workplac random 4 abcd

Analysis of Variance for Time, using Adjusted SS for Tests

Source DF Seq SS Adj SS Adj MS F P

Method 3 95.500 95.500 31.833 3.47 0.167
Order 3 0.500 0.500 0.167 0.02 0.996
Operator 3 19.000 19.000 6.333 0.69 0.616
Workplac 3 7.500 7.500 2.500 0.27 0.843
Error 3 27.500 27.500 9.167
Total 15 150.000

Method and workplace do not have a significant effect on assembly time. However, there are
only three degrees of freedom for error, so the test is not very sensitive.