Statistics Problems Solving by using

SPSS
BILAL MASOOD
Problem -3.1

Answer:-

Ho: Temperature does not effects the seal strength

H1: Temperature effects the seal strength
As it is Completely Randomized design(CRD), sowe will use One way ANNOVA to test
the hypothesis

We have the different observations of seal strength against different temperatures

and we have to test the Hypothesis that temperature effects the seal strength.
a) One way ANOVA

ANOVA
observations

Sum of Squares df Mean Square F Sig.

Between Groups 124.408 4 31.102 584.991 .000

Within Groups .797 15 .053
Total 125.206 19

Temperature has significant effect on the seal strength at 5% level of significance. As we

can see from above table Sig. is less than, As P-value is less than α-value (0.05) for
between groups, so we accept H1 which states that temperature effects the seal strength.
So temperature has significant effect on the seal strength at 5% level of significance and
F-ratio of 584.991.To check the significance we will apply the post doc test.

b) Find best temperature to maximize the seal strength:

observations
Duncana

Subset for alpha = 0.05

factor N 1 2 3 4 5

1.00 4 4.9750
2.00 4 7.0500
5.00 4 9.4500
3.00 4 10.2500
4.00 4 12.1000
Sig. 1.000 1.000 1.000 1.000 1.000
Means for groups in homogeneous subsets are displayed.
Uses Harmonic Mean Sample Size = 4.000.because it has the highest
value. Factor 4 gives us maximum seal strength so, at 120 oC seal
strength is maximized
a) 95% Confidence interval for the mean seal strength
corresponding to the temperature 120 oC:

Multiple Comparisons
LSD

Mean Difference 95% Confidence Interval

(I) factor (J) factor (I-J) Std. Error Sig. Lower Bound Upper Bound

4.00 1.00 7.12500* .16304 .000 6.7775 7.4725

2.00 5.05000* .16304 .000 4.7025 5.3975
3.00 1.85000* .16304 .000 1.5025 2.1975
5.00 2.65000* .16304 .000 2.3025 2.9975
*. The mean difference is significant at the 0.05 level.
Above table shows the 95% Confidence Interval lower and upper bound for all
temperatures. lower and upper bound of 95% Confidence Interval for 120 oC is 11.854
and 12.346 respectively.

Problem No. 3.3:

A study has been conducted to test the effect of type of fixture………
Ho:The type of fixture does not affects the measurements
H1:The type of fixture affects the measurements

As it is Completely Randomized design, so we will use One way ANNOVA to test the
hypothesis

ANOVA
observation

Sum of Squares df Mean Square F Sig.

Between Groups .017 4 .004 .776 .554

Within Groups .111 20 .006
Total .128 24

As the P-value is greater than α-value (0.05) between groups, so we accept Howhich
states that the type of fixture does not affects the measurements (strokes length) and
have the F-ratio of 0.776.so here we do not need for post hoc test.

Problem -3.7
 Four different fuels are being evaluated based on the emission
rate. For this purpose four IC engines have been used in the study…….

Ho : Fuels have same effect on the emission rate

H1: Fuels have different effect on the emission rate
As it is Completely Randomized Blocked design, so we will use General Linear model
to test the hypothesis
Univariate Analysis of Variance

Between-Subjects Factors

Value Label N

Factor 1.00 F1 4

2.00 F2 4

3.00 F3 4

4.00 F4 4
Block 1.00 Engine 1 4

2.00 Engine 2 4

3.00 Engine 3 4

4.00 Engine 4 4

Tests of Between-Subjects Effects

Dependent Variable: Observation

Type III Sum of

Source Squares df Mean Square F Sig.

Model 3.307a 7 .472 20.476 .000

Factor .479 3 .160 6.923 .010
Block .064 3 .021 .924 .468
Error .208 9 .023
Total 3.515 16

Above table shows that Factor have P-value is less than α-value (0.05), so we accept H1
which states that Fuels have different effect on the emission rate. So Fuels have
significant effect on the emission rate at 5% level of significance and F-ratio of 20476.
Whereas Engines (Block) have no significant effect on the emission rate as its P-value
is greater than α-value (0.05).
Now to check weather all fuels have different effect or some type of fuels have different
effect and some have same effect on emission rate, we will apply Post Hoc test.

Observation

Subset

Factor N 1 2

Duncana,b F4 4 .2075

F1 4 .2825

F2 4 .5700

F3 4 .6025

Sig. .503 .769

Observation

Subset

Block N 1

Duncana,b Engine 2 4 .3675

Engine 1 4 .3700

Engine 3 4 .4025

Engine 4 4 .5225

Sig. .210
After applying post doc Duncan test it is better to use engine 4 with fuel 3.
Ho: All Relational Algebraic joint operation produce same execution time
H1: All Relational Algebraic joint operation produce different execution time
As it is Completely Randomized Blocked design, So we will use General Linear model to
test the hypothesis
Between-Subjects Factors

factor 1.00 5

2.00 5

3.00 5

4.00 5
block 1.00 4

2.00 4
3.00 4

4.00 4

5.00 4

Tests of Between-Subjects Effects

Dependent Variable: observations

Type III Sum of

Source Squares df Mean Square F Sig.

Model 101.569a 8 12.696 160.203 .000

factor 18.044 3 6.015 75.895 .000
block 6.693 4 1.673 21.114 .000
Error .951 12 .079
Total 102.520 20

a. R Squared = .991 (Adjusted R Squared = .985)

The given experiment data is CRBD. The result shows that the different type of
Relational Algebric joint operation significantly effecting the execution time with F ratio
of 75.895 and error is 0.951.
Different types of queries are also significantly effecting the execution time.

Above table shows that Factor have P-value is less than α-value (0.05), so we accept H1
which states that All Relational Algebraic joint operation produce different execution
time. So All Relational Algebraic joint operation have significant effect on the
execution time at 5% level of significance and F-ratio of 75.895.
P-value of block is also less than α-value (0.05) which proves that all the queries also
have significant effect on the execution time with the F-Ratio of 21.114.
Table also shows that model is significant to analyze the data.

observations

Subset

block N 1 2 3

Duncana,b 3.00 4 .8750

5.00 4 1.9000

4.00 4 2.2000 2.2000

1.00 4 2.3000 2.3000

2.00 4 2.5250

Sig. 1.000 .079 .146

observations

Subset

factor N 1 2 3

Duncana,b 1.00 5 1.1400

3.00 5 1.3800 1.3800

2.00 5 1.7600

4.00 5 3.5600

Sig. .203 .054 1.000

The factor 4 with block 2 is selected which has the highest value in the Duncan test.

Problem 3.11
An Industrial engineer is trying to assess the effect of four different types of fixtures
(A, B, C, D) on the assembly time of a product……
Ho: Type of fixtures does not affect the assembly time of a product
H1: Type of fixtures have effect on the assembly time of a product

As it is Latin Square design, so we will use General Linear model to test the hypothesis

Between-Subjects Factors

Value Label N

Operators 1.00 4

2.00 4

3.00 4

4.00 4
Product 1.00 I 4
2.00 II 4
3.00 III 4
4.00 IV 4
Fxitures 1.00 A 4

2.00 B 4

3.00 C 4

4.00 D 4

Tests of Between-Subjects Effects

Dependent Variable: Time

Type III Sum of

Source Squares df Mean Square F Sig.

Model 159.916a 10 15.992 46.269 .000

Operators .147 3 .049 .142 .931
Product .162 3 .054 .156 .922
Treatment 13.802 3 4.601 13.311 .005
Error 2.074 6 .346
Total 161.990 16

a. R Squared = .987 (Adjusted R Squared = .966)

The given experiment data is Latin Square Design. The result shows that the different
operators and type of product is not significantly effecting the different product assembly
time. The treatment i.e different type of fixtures are significantly effecting the product
assembly time with F ratio 13.311.
As P-value is less than α-value (0.05) for treatments or type of fixtures which is 0.005
with F-Ratio of 13.311, so H1 will be accepted which States that all type of fixtures
have significant effect on the assembly time of a product.
The result also shows that the different operators and type of product is not
significantly effecting the different product assembly time as P-value is greater than
α-value (0.05) which is 0.931 with F-Ratio 0.142and 0.922 with F-Ratio 0.156
respectively.
We apply the Duncan test to see the effect of treatment.

Time

Subset

Product N 1

Duncana,b I 4 2.9000

II 4 2.9750

III 4 3.0250

IV 4 3.1750

Sig. .548

Time

Subset

Fxitures N 1 2 3

Duncana,b B 4 1.6750

A 4 3.0250

C 4 3.0750

D 4 4.3000

Sig. 1.000 .908 1.000

Time
 Subset

Operators N 1

Duncana,b 1.00 4 2.9000

3.00 4 2.9500

2.00 4 3.1000

4.00 4 3.1250

Sig. .621

Product 4,fixture D with operator 4 will give the maximum results

 CHAPTER-4
Problem 4.2

Between-Subjects Factors

tool type 1.00 6

2.00 6
depth of cut 1.00 4

2.00 4

3.00 4

Tests of Between-Subjects Effects

Dependent Variable: power consumption

Type III Sum of

Source Squares df Mean Square F Sig.

Corrected Model 318.784a 5 63.757 2638.214 .000

Intercept 1584.701 1 1584.701 65573.828 .000
tool 2.001 1 2.001 82.793 .000
cut 316.522 2 158.261 6548.724 .000
tool * cut .262 2 .131 5.414 .045
Error .145 6 .024
Total 1903.630 12
Corrected Total 318.929 11

a. R Squared = 1.000 (Adjusted R Squared = .999)

After analyzing the data result above result shows that Type of tool and Depth cut are
significantly effecting the Power consumption.
And we can also see the combined variation effect of both Tool type and depth of cut on
power consumption. Power consumption is significantly effected with combined
variation of tool and cut.
As P-value is less than α-value (0.05) for both tool type and depth of cut which is 0.000
and 0.000 with F-Ratio of 82.93 and 6548.724 respectively, so H1 will be accepted which
States that type of tool and depth of cut have significant effect on the power
consumption.
The result shows that the combined effect of tool and depth of cut alsohave significant
effect on the power consumptionas P-value is smaller than α-value (0.05) which is
0.045 with F-Ratio.we apply the Duncan test to further analyze the results.

observations
Duncana,b

Subset

depth N 1 2 3

1.00 4 5.1000
2.00 4 11.7000
3.00 4 17.6750
Sig. 1.000 1.000 1.000

Problem No 4.3:
The bonding strength of an adhesive has been studied as a function of
temperature and pressure…….
a)
Ho: Temperature and Pressure does not affects the bonding strength of an adhesive
H1: Temperature and Pressure affects the bonding strength of an adhesive

Tests of Between-Subjects Effects

Dependent Variable: Observations

Type III Sum of

Source Squares df Mean Square F Sig.

Corrected Model 29330.111a 8 3666.264 8.242 .002

Intercept 175429.389 1 175429.389 394.371 .000
Pressure 14463.444 2 7231.722 16.257 .001
Temperature 12456.444 2 6228.222 14.001 .002
Pressure * Temperature 2410.222 4 602.556 1.355 .323
Error 4003.500 9 444.833
Total 208763.000 18
Corrected Total 33333.611 17

a. R Squared = .880 (Adjusted R Squared = .773)

The bonding strength of an adhesive has been studied as function of Temperature and
Pressure. As P-Value is .001 and 0.002 for temperature and pressure with F-Ratio of
16.257 and 14.001 respectively, So H1 is accepted which explains that independently
Temperature and Pressure are significantly effecting the adhesive strength at 5%
level of significance. But combined variation of Temperature and Pressure is not
effecting adhesive strength significantly as P-Value 0.323 is greater than α-value.here
we apply post doc test.

Observations
Duncana,b

Subset

Temperature N 1 2 3

90 6 65.5000
60 6 100.8333
30 6 129.8333
Sig. 1.000 1.000 1.000

Observations
Duncana,b

Subset

Pressure N 1 2

120 6 61.0000
130 6 105.8333
140 6 129.3333
Sig. 1.000 .086

Pressure 140 and temperature 30 has the highest values in Duncan test.so we select these
values.

b) Appropriate Residual Plots

Excluded…….

c) Under what conditions process should be operated and Why?

We have to find the most consistent and efficient solution with least coefficient of
variance.
Case Processing Summarya

Cases
Included Excluded Total
 N Percent N Percent N Percent

Observations * Pressure *
18 100.0% 0 0.0% 18 100.0%
Temperature

a. Limited to first 100 cases.

Case Summariesa

Observations

Pressure 120 Temperature 30 1 90.00

2 74.00

Total N 2

Mean 82.0000

Std. Deviation 11.31371

60 1 34.00

2 80.00

Total N 2

Mean 57.0000

Std. Deviation 32.52691

90 1 58.00

2 30.00

Total N 2

Mean 44.0000

Std. Deviation 19.79899

Total N 6

Mean 61.0000

Std. Deviation 24.77902

130 Temperature 30 1 150.00

2 159.00

Total N 2

Mean 154.5000

Std. Deviation 6.36396

60 1 106.00

2 115.00

Total N 2

Mean 110.5000

Std. Deviation 6.36396

90 1 35.00
 2 70.00

Total N 2

Mean 52.5000

Std. Deviation 24.74874

Total N 6

Mean 105.8333

Std. Deviation 47.25004

140 Temperature 30 1 138.00

2 168.00

Total N 2

Mean 153.0000

Std. Deviation 21.21320

60 1 110.00

2 160.00

Total N 2

Mean 135.0000

Std. Deviation 35.35534

90 1 96.00

2 104.00

Total N 2

Mean 100.0000

Std. Deviation 5.65685

Total N 6

Mean 129.3333

Std. Deviation 30.45434

Total N 18

Mean 98.7222

Std. Deviation 44.28093

a. Limited to first 100 cases.

Coefficient of Variance
Pressure Temperature
(S.D/Mean)×100

120 30 13.7

120 60 57.0
120 90 44.9

130 30 4.1

130 60 5.7

130 90 47.1

140 30 13.87

140 60 26.1

140 90 5.15

So above table shows the minimum Coefficient of variance is 4.1 at 130 Pressure and 30
Temperature. So results are most efficient and consistent at these conditions So bonding
strength of adhesive will be maximum at 130 Pressure and 30 Temperature.