Republic of the Philippines

CENTRAL MINDANAO UNIVERSITY

College of Education
Graduate School Program
University Town, Musuan, Bukidnon, 8710

June 30, 2023

Student: VICTORIANO A. GAPAS JR.

Instructor: DR. RAUL ORONGAN
Program: Master of Arts in Language Education (English)
Subject: Statistical Methods in Educational Research (EDUC254)

PROBLEM SET 5
ONE-WAY ANOVA
INSTRUCTIONS:
1. Use the six steps hypothesis test in analysing the ff problems.
2. Submit the summarized result and the outputs after analysis. (TURN IN THE
GOOGLE CLASS IN PDF)

PART I: EXERCISES
ANOVA
1. A consumer group wishes to test the longevity of various brands of lightbulbs.
They purchase six lightbulbs each of three brands and burn them until they fail.
Here are the results (numbers are hours that the bulb lasted):

Brand X Brand Y Brand Z

bulb #1 2588 2020 2423
bulb #2 2725 2084 2328
bulb #3 2827 2027 2247
bulb #4 2986 1117 1304
bulb #5 2622 1081 1118
bulb #6 2883 1017 1374

Brand Z looks better than the others, but it might just be chance variation - the
brands might all be the same. State an explicit null hypothesis and test to see.
(You may treat each column as a random sample of 6 bulbs from the given
brand's production, for the purposes of this problem. Actually choosing such a
sample is a tricky business.)
A. IDENTIFY THE DEPENDENT AND INDEPENDENT VARIABLES.
DV: Longevity of lightbulbs (hours)
IV: Brands of lightbulbs (Brand X, Brand Y, Brand Z)

B. STATE THE NULL AND ALTERNATIVE HYPOTHESIS

HO: The mean longevity of the lightbulbs is the same for all three brands.
HA: The mean longevity of the lightbulbs differs across the three brands.
 C. ESTABLISH THE LEVEL OF SIGNIFICANCE (USUALLY SET TO 0.05
LEVEL FOR SOCIAL RESEARCHES UNLESS SPECIFIED)
Specify the α level: α = .05
D. DETERMINE THE APPROPRIATE TEST STATISTICS
One-way Anova because it is the appropriate test statistics to
compare the means of more than two groups.

E. COMPUTATION / ANALYSIS ( USE BASIC AVAILABLE STATISTICAL

SOFTWARE)
SUMMARY TABLE/S:
VARIABLE STANDARD SD F-VALUE PROB (Sig.
MEAN 2- tailed)
BRAND X
2771.8333 154.72998
BULBS
BRAND Y
1557.6667 533.81108 11.254 .001**
BULBS
BRAND Z
1799.0000 593.18328
BULBS

F. INTERPRETATION / IMPLICATION / CONCLUSION

Since the probability (2-tailed significance) is less than our desired
significance threshold (P <0.05), we reject the null hypothesis and accept the
alternative hypothesis. This means that there is a significant difference in the
mean longevity of lightbulbs across the three bulb brands where Brand X has
a significantly higher mean longevity compared to Brand Y and Brand Z, with a
mean difference of 1214.16667 (p < 0.001) and 972.83333 (p = 0.007)
respectively while Brand Y does not have a significantly different mean
longevity compared to Brand Z, with a mean difference of -241.33333 (p =
0.654).

Descriptives
HOURS

N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Minimum Maximum

Lower Bound Upper Bound

1.00 6 2771.8333 154.72998 63.16825 2609.4542 2934.2125 2588.00 2986.00

2.00 6 1557.6667 533.81108 217.92746 997.4663 2117.8670 1017.00 2084.00
3.00 6 1799.0000 593.18328 242.16606 1176.4923 2421.5077 1118.00 2423.00
Total 18 2042.8333 697.11610 164.31184 1696.1657 2389.5010 1017.00 2986.00

ANOVA
HOURS

Sum of Squares df Mean Square F Sig.

Between Groups 4957694.333 2 2478847.167 11.254 .001

 Within Groups 3303810.167 15 220254.011
Total 8261504.500 17

Multiple Comparisons
Dependent Variable: HOURS
Tukey HSD

(I) BULBBRANDS (J) BULBBRANDS Mean Difference Std. Error Sig. 95% Confidence Interval
(I-J) Lower Bound Upper Bound

2.00 1214.16667* 270.95757 .001 510.3624 1917.9709

1.00
3.00 972.83333* 270.95757 .007 269.0291 1676.6376
1.00 -1214.16667* 270.95757 .001 -1917.9709 -510.3624
2.00
3.00 -241.33333 270.95757 .654 -945.1376 462.4709
1.00 -972.83333* 270.95757 .007 -1676.6376 -269.0291
3.00
2.00 241.33333 270.95757 .654 -462.4709 945.1376

*. The mean difference is significant at the 0.05 level.

HOURS
Tukey HSD

BULBBRANDS N Subset for alpha = 0.05

1 2

2.00 6 1557.6667
3.00 6 1799.0000
1.00 6 2771.8333
Sig. .654 1.000

Means for groups in homogeneous subsets are displayed.

a. Uses Harmonic Mean Sample Size = 6.000.

2. Someone claims that the percentage of their incomes that people pay in total
taxes is the same for different income levels. A "random sample" of tax data
showing the average taxes paid in six different cities is given below; the column
heading is the income level. Form an explicit null hypothesis and perform an
ANOVA test to evaluate this claim. Bonus: discuss some difficulties with this
"random" sample - since we are picking cities, how does it affect the statement
of the null hypothesis?
City 25000 income 50000 income 75000 income
level level level
A 5.4 8.5 10.0
B 4.4 9.4 11.8
 C 4.9 5.8 14.0
D 2.3 7.4 18.1
E 10.8 25.5 16.1
F 9.8 20.9 28.4

A. IDENTIFY THE DEPENDENT AND INDEPENDENT VARIABLES.

DV: average taxes paid in six different cities
IV: income levels

Difficulties: A potential bias is introduced in this sample problem given that we

are selecting cities since tax rates and policies can differ among cities,
potentially impacting the average taxes paid. As a result, the assumption of
randomly sampling cities may be questionable. Instead, it may be more suitable
to randomly sample individuals within each city to acquire a representative
sample for the hypothesis test.

B. STATE THE NULL AND ALTERNATIVE HYPOTHESIS

HO: The average taxes paid are equal across all income levels.
HA: The average taxes paid are not equal across all income levels.

C. ESTABLISH THE LEVEL OF SIGNIFICANCE (USUALLY SET TO 0.05

LEVEL FOR SOCIAL RESEARCHES UNLESS SPECIFIED)
Specify the α level: α = .05
D. DETERMINE THE APPROPRIATE TEST STATISTICS
One-way Anova because it is the appropriate test statistics to
compare the means of more than two groups.
E. COMPUTATION / ANALYSIS ( USE BASIC AVAILABLE STATISTICAL
SOFTWARE)
SUMMARY TABLE/S:
VARIABLE STANDARD SD F-VALUE PROB (Sig.
MEAN 2- tailed)
25K INCOME
6.2667 3.31280
LEVEL
50K INCOME
12.9167 8.18570 3.945 .042*
LEVEL
75K INCOME
16.4000 6.55530
LEVEL

F. INTERPRETATION / IMPLICATION / CONCLUSION

Since the probability (2-tailed significance) is less than our desired
significance threshold (P <0.05), we reject the null hypothesis and accept the
alternative hypothesis. This means that there is a significant difference in the
average city taxes with respect to income levels.

Descriptives
INCOMELEVEL

N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Minimum Maximum

Lower Bound Upper Bound

1.00 6 6.2667 3.31280 1.35245 2.7901 9.7432 2.30 10.80

2.00 6 12.9167 8.18570 3.34180 4.3263 21.5070 5.80 25.50
3.00 6 16.4000 6.55530 2.67619 9.5206 23.2794 10.00 28.40
Total 18 11.8611 7.36783 1.73661 8.1972 15.5250 2.30 28.40

ANOVA
INCOMELEVEL

Sum of Squares df Mean Square F Sig.

Between Groups 318.081 2 159.041 3.945 .042

Within Groups 604.762 15 40.317
Total 922.843 17

Multiple Comparisons
Dependent Variable: INCOMELEVEL
Tukey HSD

(I) CITYTAXES (J) CITYTAXES Mean Std. Error Sig. 95% Confidence Interval
Difference (I-J) Lower Bound Upper Bound

2.00 -6.65000 3.66594 .199 -16.1722 2.8722

1.00
3.00 -10.13333* 3.66594 .036 -19.6555 -.6112
1.00 6.65000 3.66594 .199 -2.8722 16.1722
2.00
3.00 -3.48333 3.66594 .618 -13.0055 6.0388
1.00 10.13333* 3.66594 .036 .6112 19.6555
3.00
2.00 3.48333 3.66594 .618 -6.0388 13.0055

*. The mean difference is significant at the 0.05 level.

INCOMELEVEL
Tukey HSD

CITYTAXES N Subset for alpha = 0.05

1 2

1.00 6 6.2667
2.00 6 12.9167 12.9167
3.00 6 16.4000
Sig. .199 .618

Means for groups in homogeneous subsets are

displayed.
a. Uses Harmonic Mean Sample Size = 6.000.
 3. Sixteen midsize cars (divided into four groups) were chosen (and divided) at
random to test the mileage claims of four different brands of gasoline. The table
lists the mileage (in miles/gallon) achieved with each brand. Is there any
evidence here that all gasolines are NOT the same? Use ANOVA to find out.
Car # gas brand A gas brand B gas brand C gas brand D
1 20.2 33.9 46.0 19.7
2 28.9 33.7 48.0 16.3
3 27.5 35.0 44.5 17.9
4 29.0 33.9 45.4 15.3

A. IDENTIFY THE DEPENDENT AND INDEPENDENT VARIABLES.

DV: mileage achieved by the cars
IV: brand of gasoline used

B. STATE THE NULL AND ALTERNATIVE HYPOTHESIS

HO: The mean mileage achieved with different brands of gasoline is the same.
HA: The mean mileage achieved with different brands of gasoline is not the
same.

C. ESTABLISH THE LEVEL OF SIGNIFICANCE (USUALLY SET TO 0.05

LEVEL FOR SOCIAL RESEARCHES UNLESS SPECIFIED)
Specify the α level: α = .05
D. DETERMINE THE APPROPRIATE TEST STATISTICS
One-way Anova because it is the appropriate test statistics to
compare the means of more than two groups.

E. COMPUTATION / ANALYSIS (USE BASIC AVAILABLE STATISTICAL

SOFTWARE)
SUMMARY TABLE/S:
VARIABLE STANDARD SD F-VALUE PROB (Sig.
MEAN 2- tailed)
GAS BRAND
26.4000 4.18967
A
GAS BRAND
34.1250 .59090
B 99.191 .000**
GAS BRAND
45.9750 1.48408
C
GAS BRAND
17.3000 1.92527
D
 F. INTERPRETATION / IMPLICATION / CONCLUSION
Since the probability (2-tailed significance) is less than our desired
significance threshold (P <0.05), we reject the alternative hypothesis and
accept the null hypothesis. This means that there is no statistical evidence to
suggest that mileage differ significantly between the gas brands. In other words,
there is no basis to conclude that one gas brand performs better or worse in
terms of mileage claims compared to the others.

Descriptives
MILEAGECLAIMS

N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Minimum Maximum

Lower Bound Upper Bound

1.00 4 26.4000 4.18967 2.09483 19.7333 33.0667 20.20 29.00

2.00 4 34.1250 .59090 .29545 33.1847 35.0653 33.70 35.00
3.00 4 45.9750 1.48408 .74204 43.6135 48.3365 44.50 48.00
4.00 4 17.3000 1.92527 .96264 14.2365 20.3635 15.30 19.70
Total 16 30.9500 11.08410 2.77103 25.0437 36.8563 15.30 48.00

ANOVA
MILEAGECLAIMS

Sum of Squares df Mean Square F Sig.

Between Groups 1771.425 3 590.475 99.191 .000

Within Groups 71.435 12 5.953
Total 1842.860 15

Multiple Comparisons
Dependent Variable: MILEAGECLAIMS
Tukey HSD

(I) GASBRAND (J) GASBRAND Mean Std. Error Sig. 95% Confidence Interval
Difference (I-J) Lower Bound Upper Bound

2.00 -7.72500* 1.72524 .004 -12.8471 -2.6029

1.00 3.00 -19.57500* 1.72524 .000 -24.6971 -14.4529

4.00 9.10000* 1.72524 .001 3.9779 14.2221

1.00 7.72500* 1.72524 .004 2.6029 12.8471
2.00 3.00 -11.85000* 1.72524 .000 -16.9721 -6.7279
4.00 16.82500* 1.72524 .000 11.7029 21.9471
1.00 19.57500* 1.72524 .000 14.4529 24.6971
3.00 2.00 11.85000* 1.72524 .000 6.7279 16.9721
4.00 28.67500* 1.72524 .000 23.5529 33.7971
1.00 -9.10000* 1.72524 .001 -14.2221 -3.9779
4.00
2.00 -16.82500* 1.72524 .000 -21.9471 -11.7029
 3.00 -28.67500* 1.72524 .000 -33.7971 -23.5529

*. The mean difference is significant at the 0.05 level.

MILEAGECLAIMS
Tukey HSD

GASBRAND N Subset for alpha = 0.05

1 2 3 4

4.00 4 17.3000
1.00 4 26.4000
2.00 4 34.1250
3.00 4 45.9750
Sig. 1.000 1.000 1.000 1.000

Means for groups in homogeneous subsets are displayed.

a. Uses Harmonic Mean Sample Size = 4.000.

4. Three methods of teaching infection control are being compared using three
groups of students selected at random. Their scores out of 20 are shown. Is
there a significant difference between the means of each group? Explain
clearly.
Method A 5.5 9.5 5.5 8.5 5.5 7.5 6.5 7.5 6.5 6.5 6.5 7.5
Method B 12.5 6.5 11.5 7.5 15.5 7.5 15.5 8.5 15.5 9.5 9.5 9.5
Method C 22.2 24.2 22.2 27.2 24.2 22.2 9.2 22.2 22.2 8.2 7.2
9.2

A. IDENTIFY THE DEPENDENT AND INDEPENDENT VARIABLES.

DV: method of teaching infection control
IV: students' scores out of 20
B. STATE THE NULL AND ALTERNATIVE HYPOTHESIS
HO: There is no significant difference between the means of each group
HA: There is a significant difference between the means of at least one pair of
groups

C. ESTABLISH THE LEVEL OF SIGNIFICANCE (USUALLY SET TO 0.05

LEVEL FOR SOCIAL RESEARCHES UNLESS SPECIFIED)
Specify the α level: α = .05
D. DETERMINE THE APPROPRIATE TEST STATISTICS
One-way Anova because it is the appropriate test statistics to
compare the means of more than two groups.
 E. COMPUTATION / ANALYSIS (USE BASIC AVAILABLE STATISTICAL
SOFTWARE)
SUMMARY TABLE/S:
VARIABLE STANDARD SD F-VALUE PROB (Sig.
MEAN 2- tailed)
METHOD A 6.9167 1.24011
METHOD B 10.7500 3.30633 17.868 .000**
METHOD C 18.3667 7.48129

F. INTERPRETATION / IMPLICATION / CONCLUSION

Since the probability (2-tailed significance) is less than our desired
significance threshold (P <0.05), we reject the alternative hypothesis (There is
a significant difference between the means of at least one pair of groups) and
accept the null hypothesis (There is no significant difference between the
means of each group). This means that there are no significant differences
between any of the three methods being compared.

Descriptives
STUDENTSCORES

N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Minimum Maximum

Lower Bound Upper Bound

1.00 12 6.9167 1.24011 .35799 6.1287 7.7046 5.50 9.50

2.00 12 10.7500 3.30633 .95446 8.6493 12.8507 6.50 15.50
3.00 12 18.3667 7.48129 2.15966 13.6133 23.1201 7.20 27.20
Total 36 12.0111 6.69344 1.11557 9.7464 14.2758 5.50 27.20

ANOVA
STUDENTSCORES

Sum of Squares df Mean Square F Sig.

Between Groups 815.242 2 407.621 17.868 .000

Within Groups 752.833 33 22.813
Total 1568.076 35

Multiple Comparisons
Dependent Variable: STUDENTSCORES
Tukey HSD

(I) METHODS (J) METHODS Mean Std. Error Sig. 95% Confidence Interval
Difference (I-J) Lower Bound Upper Bound

2.00 -3.83333 1.94992 .137 -8.6180 .9514

1.00
3.00 -11.45000* 1.94992 .000 -16.2347 -6.6653
2.00 1.00 3.83333 1.94992 .137 -.9514 8.6180
 3.00 -7.61667* 1.94992 .001 -12.4014 -2.8320
1.00 11.45000* 1.94992 .000 6.6653 16.2347
3.00
2.00 7.61667* 1.94992 .001 2.8320 12.4014

*. The mean difference is significant at the 0.05 level.

STUDENTSCORES
Tukey HSD

METHODS N Subset for alpha = 0.05

1 2

1.00 12 6.9167
2.00 12 10.7500
3.00 12 18.3667
Sig. .137 1.000

Means for groups in homogeneous subsets are

displayed.
a. Uses Harmonic Mean Sample Size = 12.000.

PART II:

USE YOUR 180 GENERATED RANDOM SAMPLES IN PS1, PERFORM THE

SUGGESTED STEPS IN HYPOTHEIS TEST:

1. DO MOTIVE AND JOBPERF DIFFER SIGNIFICANTLY WHEN GROUPED

BY EDUC?
NOTE: JUST PRESENT THE SUMMARY TABLE AS SUGGESTED AND
BE SURE TO ATTACH OR APPEND THE OUTPUTS OF YOUR ANALYSIS
AFTER THE HYPOTHESIS TEST.

A. IDENTIFY THE DEPENDENT AND INDEPENDENT VARIABLES.

DV: Motive and Job Performance

IV: Education Status

B. STATE THE NULL AND ALTERNATIVE HYPOTHESIS

HO: Motive and job performance do not differ significantly when grouped by
educational status.
HA: Motive and job performance differ significantly when grouped by
educational status.
 C. ESTABLISH THE LEVEL OF SIGNIFICANCE (USUALLY SET TO 0.05 LEVEL FOR
SOCIAL RESEARCHES UNLESS SPECIFIED)
Specify the α level: α = .05
D. DETERMINE THE APPROPRIATE TEST STATISTICS

One-way Anova because it is the appropriate test statistics to

compare the means of more than two groups.

E. COMPUTATION / ANALYSIS (USE BASIC AVAILABLE STATISTICAL SOFTWARE)

SUMMARY TABLE/S:
VARIABLE STANDARD SD F-VALUE PROB
MEAN (Sig. 2-
tailed)
ELEMENTARY 1.0000 .00000
LEVEL
SECONDARY 1.0000 .00000
.187 .830
LEVEL
TERTIARY LEVEL 1.0000 .00000

F. INTERPRETATION / IMPLICATION / CONCLUSION

Since the P-value is greater than our desired significance threshold (P
<0.05), we accept the null hypothesis and reject the alternative one. This means
that motive and job performance do not differ significantly when grouped by
educational status.

Descriptives

N Mean Std. Std. 95% Confidence Interval for Minimum Maximu

Deviation Error Mean m

Lower Bound Upper Bound

1.00 46 1.0000 .00000 .00000 1.0000 1.0000 1.00 1.00

2.00 79 1.0000 .00000 .00000 1.0000 1.0000 1.00 1.00

MOTIVES
3.00 55 1.0000 .00000 .00000 1.0000 1.0000 1.00 1.00

Total 180 1.0000 .00000 .00000 1.0000 1.0000 1.00 1.00

1.00 46 4.0861 .41319 .06092 3.9634 4.2088 3.26 5.00

JOBPERF 2.00 79 4.1305 .46184 .05196 4.0271 4.2340 3.24 5.00

S 3.00 55 4.1325 .40689 .05486 4.0225 4.2425 3.36 5.00

Total 180 4.1198 .43151 .03216 4.0563 4.1832 3.24 5.00

ANOVA

Sum of Squares df Mean Square F Sig.

Between Groups .000 2 .000 . .

MOTIVES Within Groups .000 177 .000

Total .000 179

 Between Groups .070 2 .035 .187 .830

JOBPERFS Within Groups 33.260 177 .188

Total 33.330 179

Multiple Comparisons
Tukey HSD

Dependent (I) (J) Mean Std. Error Sig. 95% Confidence Interval
Variable EDUCS EDUCST Difference (I- Lower Bound Upper Bound
TATUS ATUS J)

2.00 -.04442 .08040 .845 -.2344 .1456

1.00
3.00 -.04646 .08661 .854 -.2512 .1583

1.00 .04442 .08040 .845 -.1456 .2344

JOBPERFS 2.00
3.00 -.00204 .07613 1.000 -.1820 .1779

1.00 .04646 .08661 .854 -.1583 .2512

3.00
2.00 .00204 .07613 1.000 -.1779 .1820

JOBPERFS
Tukey HSD

EDUCSTATUS N Subset for alpha

= 0.05

1.00 46 4.0861
2.00 79 4.1305
3.00 55 4.1325
Sig. .835

Means for groups in homogeneous subsets are

displayed.
a. Uses Harmonic Mean Sample Size = 57.057.
b. The group sizes are unequal. The harmonic
mean of the group sizes is used. Type I error
levels are not guaranteed.