ANALYSIS OF VARIANCE (ANOVA)

Example:
A manufacturer of paper used for making grocery bags is interested in improving the tensile strength of the product. Product
engineering thinks that tensile strength is a function of the hardwood concentration in the pulp and that the range of hardwood
concentrations of practical interest is between 5 and 20%. A team of engineers responsible for the study decides to investigate four
levels of hardwood concentration: 5%, 10%, 15%, and 20%. They decide to make up six test specimens at each concentration level,
using a pilot plant. All 24 specimens are tested on a laboratory tensile tester, in random order. The data from this experiment are
shown below:

Hardwood Concentration (%)

Observations 5 10 15 20 Totals
1 7 12 14 19
2 8 17 18 19
3 15 13 19 22
4 11 18 17 23
5 9 19 16 18
6 10 15 18 20
Totals 60 94 102 121 377
Means 10 15.67 17 20.17 15.71
n 6 6 6 6 24
Sum of Squares 640 1,512 1,750 2,459 6,361

Step 1: Statement of the null and alternative hypotheses.

Ho: There is no significant difference in the tensile strength of the different hardwood concentrations.

Ha: There is a significant difference in the tensile strength of the different hardwood concentrations.
There is an effect of the hardwood concentration in the tensile strength of the product.

Step 2: Calculate the test statistics (Raw Score Method).

a. Compute the raw scores.

b. Compute for the sums of N for each group, the sums of raw scores X, and the sums of squared scores X2.

c. Compute Sums of Squares

SST (Sums of Squares for the SSB (Sums of Squares for SSW (Sums of Squares for Within
Total Variability) Between Group Variability) Group Variability)
( ) ( ) ((∑ ) )
{∑ }
( )

d. Determine the degree of freedom.

Between groups: Within groups: Total:

e. Find the Mean Squares (MS)

f. Compute for the F-ratio.

Step 3: Determine the critical value of F.
At =0.05, Fcrit(4,20) = 3.10.

Step 4: Construct ANOVA Summary Table.

Sums of Mean of
Source of Variation dF F-Ratio
Squares Squares
Between Groups 324.79 3 108.26 18.96
Within Groups 114.17 20 5.71
Total 438.96 23

Step 5: State your decision rule.

Since the computed value of F (Fcomp=18.96) is greater than the critical value of F (Fcrit=3.10), Fcomp>Fcrit,
reject the null hypothesis.

Step 6: Determine the p-value for your conclusion.

Step 7: Interpretation:
The significant F-ratio reveals the rejection of the null hypothesis. There is a significant difference in the tensile
strength of the different hardwood concentrations. That is, there is an effect of the hardwood concentration in the
tensile strength of the product.

MULTIPLE COMPARISONS TEST (POST –HOC TEST)

The formula for the least significant difference is:

( )
√ ( )

where:
= critical value from the t-distribution table referred at dfw
MSw = mean square within, obtained from the results of your ANOVA test
n = number of scores used to calculate the means

( )

√ ( )

Treatments Mean Difference LSD Remarks

̅ ̅ 5.67 2.88 sd
̅ ̅ 7 2.88 sd
̅ ̅ 10.17 2.88 sd
̅ ̅ 1.33 2.88 nsd
̅ ̅ 4.5 2.88 sd
̅ ̅ 3.17 2.88 sd

Result of Multiple Comparisons Test

Hardwood Mean
Concentration Tensile
(%) Strength
5 10c
10 15.67b
15 17b
20 20.17a
F-test *
*The mean difference is statistically significant at 0.05 level.
Means of the same letter are not significantly different.
Review Problem 02: A researcher used different laboratory tests in an experiment involving volunteer patients, with the fast results, in
hours, given below. Test the hypothesis that the different laboratory test results has the same mean at the 0.01 level of significance.

Test A Test B Test C Totals

5 8 10
7 7 8
10 5 9
8 5 7
6 8 10
3 8
11
13
Sums 39 33 76 148
n 6 5 8 19
Mean 6.5 6.6 9.5 7.79
SS 283 227 748 1,258

Ho: The different laboratory test results have the same mean.
Ha: The different laboratory test results have different means.

ANOVA Summary Table

Sums of Mean of
Source of Variation dF F-Ratio
Squares Squares
Between Groups 40.46 2 20.23 5.01
Within Groups 64.70 16 4.04
Total 105.16 18

At =0.05, Fcrit(2,16) = 3.63.

Since the computed value of F (Fcomp=5.01) is greater than the critical value of F (Fcrit=3.63), Fcomp>Fcrit, reject the null
hypothesis.

Conclude that the different laboratory test results have different means.
(Perform post-hoc test: Fisher’s LSD)

( )

Comparisons |Mean Difference| LSD Remarks

nsd
A&B 0.1 √ ( ) (=)

sd
A&C 3.0 √ ( ) (≠)

sd
B&C 2.9 √ ( ) (≠)

Result of Multiple Comparisons Test

Laboratory Mean
Tests Lab Results
A 6.5b
B 6.6b
C 9.5a
F-test *
*The mean difference is statistically significant at 0.05 level.
Means of the same letter are not significantly different.