Anova Post Hoc
Anova Post Hoc
It is used when we wish to test the significance of the differences between three
or more means obtained from independent samples.
The One-Way F-test or One-Way ANOVA is called such because there is
only one factor being studied as the independent variable.
Sample Problem
Suppose a teacher aims to test the hypothesis that academic performance varies
from IQ Level. The following academic grades are:
Above Average: 90, 92, 89, 95, 93
Average: 85, 87, 83, 84, 80
Below Average: 78, 75, 77, 79, 76
At a .05 level of significance, test if the mean differences between the
three groups are significant.
Procedures:
A. Identify the Variables
Independent Variable: IQ Level (Above Average, Average, and Below Average)
Dependent Variable: Academic Performance
E. Computations
Step 1. Determine the following:
a. k = 3 (there are 3 groups in the research problem)
b. n1 = 5 (there are 5 samples in group 1-Above average IQ level)
n2 = 5 (there are 5 samples in group 2-Average IQ level)
n3 = 5 (there are 5 samples in group 3-Below average IQ level)
c. N = n1 + n2 + n3
=5+5+5
= 15
2
Step 2. Compute the mean for each group.
b. Compute the Sum of Square for each group using this formula:
3
c. Compute the Sum of Square within group (SSw) using this formula:
e. Compute the Sum of Square between group (SSb) using this formula:
4
Step 7. Compute the F-ratio
5
Note: In the F-critical table, the degree of freedom in the numerator is the degree of
freedom between groups (dfb), while the degree of freedom in the denominator
is the degree of freedom within groups (dfw).
6
7
8
Decision:
Since the computed value = 55.21 is greater than the tabular
value = 3.89, the Null Hypothesis is rejected and the Alternative
Hypothesis is accepted.
NOTE!
1. If the result of the Analysis of Variance (ANOVA) is not
significant (accept Ho and reject Ha, immediately make an
implication (Procedure G).
2. If the result of the Analysis of Variance (ANOVA) is significant
(reject Ho and accept Ha, DO NOT make an implication;
instead, proceed to Test for Multiple Comparison of Means
(Module 6).
Sample Problem
9
Suppose a teacher aims to test the hypothesis that academic
performance varies with IQ Level. The following academic grades are:
Above Average: 90, 92, 89, 95, 93
Average: 85, 87, 83, 84, 80
Below Average: 78, 75, 77, 79, 76
At a .05 level of significance, test if the mean differences of the
three groups are significant.
Results
Decision:
Since the computed value = 55.21 is greater than the tabular value
= 3.89, the Null Hypothesis is rejected and the Alternative Hypothesis is
accepted. Thus, proceed to Test for Multiple Comparisons of Means
(Post-Hoc Test).
10
Based on the ANOVA findings, the overall difference among the three
groups is significant different. Our task is to find which specific pair has a
significant difference. Hence, we can come up with these hypotheses.
1. Ho: 𝑋̅1 = 𝑋̅2; there is no significant difference between group 1 and
group 2 means.
Ha: 𝑋̅1 ≠ 𝑋̅2; there is a significant difference between group 1 and
group 2.
2. Ho: 𝑋̅2 = 𝑋̅3; there is no significant difference between group 2 and
group 3 means.
Ha: 𝑋̅2 ≠ 𝑋̅3, there is a significant difference between group 2 and
group 3.
3. Ho: 𝑋̅1= 𝑋̅3; there is no significant difference between group 1 and
group 3 means.
Ha: 𝑋̅1≠ 𝑋̅3; there is a significant difference between group 1 and group
3 means.
11
Procedure for the Scheffé Test
1. Recall the following values in the Sample Problem in Module 5.
12
b. X ̅2 vs. X ̅3 (Average versus Below Average)
13
4. Decision
Decision Rules:
1. If the F-observed (Fo) value is less than the F-critical (Fc) value, it fails
to reject Ho (Null Hypothesis) and reject the Ha (Alternative Hypothesis).
2. If the F-observed (Fo) value is greater than the F-critical (Fc) value,
reject the Ho (Null Hypothesis)
Decisions:
̅1 vs. X
1. For X ̅2 (Above average versus Average), since the F-
observed (Fo) =32.16 is greater than the F-critical (Fc) value =
7.78, Ho is rejected.
̅2 vs. X
2. For X ̅3 (Average versus Below Average), since the F-
observed (Fo) =23.24 is greater than the F-critical (Fc) value =
7.78, Ho is rejected.
̅1 vs. X
3. For X ̅3 (Above Average versus Below Average), since the
F-observed (Fo) =110.07 is greater than the F-critical (Fc) value
= 7.78, Ho is rejected
5. Implication
14
levels is significantly higher compared to students who have
below-average IQ levels.
3. The academic performance of students who have above-
average IQ levels is significantly higher compared to students
who have below-average IQ levels.
15