Analysis of Variance

(ANOVA)

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science slide 1

 Why ANOVA?
To compare differences between group means
Example: Is there any significant difference in satisfaction
between users of the following three computer operating
systems

The variables Type Scale

Satisfaction Dependent Interval
Operating systems Independent Nominal
‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 2
 Basis for the comparison
Total variance which is partitioned into Between
and Within-group

Between group

Total Variance

Within group

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 3
 Which of the following data sets indicate differences between the
two groups?

Data 1: Data 2:
Group Group
A B A B
30 21 13 25
14 28 12 26
26 14 15 28
15 31 14 27
27 16 15 26
13 12 12 30
25 27 17 29
Range 17 19 Range 5 5
Mean 21.4 21.3 Mean 14 27.3
‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 4
 Assumptions Underlying
One-way ANOVA
1. The dependent variable is normally distributed for each
of the populations as defined by the different levels of
the factor (independent variable)
2. The variances of the dependent variable are the same for
all populations
3. The cases represent random samples from the
populations and the scores on the test variable are
independent of each other

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 5
 What to expect?

Post hoc
comparison

re n ce
i f fe
Sig .d
Test Omega and
hypothesis Eta squares
No
s ig
. di
ffe
r en
ce

‫ﺒﻬﺎﻣن‬
z
bas statistics for Social Science
slide 6
 Steps in ANOVA
1. State the null and alternative hypotheses
HO: µ1 = µ2 = µ3
HA: Not all means are equal
2. Calculate the test statistic
F-ratio
3. Determine critical value
4. Make your decision
5. Make conclusion

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 7
 Step 1: State the hypotheses
HO: µ1 = µ2 = µ3 = ……..
HA: Not all means are equal

Depends on number of
groups

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 8
 Step 2: Calculate test statistic
1. Calculate sum of squares

a. SST = ∑ X 2
−
(∑X)
2

b. SSB = ∑
T j2
−
(∑X)
2

n N
c. SSW = SST − SSB

2. Determine degrees of freedom

a. df B = k − 1
b. dfW = N − k
c. dfT = N − 1

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 9
 3. Calculate Mean Squares 4. Calculate F-ratio
SSB
a. MSB = F=
MSB
df B MSW
SSW
b. MSW =
dfW

Summary ANOVA table:

Source SS df MS F
Between-group SSB k-1 MSB F-ratio
Within-group SSW N-k MSW
Total SST N-1

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 10
 Step 3: Determine critical value
Critical value is based on:
n Significance level (α)
o Degrees of freedom Fdfdfwb = FNk −−1k

Step 4: Decision
Manual:
Reject HO: Fcal > Fcritical
Fail to reject HO: Fcal ≤ Fcritical
SPSS:
Reject HO: sig-F < α
Fail to reject HO: sig-F ≥ α

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 11
 Step 5: Conclusion
Reject HO:
Significant difference between the groups
Fail to reject HO:
No significant difference between the groups

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 12
Additional Analysis I:

Post-hoc Comparison
Tukey’s HSD (Honestly Significant Difference)

MSW
HSD = qk , dfW , α
n

Criteria to determine difference:

Sig difference: Mean difference > HSD
No sig. difference: Mean difference ≤ HSD

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 13
Additional Analysis II:

Omega Square
Amount of variance of the dependent variable attributed
to the independent variables

df B ( F − 1)
ωˆ 2 =
df B (F − 1) + N

Eta Square
Proportion of variance of dependent variable that is
attributed to the independent variable
SSB
η2 =
SST

‫ﺒﻬﺎﻣن‬

bas statistics for Social Science

slide 14