CHAPTER 12

INTRODUCTION TO
ANALYSIS OF VARIANCE
(ANOVA)
1. SAIFUL BAHARI BIN ABD RAHMAN
45615
2. NOR AMIRAH BINTI THALATHA GS
3. SARAWATHY CHANDIRAN
GS
4. NURUL ATIKAH BINTI MISSNAN GS

GS
45557
46065
45647

12.1 Introduction to Analysis

of Variance

Analysis of variance is a
Hypothesis-testing procedure

Used to evaluate mean differences between

two or more treatments
Uses sample data as basis for drawing
general conclusions about populations

Clear advantage over a t- test: it

can be used to compare more than
two treatments at the same time
This is an extension of the two
independent sample t-test

Figure 12.1 :Typical Situation

for Using ANOVA

Terminology

Factor

Levels

The independent (or quasi-independent)

variable that designates the groups being
compared
Individual conditions or values that make
up
a factor

Factorial design

A study that combines two or more factors

Sources of Variability Between Treatments

Factor 2

Factor 1

Hypothetical data
From an experiment examining driving
performance under three telephone condition

Between treatment

Within
treatment

Statistical Hypotheses for ANOVA

Null hypothesis: the level or
value on the factor does not
affect the dependent variable
In the population, this is equivalent to
saying that the means of the groups do
not differ from each other.
The means are all equal.

H 0 : 1 2 3

Alternate Hypothesis for

ANOVA

H1: There is at least one mean

difference
among the populations
(Acceptable shorthand is Not
H0)
Issue: how many ways can H0 be
wrong?

All means are different from every other

mean

Test statistic for ANOVA

F-ratio is based on variance
instead of sample mean
differences
variance(differences) between sample means
F
variance(differences) expected with no treatment effect

Test statistic for

ANOVA
Not possible to compute a
sample mean difference
between more than two
samples
F-ratio based on variance
instead of sample mean
difference

Variance used to define and measure

the size of differences among sample
means (numerator)

Type I Errors and MultipleHypothesis tests

Why ANOVA (if t can compare two

means)?

Experiments often require multiple hypothesis

testseach with Type I error (testwise alpha)
Type I error for a set of tests accumulates testwise
alpha experimentwise alpha > testwise alpha

ANOVA evaluates all mean differences

simultaneously with one test
regardless of the number of means
and thereby avoids the problem of
inflated experiment wise alpha

12.2 Analysis of
Variance Logic

Between-treatments variance - SS(B)

Variability results from general differences
between the treatment conditions
Variance between treatments measures
differences among sample means
Some of square (variation) between the
group.
Within-treatments variance SS(W)
Variability within each sample
Individual scores are not the same within
each sample

Sources of Variability
Between Treatments

Systematic differences caused

by treatments
Random, unsystematic
differences
Individual differences
Experimental (measurement) error

Sources of Variability
Within Treatments

No systematic differences related

to treatment groups occur within
each group
Random, unsystematic differences

Individual differences
Experimental (measurement) error

differences including any treatment effects

F
differences with no treatment effects

12.3 ANOVA NOTATION

AND FORMULAS
NOR AMIRAH BINTI THALATHA
GS 45557

ANOVA NOTATION

k = Number of treatment conditions / number of separate samples

N = Total number of score in the entire study
G = Sum of all score

THE STRUCTURE AND SEQUENCE OF

CALCULATIONS FOR THE ANOVA

Anova Formula
Variance between treatments

F=

Variance within treatments

Sample variance = =

1. Total Sum Of Squares, SS.

(To make it consistent with ANOVA Notation)

= 106
= 46

2. Within-Treatments Sum Of Squares, SS.

= 6+6+4
= 16

3. Between-Treatments Sum Of Squares

= = 46-16
= 30

=n
= 5 (4+1+1)
= 5 (6)
= 30

Degree Of Freedom (df)

1. Total Degree Of Freedom,.

= 15 - 1
= 14

2. Within-Treatment Degree Of Freedom,.

= (5-1) + (5-1) + (5-1)

= 12

= 12

3. Between-Treatment Degree Of Freedom

=k-1

= 3-1
=2

Calculation of variance and the F-ratio

Variance Between Treatments

=
= 15

Variance Within Treatments

=
= 1.33

Variance between treatments

F=

F=

= 11.28

Variance within treatments

ANOVA SUMMARY TABLES

Source
Between treatments
Within treatments
Total

SS
30
16
46

df
2
12
14

MS
15
1.33

F = 11.28

12.4. THE DISTRIBUTION OF F-RATIOS

1.The numerator and denominator of the ratio are measuring

exactly the same variance when the null hypothesis is true.
F= systematic treatment effects+random,unsystematic differences
random,unsystematic differences
(the value obtained for the F-ratio helps to determine whether any
treatments effects exist.)

2. For eg. We expect F to be around 1.00, if we obtain an F-ratio

that is >1.00, then it is evidence that a treatment effects exists
and the null hypothesis is false.

3. The problem now is to define precisely which values are

around 1.00 and which are much greater than 1.00.
4. So, we need to have a look at all possible F values and H0
is true = that is the distribution of F-ratios.
Note two characteristics:
a) Because F-ratios are computed from two variances (the
numerator and denominator of the ratio) f -values are always
positive
b) When H0 is true , the numerator and denominator of the fratio are measuring the same variance. In this case, two sample
variances should be about the same size, so the ratio should be
near 1. which is the distribution should be pile up around 1.00.

F DISTRIBUTION TABLE

EXAMPLE:

For the experiment we have been considering , the numerator of F

ratio (between treatments ) has df=2 and the denominator of f-ratio
(within treatments) has df=12.So, the degrees of freedom equal to
2 and 12 which is written as 2,12.
Step 1: find df=2 across the top of the table and df=12 in the first
column.
Step 2: when line up this two, they point to a pair of numbers in the
middle

Step 3: This number give critical cut off for = 0.05 and
=0.01

12.5 EXAMPLES OF HYPOTHESIS TESTING

AND EFFECT SIZE WITH ANOVA
Quiz scores for students using four different study strategies
Read once

Read and
reread

Answer
prepared
question

Create and answer

question

N=24

G=168

10

X2=1298

10

10

10

T=30
T=36
T=48
T=54
STEP1:
State theM=6
hypotheses andM=8
select alpha
level
M=5
M=9
SS=20
SS=14 effect)
SS=8
H 1= 2= 3=SS=20
4(there is no treatment
0:

H1: Atleast one of the treatment means is different.

We use =0.01
STEP 2: LOCATE CRITICAL REGION
We must determine the degrees of freedom for MS(between treatments)
MS(within treatments) which is the numerator and the denominator of F-ratio
df (total)= N-1 =24-1
Analyze into two components
df (between)= k-1=4-1=3
df (within)=df (inside each
treatment)=5+5+5+5=20
SO, The F-ratio for this data has df=3,20

STEP 3: COMPUTE THE F-RATIO

a. analyzed the SS to obtain SS(between) and SS(within).
b. Use the SS values and the df values(frm step 2) to calculate two
variances MS(between), MS(within)
c. Finally , use two MS value(the variance )to compute the F-Ratios.

STEP4: MAKE DECISION

The F value obtained is 6.45 is in critical region and it is very
unlikely (p<.01) that we obtain value this large if H0 is true.
Therefore we reject H0 and conclude that there is significant
treatment effect.

MEASURING EFFECT
SIZE FOR ANOVA
Example : SOURCE

SS

df

MS

Between
Treatment

60

20.00 F=6.45

Within
Treatment

62

20

3.10

TOTAL

122

23

The percentage of variance accounted for= SS (between treatments)

SS(total)
=60/122 = 0.492
The percentage of variance accounted for by the treatment effect is
usually called 2 . Thus , the 2 for the study is 0.492.

AN EXAMPLE WITH
UNEQUAL SAMPLE SIZES
A researcher is interested in the amount of homework
required by different academic majors. Students are
recruited from Biology, English and Psychology to
participate in the study. The researcher randomly selects one
course that each student is currently taking and asks the
student to record the amount of out of class work required
each week for the course. The researcher used all the
volunteer participants , which resulted in unequal sample
sizes. The data are summarized in Table 12.5.

Finally , compute the MS values and the F-ratio

MS(between)= SS/df = 65/2 = 32.5
MS(within) = SS/ df = 187/17 =11
F= MS(between)/MS(within) = 32.5/11 = 2.95
STEP 4: MAKE DECISION
Because the obtained F-ratio is not in the critical region, we
fail to reject the null hypothesis and conclude that there is
no significant differences among the three population.

12.6 : POST HOC TEST

Post hoc test (posttest) are additional

hypothesis tests that are done after an ANOVA
To determine exactly which mean differences
are significant and which are not.
Pairwise comparisons.
The tests are done after ANOVA when:
1.
2.

Reject H0 and
There are three or more treatments (k

Rejecting H0 indicates that at least one

difference exists among the treatment.

3)

TURKEYS HONESTLY SIGNIFICANT

DIFFERENCE (HSD) TEST

Used to compare any two treatment

conditions
If the mean difference exceed Turkeys
HSD, there is a significant difference
between the treatment,

Table 12.6 (page 376)

Treatment A

Treatment B

Treatment c

n=9

n=9

n=9

T = 27

T = 49

T = 63

M = 3.00

M = 5.44

M = 7.00

Source

SS

df

MS

Between

73.19

36.00

Within

96.00

24

4.00

Total

169.19

26

Overall F(2,24) = 9.15

k = 3, dfwithin = 24, = .05 and q = 3.53

1. Treatment A is significantly different from

treatment B
(MA MB = 2.44)
2. Treatment A is also significantly different from
treatment C (MA MC = 4.00)
3. Treatment B is not significantly different from
treatment C (MB MC = 1.56)

THE SCHEFF TEST

Use F Ratio to evaluate the significance

of the difference between any two
treatment condition.

FA versus B = MS between
MSWithin

df(2,24), = .05, F value is

3.40

Treatment A versus B : F (2,24) = 3.36

# there is no significant difference
between treatment A and treatment B
Treatment A versus C : F (2,24) = 9.00
# there is a significant difference
between treatment A and treatment C
Treatment B versus C : F (2,24) = 1.36
# there is no significant difference
between treatment B and treatment C

12.7 RELATIONSHIP BETWEEN ANOVA

AND t TESTS

T test or F can be used to evaluating two

independent samples.

F = t

Testing the same hypothesis

Ho : 1 = 2
H 1 : 1 2

Figure 12.10 (page 380)

The distribution
of
t
statistics
with df = 18 and
the
responding
distribution of
F-ratios with df =
1,18. Notice that
the
critical
values for