Chapter One way ANOVA

For B. V. Sc. A. H.

for B. V. Sc. and A. H. 1

 Introduction
• 20 children are divided into 5 groups and they fed four diets.
The data are weight of children (in Kg), after being raised on
these diets. We wish to know whether the weights of children
are same for all four diets.
• To test whether the weights of children are same for all four
diet?
Diet 1 Diet 2 Diet 3 Diet 4
A 60 68 102 87
B 57 67 102 84
C 65 74 100 83
D 58 66 96 85
E 61 69 101 90
• Is ‘t’ test convenient for the test?
• Correct procedure is provided by the technique of analysis of
variance.
for B. V. Sc. and A. H. 2
 Contd…
• Prof. R. A. Fisher introduced the term ‘Analysis of
Variance’ (ANOVA)in 1920’s to deal with problems in
the analysis of agronomical data.
• In large sample test ratio of difference in sample
means to its standard error followed normal
distribution
• In small sample test ratio followed t-distribution
• Study of more than two samples at a time is not
possible using ‘t’ or ‘Z’ test.
• To study over the more than two samples at a time,
ANOVA (analysis of variance) is used.
for B. V. Sc. and A. H. 3
 Contd…
• The statistical technique used to compare means of
variations of more than two populations is called
ANOVA.
• Variation is based on two types
i. Variation within samples
ii. Variation between samples(group)
• If the variation within the samples and between the
samples do not differ from each other then they
belongs to same population.
• Large variation between and within samples indicates
that samples comes from divergent population.

for B. V. Sc. and A. H. 4

 Definition
• Analysis of Variance (ANOVA) is the “Separation of variance
ascribable to one group of causes from the variance ascribable to
the other group.” Prof. R. A. Fisher
• Note:
• F- test is widely used in the ANOVA.
• It is a Statistical technique used to evaluate the variances between
three or more sample means.
• The test of differences between more than two sample means is
also accomplished with F- distribution.
• The null hypothesis is taken as all means are equal while the
alternative hypothesis is not all of them are equal, or at least two of
them are different.
• The table showing the source of variance, the sum of squares,
degrees of freedom, mean square (variance) and the formula for the
F- ratio is known as ANOVA table.
for B. V. Sc. and A. H. 5
 Assumptions
•   Normality: all samples are drawn form normally distributed
1.
populations.
2. Homogeneity of variances: all populations have common
variance.
3. Independences: all samples are independently of each other.

4. It has no significant outlier.

5. All the samples must be randomly selected and
independently.
6. Factors affects are additive.
7. i.e. Variance ratio must be greater than 1.
8. 0 ≤ F ≤ ∞ i.e. the value of F must lies in between zero to
infinity.

for B. V. Sc. and A. H. 6

 Contd…
• Note:
  F- Statistics is the ratio obtained by
dividing the variance between samples by the
variances within samples is given by:
• .
• One way ANOVA and Two Way ANOVA
• In one way classification observations are
classified into only one criteria.
• In two way classification observations are
classified into two way criteria.

for B. V. Sc. and A. H. 7

 ANOVA in One-Way Classification
Class Sample Observation Total Mean
1 X11, X12,…, T1
2 X21, X22,…, T2
…. … … …
i Xi1, Xi2,…,
…. … … …
k Xk1, Xk2,…, Tk

•• ‘n’ = sample observations of a random variable X are divided into ‘k’

 classes( n = n + n +…+n = )
1 2 k

• Xij = jth member of the ith class; (j = 1, 2, …, ni & i = 1, 2,…, k) &

• Note: In one-way classification, the effect of any one factor is
consideration.

for B. V. Sc. and A. H. 8

 Model

•  
• Where
• = the observation obtained from the jth plot to
which ith treatment has been applied.
• = the means of all the ‘k’ population means is
termed as ‘Grand Mean’.
• =represents the difference between the mean of
the j th population and the grand mean termed as
‘treatment effect’.
• =represents the amount by which an individual
measurement differs from the mean of the
population to which it belongs is termed as ‘error
term’.
for B. V. Sc. and A. H. 9
 Steps of Calculation(by short cut method)
Step 1: Formulate null and alternative hypothesis.
• Null Hypothesis: (H₀): μ₁= μ₂ =….= μk. That is the
arithmetic means of population from which ‘k’
samples are drawn are all equal to one another.
Or there is no significant difference of ‘k’
population means.
• Alternative Hypothesis: (H₁): μ₁≠μ₂≠……≠μk .That is
the arithmetic means of population from which
‘k’ samples are drawn are not all equal to one
another. Or at least one of the population mean
is different significantly from the other means.
for B. V. Sc. and A. H. 10
 Contd…
•   2: Compute the test statistic ‘F’. For this, we use the
Step
following procedures:
i. Find the sum of all the observations (G).
ii. Calculate the correction factor = (G2 ÷ N)
iii. Find the sum of square of all the observations.
iv. Find out the total sum of squares (SST) = (sum of squares of
all the observations) - (G2 ÷ N)
v. Find out the sum of squares between the samples (SSC) by
subtracting (G2 ÷ N) from the sum of the squares of the totals
divided by the numbers of items in each samples. (SSC) =
vi. Find the sum of squares within sample (SSE) by using SSE =
SST – SSC.
vii. Prepare ANOVA table to compute F ratio.
for B. V. Sc. and A. H. 11
 ANOVA table
Sources of Sum of squares d. f. Mean sum F- ratio
variations of squares
Between SSC K –1 MSC = SSC/K – 1 F = MSC/MSE
samples
Within SSE = SST – SSC N–K MSE = SSE/N – K
samples
Total SST N –1    

 Step 3: Fix the level of significance 1% or 5%

Step 4: Degrees of freedom = (K – 1, N – K)
Step 5: Write down the critical value of F for (K – 1, N – K) d. f.
at α% level of significance.
Step 6: Make decision.
(i) If F(cal) F(tab) then accept H0.
(ii) If F(cal) > F(tab) then reject H0 then accept H1.

for B. V. Sc. and A. H. 12

 Example
• To test the hypothesis that the average
number of days a patient is kept in the three
local hospitals say, A, B & C is the same,
random check on the number of days that
seven patients stayed in each hospital reveals
the following. Test the hypothesis by an
ANOVA at α = 0.05.

Hospital A 8 5 9 2 7 8 2
Hospital B 4 3 8 7 7 1 5
Hospital C 1 4 9 8 7 2 3

for B. V. Sc. and A. H. 13

 Solution
• Let
  X1j, X2j & X3j denote the number of days jth
patient stays in the hospitals A, B & C.
• H0: i. e. There is no significant difference that
the average number of days a patient is kept
in the three local hospitals say: A, B & C.
• H1: There is significant difference that the
average number of days a patient is kept in
the three local hospitals say: A, B & C.

for B. V. Sc. and A. H. 14

 Contd…

X1j X2j X3j

8 4 1
5 3 4
9 8 9
2 7 8
7 7 7
8 1 2
2 5 3
∑ X1j = 41 ∑ X2j = 35 ∑ X3j = 34

for B. V. Sc. and A. H. 15

 Contd…
• Sum
  of all observations (G) = ∑ X1j +∑ X2j+ ∑ X3j =
41 + 35 + 34 = 110
• Correction factor (C. F.) = G2 / N = 1102 / 21 =
576.2
• Total sum of squares (SST) = sum of square of all
observations – C. F. = 82 + 52 + 92 + 22 + 72 + 82 +
22 + 4 2 + 3 2 + 8 2 + 7 2 + 7 2 + 1 2 + 5 2 + 1 2 + 4 2 + 9 2 +
82 + 72 + 22 + 32 – 576.2 = 728 – 576.2 = 151.8
• Sum of squares between samples (SSC) =
•
for B. V. Sc. and A. H. 16
 Contd…
•  Sum of squares within samples (SSE) = SST –
SSC = 151.8 – 4.085 = 147.715
• Mean sum of squares between samples (MSC)
= SSC / k – 1= 4.085/ 3 – 1 = 4.085 / 2 = 2.0425
• Mean sum of squares between samples (MSE)
= SSE / N – k = 147.715 / 21 – 3 = 147.715 / 18
= MSE = 8.206
• d. f. = (k – 1, N – k) = (3 – 1, 21 – 3) = (2, 18)

for B. V. Sc. and A. H. 17

 ANOVA Table
Sources Sum of d. f. Mean sum F- Ratio
of squares of squares
variation
Between SSC = 4.085 K – 1 = 3 – 1 = MSC = F=
samples 2 2.0425 MSC/MSE
Within SSE = N – k = 21 – 3 = MSE = = 2.0425 /
samples 147.715 18 8.206 8.206
= 0.2489
Totals SST = 151.8      

Decisions: F(calculated) = 0.2489 < F(tabulated (2, 18) d. f. and α = 0.05 ) = 3.55,
then it is not significant. Therefore, we accept H0.
Conclusion: There is no significant difference in the average stay at each of the
three hospitals.

for B. V. Sc. and A. H. 18

THANK YOU

for B. V. Sc. and A. H. 19