UNIT II

Design of Experiments
ANOVA
 General principles – Analysis of variance
(ANOVA) - One way classification -
Completely randomized design (CRD) – Two
way classification - Randomized block design
(RBD) – Three way classification -Latin square
design(LSD) – Two factor experiments: 2 2
factorial design.
Definition

Design of experiments is a systematic

method to determine the relationship
between factors affecting a process and
output of that process
 Experiment:
 - device of getting an answer to the
problem under considerations.

 Treatments:
- the objects(factors) of comparison in
an experimental design
 E.g. different fertilizers, crops
 Experimental error:
 -the variations from plot to plot

caused by uncontrolled factors

Experimental Unit:
- the smallest division of the
experimental material to which we apply
the treatments.
e.g, plot of lands

* Block:
- Homogeneous subgroups or layers of
the whole experimental Unit
Basic principles of design of
Experiments
 To reduce the experimental error we
adopt certain principles known as
basic principles of experimental
design.

 Randomization
 Replication
 Local control
Randomization


When all the treatments have
equal chance of being allocated to
different experimental units it is
known as randomization.
Replication

 Repeated application of the

treatments is known as replication.

 Repetition
of the treatments
under investigation.
Local control

 The process of reducing the

experimental error by grouping the
homogenous experimental units into
blocks is known as local control of
error.
ANOVA – Analysis of Variance
 If there are more than two populations, for
testing the equality of their means the
analysis of variance method is applied

 It is a statistical technique specially designed

to test whether the means of more than two
quantitative populations are equal.
ANOVA
 It is the separation of variance
ascribable to one group of causes
from the variance ascribable to other
groups. – R. A. Fisher

 It is achieved by separating total

variation into 2 sections
 i) variation between the samples
 ii) variation within samples
ANOVA
 Variance ratio or F- ratio

F =
Assumptions for ANOVA
 The observations are independent
 The populations are normally distributed
 The variances of all populations are equal
 The samples have been drawn from the

normal population.
 Notations:
 N – Total number of elements
 c – number of columns
 r – number of rows
 T – grand total
 C – correction factor
 TSS – total sum of squares
 SSC – sum of squares between columns
 SSR - sum of squares between rows
 SST - sum of squares between treatments
 SSE - sum of squares between errors
 MSC – mean sum of squares between columns
 MSR - mean sum of squares between rows
 MST – mean sum of squares between treatments
 MSE - mean sum of squares between errors
Some Important Experimental
designs

 Completely Randomized Design(CRD).

 Randomized Block Design(RBD).
 Latin Square Design(LSD).
 Two square factorial Design.
Completely Randomized Design
(C.R.D) or ANOVA for One-way
Classification
 Null hypothesis:H0:There is no significance
difference between columns.

 Alternative hypothesis:H1 :There is a

significance difference between columns.
 Calculation Table

Factors X1 X2 X3 Grand

Total

. . .

. . .

. . .

Column C1 = C2 = C3 = T=

total Ci

=
Calculation Steps
 c = Number of columns,
number of elements in first column ,
, ,… N = total number of elements
 Grand total T =
 Correction factor C =
 TSS = - C
 SSC = - C = + + – C
 SSE = TSS - SSC
ANOVA Table (CRD)
Source of Degrees Sum of Mean Variance Table
variation of squares Squares Ratio value
freedo (SS) (MS) (F-ratio)
m (d.f)
Between c -1 SSC MSC = F=
Columns
(Between Nr > Dr
Samples)
Between Error N - c SSE MSE =
(Within in
Samples)
Problem:1
 The following are the number of mistakes
made in 5 successive days by 4 technicians
working for a photographic laboratory .Test
whether the difference among the four
samples means can be attributed to
chance(test at a level of significance
alpha=0.01)
 Technician

I II III IV
6 14 10 9
14 9 12 12
10 12 7 8
8 10 15 10
11 14 11 11
Problem:2
 A completely randomized design experiment
with 10 plots and 3 treatments give the
following results: Analyse the results of
treatment effects

Plot NO 1 2 3 4 5 6 7 8 9 10

Treatment A B C A C C A B A B

Yield 5 4 3 7 5 1 3 4 1 7
 A completely randomized design experiment
with 10 plots and 3 treatments gives the
following results:

Treatments
A B C
5 4 3
Replication

7 4 5
3 7 1
1
Randomized Block Design (RBD) (or)
Two-way Classification
 Null hypothesis:H0: There is no significance
difference between rows.
 There is no significance difference between

columns.
 Alternative hypothesis:H : There is a
1
significance difference between rows.
 There is a significance difference between

columns.
 Calculation Table

Factor 1 -> X Row

Factor2 A B C D Total
Y X1 X2 X3 X4 Ri

Y1 R1 =
Y2 R2 =
Y3 R3 =
Column C1 = C2 = C3 = C4 = T=
total Ci
Calculation Steps
 c = Number of columns
 r = Number of rows

N = r c = number of elements
 Grand total T =
 Correction factor C =
 TSS = - C
 SSC = - C = + + + – C
 SSR = - C = + + –C
 SSE = TSS – SSC - SSR
ANOVA Table (RBD)
Source of Degrees Sum of Mean Variance Table value
variation of squares Squares Ratio
freedom (SS) (MS) (F-ratio)
(d.f)
Between c -1 SSC MSC =
Columns
Fc =
Between r-1 SSR MSR =
Rows

Between (c-1)(r-1) SSE MSE = Fr =

Error
Nr > Dr
Problem: A company appoints 4 salesman A, B, C
and D and observes their sales in 3 seasons
summer, winter and monsoon. The figure are
given in the following table: Carry out an analysis
of variance.
Season Salesmen
A B C D

Summer 45 40 28 37

Winter 43 41 45 38

monsoon 39 39 43 41
Problem: Three varieties of coal were analyzed
by 4 chemists and the ash content is tabulated
here. Perform an analysis of variance.

Coal Chemists
A B C D

I 8 5 5 7

II 7 6 4 4

III 3 6 5 4
Latin Square Design
 Null hypothesis:H0: There is no significance
difference between rows.
 There is no significance difference between

columns.
 There is no significance difference between

treatments.
 Alternative hypothesis:H : There is a
1
significance difference between rows.
 There is a significance difference between

columns.
 There is a significance difference between

treatments.
 Calculation Table - I

Factor 1 - X Row
Factor2 X1 X2 X3 X4 Total
-Y Ri

Y1 R1 =
Y2 R2 =
Y3 R3 =
Y4 R4 =
Column C1 = C2 = C3 = C4 = T=
total C
 Calculation Table - II

D
Treatments A B C

. . .

. . .

. . .

Column T1 = T2 = T3 = T4 =

total Ti
Calculation steps
 n = no. of columns = no. of rows = no. of
treatments
 N = total number of elements
 Grand total T =
 Correction factor C =
 TSS = - C
 SSC = - C = + + + – C
 SSR = - C = + + +– C
 SST = - C = + + + –C
 SSE = TSS – SSC – SSR-SST
 ANOVA Table (LSD)
Source of Degrees Sum of Mean Squares Variance Table
variation of squares (MS) Ratio (F- value
freedom (SS) ratio)
(d.f)
Between n -1 SSC
Columns Fc
Between n-1 SSR
Rows Fr
Between n-1 SST
treatnents Ft
Between (n-1)(n-2) SSE
Problem: Analyze the variance in the Latin square of
yields(in kgs) paddy where P, Q, R and S denote the
different methods of cultivation. Examine whether the
different methods of cultivation have given significantly
different yields.

S 122 P 121 R 123 Q 122

Q 124 R 123 P 122 S 125

P 120 Q 119 S 120 R 121

R 122 S 123 Q 121 P 122

Choose origin as 122
x-122

S 0 P -1 R 1 Q 0

Q 2 R 1 P 0 S 3

P -2 Q -3 S -2 R -1

R 0 S 1 Q -1 P 0
 Table
value

12

r= replication number,