CHAPTER – III

RANDOMIZED BLOCK DESIGN

3.1 INTRODUCTION

Randomized block design is a experimental design method used to reduce variability in an

experiment by grouping experimental units into blocks based on some similarity or
characteristic, and then randomizing treatments within each block. This helps to account
for potential sources of variability that may be related to the blocking factor. It’s commonly
used in agricultural, medical, and industrial research settings to increase the precision and
efficiency of experiments.

In this chapter we see the petrol consumption of five bikes from each of four different bike
makes. The bike will be driven on five different days at five different speeds. The data will
be used to compare the petrol consumption rates of the different bike makes.

3.2 RANDOMIZED BLOCK DESIGN

The entire experimental material is divided into a number of blocks equal to the number of
replications for each treatment. Then each block is divided into a number of plots equal to
the number of treatments. For example if we have 4 treatments A,B,C and D and each
treatment is to be replicated 3 times. Then according to the condition of RBD, we will
arrange the experimental material in three blocks each of size 4, i.e. each block consists of
4 plots. It is also known as two way classification. The method is illustrated below by the
following arrangement of 3 block and 4 treatments

Block I A B D C
Block II C A D B
Block III D B C A
3.2.1 ADVANTAGES OF RANDOMIZED BLOCK DESIGN

 This design is more efficient or accurate than CRD

 It has less experimental error.
 This design is more flexible. i.e. no restriction are placed on the number od
treatments or the number of replications
 The statistical analysis for this design is simple and rapid.
 It is easily adaptable. In an agricultural experiment it can be accommodated well
in a rectangular or square field.

3.2.2 DISADVANTAGES OF RANDOMIZED BLOCK DESIGN

 If the number of treatments is very large, than the size of the blocks will increase
and this may introduce heterogeneity within blocks.
 If the interaction are large, the experiment may yield misleading results.

3.3 ANOVA TABLE FOR RANDOMIZED BLOCK DESIGN

Source of Degree of Sum of Mean sum of Variance Ratio (F-

variation freedom (d.f) square(SS) square (MS) ratio)
Between k-1 SS treat = R1 R1 MS treat
MS treat = F1 = F
k−1 MS error
treatments
Between blocks r-1 SS blocks = R2 R2 MS blocks
MS blocks = F2 =
r−1 MS error
F
Error (k-1) (r-1) SS error = R3 MS error =
R3
( k−1 ) (r−1)
Total Kr-1 SST

To use ANOVA table, it is convenient to use the following short-cut computational

formulae for finding the various sums of squares numerically.

1. Total sum of squares

 SST = ∑ ∑ y ij - correction factor
2

j i

2
G
Where correction factor = CF =
N
2. Between treatment Sum of Squares
2
T
R1 = SS treatments = ∑ j - C.F
j r
3. Between block sum of squares
2
Bi
R2 = SS blocks = ∑ - C.F
i k
4. Error sum of squares
R3 = SS error = SST- R1 – R2

After calculating the F – ratios (F 1 – F2), compare with the table values of F for the given
degree of freedom at α %.

If the calculated value F1 is less than the tabulated value, we accept our null hypothesis
(H01) and we conclude that there is no significant difference between the treatments.
Similarly, if the calculated value F 2 is less than the tabulated value, we accept our second
null hypothesis (H02) and we conclude that there is no significant difference between the
blocks.

3.4 ANALYSIS OF PETROL CONSUMPTION RATES OF THE

DIFFERENT MAKES OF BIKES

There were 4 different makes of bikes. A problem was posed to estimate the petrol
consumption rates of the different makes of bikes for suitable average speed and compare
them. The following experiment could be conducted for an inference about the problem.

Five different bikes of each four makes were chosen at random. The five bikes of each
makes were put on road on s different days. The bikes of A make run with different speeds
on different ways. The speeds were 20, 30, 40, 50 and 60 mph. Which bike was to put on
the road on which day and what speed it should have was determined through a chance
mechanism subject to the above conditions of the experiment. The procedure was adopted
for each of the Makes of bikes. For each bike, the number of miles covered per gallon of
petrol was observed. The observation are presented below:

Table: Miles per Gallon of Petrol

Makes of Speed of the bike in miles per hour(mph)

bikes 20 30 40 50 60
A 20.6 19.5 18.1 17.9 16.0
B 19.5 19.0 15.6 16.7 14.1
C 20.5 18.5 16.3 15.2 13.7
D 16.2 16.5 15.7 14.8 12.7

3.4.1 Calculation of ANOVA Table for RBD using F-test

Let us consider the makes of the bikes are the treatments, speed of the bikes in miles per
hour are the blocks and the specific bikes used, the effect of the days, drivers and possibly
some other effects contributed to the error variance.

Makes of Speed of the bikes in miles per hour (mph) Average

bike 20 30 40 50 60 Total
A 20.6 19.5 18.1 17.9 16.0 92.1 18.42
B 19.5 19.0 15.6 16.7 14.1 84.9 16.98
C 20.5 18.5 16.3 15.2 13.7 84.2 16.84
D 16.2 16.5 15.7 14.8 12.7 75.9 15.18
Total 76.8 73.5 65.7 64.6 56.5 337.1

2
(337.1)
 Correction factor (CF) =
20
113636.41
=
20
= 5681.82
 ∑ ∑ y 2ij = (20.6)2+ (19.5)2+ (18.1)2+ (17.9)2+ (16.0)2+ (19.5)2+ (19.0)2+
j i

2 2 2 2 2 2 2 2 2
(15.6) + (16.7) + (14.1) + (20.5) + (18.5) + (16.3) + (15.2) + (13.7) + (16.2) +
2 2 2 2
(16.5) + (15.7) + (14.8) + (12.7)

= 5779.37

 Total sum of squares (TSS) = ∑ ∑ y ij - CF

2

j i

= 5779.37 – 5681.82
= 97.55
 Sum of squares due to speed (SSS)
2 2 2 2 2
(76.8) +(73.5) +(65.7) +(64.6) +(56.5)
= – CF
4
= 5745.59 – 5681.82
= 63.77
 Sum of squares due to Makes (SSM)
2 2 2 2
(92.1) +(84.9) +(84.2) +(75.9)
= – CF
5
= 5708.17 – 5681.82
= 26.35
 Sum of squares due to Errors (SSE)
= TSS – SSS – SSM
= 97.55 – 63.77 – 26.35
= 7.43
3.4.2 TWO FACTOR ANOVA TABLE

Source of Degree of Sum of Mean of F - ratio F table value

variation freedom square square at 5% level
Speeds 4 63.77 15.94 25.75 F(4,12) =
3.26
Treatments 3 26.35 8.78 14.18 F(3,12) = 3.49
(Makes)
 Error 12 7.43 0.619
Total 19 97.55

In both the case either for speeds or for makes, calculated value of F is greater than
tabulated value of F at 5% level of significance and thus null hypothesis is rejected.

3.4.3 TREATMENT MEAN

Treatment A B C D
(Makes)
Mean of 18.42 16.98 16.84 15.18
treatments

20

18

16

14

12

10

0
A B C D

Comparing the four makes of bikes with each other in the bar chart mentioned above.
Comparing the four makes the make A is significantly better than all other Makes.

3.5 MISSING PLOTS TECHNIQUE IN RBD

In randomized block design, the missing plots technique is used when there are missing
observations within blocks due to unforeseen circumstances like equipment failure or
experimental error.

Suppose without loss of generality that observation for treatments 1 on block 1 i.e. y 11 is
missing and let it is Y, then the observations for RBD may be represented as below:

T1 T2 … Ti … Tp Total
B1 Y11 =Y Y21 … Yi1 … Yp1 B’1 + Y
B2 Y12 Y22 … Yi2 … Yp2 B2
… … … … … … … …
Bj Y 1j Y2j … Y ij … Y pj Bj
… … … … … … … …
Bq Y 1q Y2q … Y iq … Y pq Bq
Total T’1 + Y T2 … Ti … Tp G’ + Y

Where, B’1 = total of all available (p-1) observations in 1st block

T’1 = total of all available(q-1) observations in 1st treatment

G = total of all available (pq-1) observations.

p T 1+ q B 1−G
Missing estimation formula = Y =
( p−1 ) (q−1)

3.5.1 EXAMPLE FOR MISSING PLOT TECHNIQUE IN RBD

Example1.

In the following data one value are missing. Estimate the missing value and
analyses the data.
 Treatments Blocks
I II III
A 12 14 12
B 10 12 8
C Y 15 10
Solution:

let as calculate the missing value

Treatments Blocks Total

I II III
A 12 14 12 TA = 38
B 10 12 8 TB = 30
C Y 15 10 TC = 25 +Y
Total B1 = 22+Y B2 = 41 B3 = 30 G = 93 + Y

Where, p =3, q = 3, B1 = 22, TC = 25, G = 93

Applying the missing estimation formula

3× 25+3 ×22−93
Y=
4

75+66−93
=
4

48
=
4

= 12

Treatments Blocks Total

I II III
A 12 14 12 TA = 38
 B 10 12 8 TB = 30
C 12 15 10 TC = 37
Total B1 = 34 B2 = 41 B3 = 30 G = 105

2
(105)
 Correction factor (CF) =
9
11025
=
9
= 1225

 ∑ ∑ y 2ij = (12)2+ (14)2+ (12)2+ (10)2+ (12)2+ (8)2+ (12)2+ 2

(15) +(10)
2
j i

= 1261

 Total sum of squares (TSS) = ∑ ∑ y ij - CF

2

j i

= 1261 – 1225
= 36
 Treatment Sum of squares (SST)
2 2 2
(38) +(30) +(37)
= – CF
3
= 1237.67 – 1225
= 12.67
 Block Sum of squares (SSB)
2 2 2
(34) +(41) +(30)
= – CF
3
= 1245.67 – 1225
= 20.67
 Error Sum of squares (SSE)
= TSS – SST – SSB
= 36 – 12.67 – 20.67
= 2.66
TWO FACTOR ANOVA TABLE

Source of Degree of Sum of Mean of F - ratio F table value

variation freedom square square at 5% level
Treatments 2 12.67 6.34 4.77 F(2,6) = 9.55
Blocks 2 20.67 10.34 7.77 F(2,6) = 9.55
Error 2 2.66 1.33
Total 6 36

In case of both treatments and blocks, calculated value of F is less than tabulated
value of F at 5% level of significance, thus treatment and block means are not
significantly different.

3.6 CONCLUSION
We discussed the analysis of petrol consumption rates of the different makes of
bikes using Randomized block design and explore which factor caused more rate.
The uses of F-distribution test to check whether difference between the treatments
where significant. From the ANOVA table we seen the f-calculated value for speed
and the makes of bikes are 25.75 and 14.18. The section 3.4.3 we seen the treatment
mean. Therefore it is significant based on speed and makes at 5% level. And also
we seen the missing plots technique on Randomized block design.