Topic

5: Latin Squares Design

- We have seen before that randomized complete block design (RCBD) is used toreduce
experimental errors. There are other designs that use the blocking principal.
- For example, suppose an experimenter studying the effects of different formulation of an
explosive mixture use in the manufacturing, each formulation is mix from different
batches of raw material. Furthermore, the formulation are prepared by several operators
and there may be substantial different in the skills and experience of the operators.
- Thus, it would seem that there are 2 extraneous factors which will affect the experimental
results, i.e. batches of raw material and operators.
- Appropriate design for this problem consists of testing each formulation exactly once in
each batch of raw material and for each formulation to be prepared exactly once by each
of five operators.
- Resulting design called Latin Square Design where the design is a square arrangement.
- Suppose we want to see the effect of 4 different methods on the employees’ productivity.
3 different days and 3 operators have an effect on the employees’ productivity.
1 2 3
Monday B A C
Tuesday A C B
Wednesday C B A

Each treatment appears once and only once in each row and column.

Latin Squares Design has following features:
- We are certain that there exist 2 extraneous factors other than the treatments that
will affect the experimental error.
- Blocking variables in two directions represented by rows and columns.
- Each treatment appears once and only once in each row and column.

Randomisation of Latin Square Design
- Yates have prepared standard Latin Squares Design.
- A Latin Squares in which the first row and the first column consist of letters written
in alphabetical order are called Latin Standard Squares Design.
- Randomisation employed as follows:
- For t=3, independently arrange rows and columns at random
- For t=4, select one of the standard square at random, then independently arrange
its rows and columns at random.
- For t=5 and higher, independently arrange rows, column and treatment of the
given standard square at random.
- Example, for t=3
Step 1: start with 3 by 3 Standard Latin Square
A B C
B C A
C A B

1
 Step 2: Rearrange the rows independently
- Obtain 3 random numbers
Order number: 1 2 3
Random number: 145 116 614

- Arrange the random number in ascending order, carrying along the original order.
Random number: 116 145 614
Order number: 2 1 3

- Rearrange the rows independently.
Original Arrangement Present Arrangement
1 2 3 Row 1 2 3
1 A B C 2 B C A
2 B C A 1 A B C
3 C A B 3 C A B

Step 3: Rearrange the column independently
- Obtain 3 random numbers
Order number: 1 2 3
Random number: 985 212 168

- Arrange the random number in ascending order, carrying along the original order.
Random number: 168 212 985
Order number: 3 2 1

- Rearrange the column independently.
Arrangement from Step 2 Present Arrangement
1 2 3 Row 3 2 1
2 B C A 2 A C B
1 A B C 1 C B A
3 C A B 3 B A C


Example, t=5
Step 1: Obtain the 5x5 Standard Latin Square
A B C D E
B A E C D
C D A E B
D E B A C
E C D B A

Step 2: Rearrange the rows independently
- Obtain 5 random numbers

2
 Order number: 1 2 3 4 5
Random number: 360 764 042 945 195

- Arrange the random number in ascending order, carrying along the original order.
Random number: 042 195 360 764 935
Order number: 3 5 1 2 4

- Rearrange the rows independently.
Rows 1 2 3 4 5
3 C D A E B
5 E C D B A
1 A B C D E
2 B A E C D
4 D E B A C

Step 3: Rearrange the columns independently
- Obtain 5 random numbers
Order number: 1 2 3 4 5
Random number: 816 413 826 773 100

- Rearrange the random number in ascending order.
Random number: 100 413 773 816 826
Order number: 5 2 4 1 3

- Arrange the columns independently.
5 2 4 1 3
B D E C A
A C B E D
E B D A C
D A C B E
C E A D B

Step 4: Rearrange the treatment independently
- The treatment designation A B C D E remain fixed
- Randomised the cells in which the specific treatment occur, given the pattern
obtain in step 3.
- Assume the treatment arrangements as
1 2 3 4 5
A B C D E

- Obtain 5 random numbers
Treatment order A B C D E
Random number 501 154 606 967 133

3
 - Arrange the random number in ascending order carrying along the treatment
order
Random number 133 154 501 606 967
Treatment number E B A C D

- Now ready to arrange treatments
- Present cell designation: A B C D E
- New cell designation: E B A C D
- Hence, final arrangement, refer to arrangement in step 3
Week M T W T F
1 B C D A E
2 E A B D C
3 D B C E A
4 C E A B D
5 A D E C B