[This qmtstion paper contain ~ 6 print ed pages .

Your Roll N o .............. .

Sr. No. of Question Paper : 4 516 E

Unique Paper Code 32371601

Name of the Paper Design of Experiments

Name of the Course B.Sc. (H) Statistics under

CBCS (LOCF)

Semester VI

Duration : 3 Hours
[
Maximum Marks : 75

Instructions for Candidates

1. Write your Roll No. on the top immediately on rec eipt

of this question paper.

2. Attempt five questions in all.

3. Question no . J is compulsory.
•

4. Attempt four questions from the remaining questions .

5. The use of a non -programmable scientific calculator

is allowed.

P.T.O.
 4516 2

1. Attempt any fivt• purls :

(a) Complete the following table for the analysis of

variance of a fixed-effects Latin Square design :
CV\ ~ Y (f) Y, ~-
,...... ~
Sources of d.f. Swnof Mean F
Variation S uares
Colwnns 36 ~l 1~
,1 -b
Rows 3 - 72 - o? 4 2
Treatments
Errors
Total
~ -
6
- I ,,..
180
- ,;. -
12
0 - J-J

- _ ~ I o'J. ~ >' ·
i, -ty (b) Calculate the m-m1mum number of replications
,,., . -~1.µ
e, v ~ _ - required so that an observed difference of 10% of
-, - '6 -bv 1
I,,-~
the mean µ will be regarded as significant at 5%
~~(y
-( level of significance, given that the coefficient of
.
-y::: \V"
variation of plot value is 12%.

(c) Show that for given v, k and A, the parameters r

,) and b given as follows must be integral :
( t-:1/)o
¥(- \'..f".,.
~ / r l's--
\1/(;\ A(v-1) ~ Av(v-1)
~v<;,~ 1-J ,k-i) - .~J k(k-1) .
~ ' ~
(d) Following is the principal block of a 24-fadori.1I
experiment :

( 0000, 0 IO I, 1010, 1111)

Identify the confounded factorial effects.

411,p ~

"I~
 (~ t-1, '-
At~·,~
LO r~
o A.I •
II ,(, , / •
I
...
~\ i~ ) 3
4516 .
l\ f>')(

{C) t\,nl\tru\.'t a ,.i dmdgl\ in llll\:: l;; block s ol' nine run s
ca~h such that information on llu; highest ord er
inh.'raction is compklcly lost.
No ~~ ~, \,,.f1 ~ -i~~'-#< --il~
(t) Define Resolution IV and Resolution V design s
giYing an example of each. (3 :1. 5)

C) ,
(a) Discuss\ th~ factors that affect the shape and size
of plots and blocks in agricultural experiments.

(b) Show that in a Latin Square Design, the mean E{k5.£)

sum of squares due to treatments is a biased @
estimator of the error_ variance: Under what ~
~ co~~tion is it unbiased? Also, derive the expected tY
value of the mean sum of squares due to errors.

(c) Derive the expre@ n to measure the efficiency

0
of RBD o~er CRD, stating clearly the assumptions
used in the derivation. (6,6J) -
3. A randomized block experiment has been carried out
in 4 blocks with 5 treatments A, B, C, D, and E. The
reading for treatment D in block 2 appears to be
missing. Give the complete statistical analysis of this \i.)
design using the missing plot technique. Furtht'r,
calculate the expression for the Htandnrd error of
the difference between two estimated trratmt:nt
means when one of them involves th~ missing
observation.
Ho - GJ
~~, - ®~
~.,,-~
~t @
 ~} 1\)--1. "i I~~\.( , '\A.:= 3 J

4516 4 6% I,_ 1 t1 e J
Lo M '" \-e ~°>j •
-:s 4. (a) Given a BIB design D, with parameters (v, b, r, k,
I -.lJ
~ .. A). Let each block of this design be replaced by
"'I) ). .x: another block containing those treatments which
I I
.. "3 ~ are not included in the original block. What is the

I-' i
N N
0 resulting design called? Derive the parameters~
(v* ~ b*, r*. k*, A*) of this new design D*. Is it
also a BIBO? Justify your answer. Ci) Q)
@) Q--..
(b) Define Resolvable and Affine Resolvable BIBDs.
(y Identify the parameters of the following BIBD
and sho~ hat it is both Resolvable and Affine
- , ...... 'I , .,_. ~1 ~ J 2- , rf"IJ.. -J ~ 1-. .
Resolvable. e,;;, ·

Blocks
I II III IV V VI VII VIII IX X XI XII
Treatments 1 4 7 1 2 3 1 2 3 1.2 3
2 5 8 4 5 6 5 6 4 6 4 5
3 6 9 7 8 9 9 7 8 8 9 7
(7,8)

5. (a) What is meant by factorial ~ eriments? Write

( ~' down the simple effects of a 2 3 factorial
~ experiment. Derive the expressions for the main
I \1 t..,,...o-, - .
(5' effects and interaction e1Tects of 23-factorial
experiment from these simple effects.

(b) An experiment for 5 treatment factors, each with

0 two levels, is qesigneo such that only a single

replicate of the 32 treaiment combinations is run.
Suggest a suitable method for analyzing such a
design. (9,6)
' 4516 s
"
k N.., ( ftt,-.,
}, J., , -Al ,,
)

CI) fvr ~t ~, l JJ, tv ,,

6. (a) nc..•s~ribL' Yat1..'S • algorithm 1'01· eom put i ng tlt c Iot nI

~n\.-cts and th~ S\ltn or sqlWl'l:S due to vnrinus

~ffr~ts for a 3-' foctorial l'Xperiment laid out in r
randomized bloc ks.

(b) In a -3 3 factorial experiment confounded in 3

..,
blocks, three elements of the key block arc

-
.L,c-1iven to be 000, 011, 101. Write other elements
of the key block and identify the confounded
(3)...

effects. -09 A{!, c:,_;

(c) A 2 5 design is to be arranged in 22 blocks of size .

2 3 each. Suggest a suitable set of 3 degrees of (!)
freedom that you would like to confound such that
information on main effects and first-order
pE,-- (0
interactions is not at all lost and write down the
ll.· J 1- (D ~/..
treatment combinations of all the blocks.
A (1 l
) f\ t)'G
~(_ ) Qi !)1.1 (6.3,6)

7. (a) Define fractional factoria~ esigns. Who first (D

proposed the idea of fractional replications in large
<vif'1 f.actorial experime nts? Explain two diffcrenl'l'S
between fraction nl facto ri a I 1..' x peri men ts and
confounded designs. ( \ \,) t,,,,..<.l, .

P.T.O.
4516 6
~
(b) Conslt \ll'l a 2' l f't'ilCI 1011:l I ra~, 01 I :ii dcl.l iv fl wi 111
defining rdat ion~

ACD and BC'h

~
Venfy that each main effect is aliased with at
least one two-factor interaction effect. What i(D
\9 the resolution of this design? Jlow many alternate
fractions can you generate for this design? Also,
give their generating relationships. GJ (6,9)

(1000)