Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
8 views17 pages

Lecture 13

Uploaded by

Lakshaya Murugesan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
8 views17 pages

Lecture 13

Uploaded by

Lakshaya Murugesan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 17

Balanced Incomplete Block

Design (BIBD)
 Notation
– a-number of treatments
– m-number of treatments/block
– r-number of treatment reps
– b-number of blocks
 N=ar=bm
 r(m-1)=l(a-1)
BIBD
 Example

Block
1 2 3 4
A A A B
B B C C
C D D D
 Verify that l=2
Generating BIBD’s
 Oneway to generate BIBD’s
(assuming a and m have been
selected) is to
a  choose
b  
 m

 This worked well in the


previous example
 It can be inefficient for large
a or m≈a/2
Least Squares critetion
for a
 These are the adjusted means
that one obtains from LSMEANS
in SAS

Yij    i   j   ij
Q   nij Yij     i   j 
2

i j
Least Squares estimator of
a
 The solution for ai:

mYi.  n Y
j
ij . j

ˆ i 
a
Intrablock estimator for a
 Theseestimators are called
intrablock estimators because
they can be expressed as
contrasts of within block
observations.  
mYi.  n Y ij . j
m  nijYij   nijY. j 
ˆ i  j
  j j 
a a
 
m  nij Yij  Y. j 
 j 
a
Properties of LS estimators

ma  1 2
V ˆ i   
a 2
Variance of LS estimators
 Thevariance estimate is not
the square of the standard
error for the LSMEANS that one
obtains in SAS (it’s not
corrected for the overall
mean).
Vˆ ˆ i  
m a  1 MSE
a 2
Example
 Region II Science Fair Winner
 Treatment is organic chemical
repellent
 Block is Day
 Response is average number of flies
on underside of lid (smaller is better)
 Originally presented as a Latin
Square (additional block was Fly
Group)
Data Table

Day
1 2 3 4 5 6 7
C=19.8 D=16.0 A=11.7 B=11.2 A=13.2 A=16.0 B=15.7
E=7.8 F=11.0 E=5.3 D=10.0 C=17.3 B=17.2 C=18.0
G=13.0 G=5.3 F=12.3 E=6.0 D=16.2 G=10.8 F=12.7

A=Ascorbic Acid B=Citric Acid C=Control D=Hesperidin


E=Lemon Juice F=Black Pepper G=Piperine
Interblock Model
 WhenBlock is a random effect,
alternative estimators of
treatment effects can be found

  j iid N (0,  2 )
Yij    i   j   ij 

 ij iid N  0,  2

a
 a

Y. j m   nij i   m j   nij ij 
i 1  i 1 
Interblock Estimators

ˆ Y..

n Y ij . j  mrY..
~
i  j

r 
Interblock Estimator
Properties
 Some properties:

~
Cov( i ,  i ) 0
ˆ

m(a  1) 2
~
V  i  
a(r   )
  m 
2

Interblock EMS
 Forthe random block model,
Source df SS EMS
Trt a-1 SSTrt
Blockadj b-1 SSBlock(adj)
 2
 a  2
 ( r  1) (b  1)

Error N-a-b+1 2
Total N-1
Variance Compoents

ˆ 2 MSE
 b 1

ˆ   a (r  1)
2 MS Block ( adj)  MSE  if MS Block ( adj)  MSE  0
 0 otherwise

m ( a  1)
ˆ ~
V  i  
a(r   )

ˆ  mˆ 
2 2

BIBD Estimators
 Supposewe want to minimize the
variance of estimators of the
form:
~
 i  (1   ) i , 0  1
ˆ

 The best weighted estimator is:


 1 1 ~  1 1 
ˆ i *  ˆ i  ~  i    ~ 
 V ˆ i  V  i    V ˆ i  V  i 
Power Analysis
 Thenoncentrality parameters look
similar to all other one factor designs
we have so far seen

You might also like