STRATIFIED

RANDOM SAMPLING
DEFINITION
A STRATIFIED RANDOM SAMPLE is one
obtained by separating the population
elements into non overlapping groups,
called strata, and then selecting a
simple random sample from each
stratum
 N1 n2
y2 N2
n1 N3
y1 n3
y3
HOW TO DRAW A STRATIFIED RANDOM
SAMPLE

• Clearly specify the strata

• After the sampling units divided into
strata, select a SRS from each stratum

You plan to stratify the sampling units, say household, into

rural and urban units. What should be done with
households in a town of 1000 inhabitants. Are these
households rural or urban?
Rural if the town is isolated in the country
Urban if the town is adjacent to a large city

Hence to specify what is meant by urban and

rural is essential so that each sampling unit
clearly falls into only one stratum
TAKSIRAN UNTUK :
•MEAN
•TOTAL
•PROPORSI
•ALOKASI SAMPEL
Notation :
L : number of strata
N i : number of sampling units in stratum i
N : number of sampling units in the population
 N1  N 2  ............  N L
yi : sample mean for the SRS selected from stratum i
ni : sample size for stratum i
i : population mean for stratum i
 i : population total for stratum i
 : population total  1   2  .........   L
From SRS 
yi is an unbiased estimator of i
N i yi is an unbiased estimator of  i  N i i

L
1 1
ˆ  yst   N1 y1  ......  N L yL   N i yi
N N i 1
 L
1
yst 
N
N
1
i yi

1 L
 N i  ni   s 
2
SE  2* 2  N i
2

i

N 1  N i   ni 
 L
ˆ  N yst  N
1
i yi

L
 N  n  s  2
V (ˆ)   N 
ˆ
i
2 i i
 
i

1  N i  i 
n

ˆ
Sampling Error  B  2 V (ˆ)
TAKSIRAN UNTUK PROPORSI
L
1
ˆ 
p
N
 N . pˆ
i 1
i i

1 L  pˆ i qˆi  N i  ni  
Vˆ  pˆ   2  N i . 
2
 
N i 1  (ni  1)  N i  
Bˆ  2 Vˆ  pˆ 

1 L  pˆ i qˆi  Ni  ni 
ˆ
B2
N2
 2
Ni  
 (ni  1)  Ni

i 1  
 UKURAN SAMPEL
Misalkan : ni  n wi
L

 i i / wi
N 2
 2

n  1
L
N D   Ni 
2
i
2

B2
D  untuk menaksir 
4

B2
D  untuk menaksir 
4N2
 ALOKASI SAMPEL
Untuk menentukan alokasi sampel harus
dipertimbangkan :
• Ukuran populasi strata; makin besar ukuran populasi
stratum makin besar ukuran sampel yang diambil dari
stratum tersebut
• Variasi observasi dalam tiap stratum  makin
bervariasi, ukuran sampel makin besar
• Biaya pengambilan data di setiap stratum  harga
makin mahal maka ukuran sampel makin kecil.
ALOKASI OPTIMAL
 
 N  / c 
ni  n  L i i i

 
 
 1
N k  k / ck 


 L
 L

  N k  k / ck    N i  i ci 
n   1   1 
L
N D   Ni  i
2 2

1
ALOKASI NEYMAN
 
 Ni  i 
ni  n  
 
L



N
1
i i 


2
 L

  Ni  i 
n   1 
L
N 2
D   Ni  i
2

1
 ALOKASI PROPORSIONAL

 
 Ni   Ni 
ni  n   n 
   N 
L




1
Ni 


2
 L

  Ni  i 
n   1 
L
N 2
D   Ni  i
2

1
Menentukan ukuran sampel untuk menaksir
proporsi dengan sampling error B
ALOKASI OPTIMAL

Ni pi qi / ci
ni  n L

N
1
k pk qk / ck

 i pi qi / wi
N 2
B2
n 1
L
dim ana D 
N 2 D   Ni pi qi
4
1
ALOKASI NEYMAN

 
 Ni pi qi 
ni  n  
 
L



N 1
i pi qi 


2
 L

  Ni pi qi 
n   1 
L
N 2
D   N i pi qi
1
ALOKASI PROPORSIONAL

 Ni 
ni  n  
 N 
Npq
n
ND  pq
2
B
D
4