SYSTEMATIC RANDOM SAMPLING Notations

yij is the observation on the j − th unit of the i − th sample, i = 1, 2, L , k ; j = 1, 2, L , n .

A sampling technique in which only the first unit is selected with the help of random numbers
and the rest get selected automatically according to some pre-determined pattern (in regular
1 n
spacing pattern) is known as systematic random sampling. Suppose N units of the yi. = ∑ yij ,
n j =1
mean of the i − th systematic sample.
population are numbered from 1 to N in some order. Let N = nk , where n is the sample
N
size, and k = , being an integer and usually called the sampling interval. Draw a random For notational convenience, we shall write y sy = mean of the systematic sample = yi.
n
number less than or equal to k , say i , and select the unit with the corresponding serial
1 k n 1 k
number and every k − th unit in the population thereafter. The resultant sample will contain Y = ∑ ∑
nk i =1 j =1
y ij = ∑ y i. , population mean.
k i =1
the n units with serial numbers i, i + k , i + 2k , L , i + (n − 1) k is called every k − th systematic
sample and such a procedure termed linear systematic sampling.
k n
1
Example: There are 50 houses on a street. If a sample of size 5 is to be chosen, then S2 = ∑ ∑ ( yij − Y ) 2 , mean square between units in the population.
nk − 1 i =1 j =1
N
k = = 10 , we select randomly one house out of the first ten, suppose we select 3rd and then
n Theorem: In Systematic sampling with interval k , sample mean y sy is an unbiased
take every 10th house after selected one, i.e. as 3rd, 13th, 23rd, 33rd, 43rd. And the possible
sample in (linear) systematic sampling will be: estimate of the population mean Y . Its variance is
N − 1 2 k (n − 1) 2 N −1 2 n −1 2
Sample number Sample units V ( y sy ) = S − S wsy = S − S wsy ,
N N N n
1 1 11 21 31 41
k n
1
2 2 12 22 32 42 2
where, S wsy = ∑ ∑ ( yij − yi. ) 2 is the mean square among the units which lies
k (n − 1) i =1 j =1
3 3 13 23 33 43
within the same systematic sample.
M M M M M M
Proof:
10 10 20 30 40 50
By definition,
Linear systematic sampling suffers from the limitation that it cannot be used when the
1 k 1
sampling interval k =
N
is not an integer. The procedure to be followed is that of circular E ( y i. ) = ∑ yi. = Y , since k is the probability of selecting i − th systematic sample.
k i =1
n
systematic sampling [proposed by Lahiri]. In this method, select first item randomly out of
To obtain the variance, we have,
N units and then take every k − th unit thereafter (where k is the nearest integer to N / n ) in
k n
a cyclical manner until n sampling units are obtained. 1 k 1
V ( y sy ) = ∑ ( yi. − Y ) 2 , and
k i =1
S2 = ∑ ∑ ( yij − Y ) 2 .
N − 1 i =1 j =1
Example: Consider a population of size N = 33 . If a sample of size n = 5 is to be drawn.
N Consider,
Here the sampling interval = 6.6 , is a fractional number, and we have to go in for circular
n
k n k n
systematic sampling. Since the integer k nearest to 6.6 is 7, we select randomly one item out
( N − 1) S 2 = ∑ ∑ ( yij − yi. + yi. − Y )2 = ∑ ∑ [( yij − yi. )2 + ( yi. − Y )2 + 2 ( yij − yi. ) ( yi. − Y )]
of N = 33 , suppose 10th item be selected and then take every 7th item from that one, i.e. as
i =1 j =1 i =1 j =1
10th, 17th, 24th, 31st, 6th.
Systematic random sampling is extensively used on account of its low cost and simplicity in k n k k  n 
the selection of the sample. Apart from its simplicity, this procedure in many situations = ∑ ∑ ( yij − yi. ) 2 + n ∑ ( yi. − Y ) 2 + 2 ∑ ( yi. − Y )  ∑ ( yij − yi. ) .
provides estimates more efficient than simple random sampling. For example, the products i =1 j =1 i =1 i =1  j =1 
coming out continuously from a manufacturing process may be sampled systematically by k n k
selecting products manufacturing at a fixed interval of time. Again, in sampling from plants
growing in rows, one may devide the whole area into a number of equal rectangular blocks,
= ∑ ∑ ( yij − yi. ) 2 + n ∑ ( yi. − Y ) 2 ) , as sum of deviations from their mean is
i =1 j =1 i =1
choose a plant at random from the first block and draw plants exactly from similar spots from zero.
other blocks. (It has been found very useful for sampling material continuously distributed
over time or in space). 2
= k (n − 1) S wsy + nk V ( y sy .
Systematic sampling

Therefore, k n

N − 1 2 k (n − 1) 2 N −1 2 n −1 2
⇒ ∑ ∑ ( yij − Y ) ( yij′ − Y ) = (n − 1) ( N − 1) ρ w S 2 (4.1)
V ( y sy ) = S − S wsy = S − S wsy . i =1 j ≠ j′
N N N n
By definition,
Corollary: E (Yˆ ) = N E ( y sy ) = N Y = Y , means that Yˆ = N y sy is an unbiased estimate of
2 2
the population total Y , and 1 k 1 k 1 n  1 k  n 
V ( y sy ) = ∑ ( yi. − Y ) 2 = ∑  ∑ y ij − Y  = ∑ ∑ y ij − nY 
 N − 1 2 k (n − 1) 2  k i =1 k i =1  n j =1  n 2 k i =1  j =1 
V (Yˆ ) = V ( N y sy ) = N 2 V ( y sy ) = N 2  S − S wsy    
 N N  2
1 k  n  1 k  n n n 
= N ( N − 1) S 2 − Nk (n − 1) S wsy
2
. = ∑ ∑ ( y ij − Y )  = ∑  ∑ ( y ij − Y ) 2 + ∑ ∑ ( y ij − Y ) ( yij′ − Y )

nN i =1 j =1  nN i =1  j =1 
   j =1 j′≠ j =1 
Theorem: The mean of the systematic sample is more precise than the mean of a simple
2
random sample if and only if S wsy > S2. 1 k n k n 
= ∑ ∑ ( y ij − Y ) 2 + ∑ ∑ ( y ij − Y ) ( y ij′ − Y ) (4.2)
nN i =1 j =1 
N −n 2  i =1 j ≠ j′ 
Proof: If y is the mean of a simple random sample of size n , then V ( y ) = S , and
nN Substitute the value of equation (4.1) in equation (4.2), we get
N − 1 2 k (n − 1) 2
y sy is the mean of systematic sample, then V ( y sy ) = S − S wsy , so that 1
N N V ( y sy ) = [( N − 1) S 2 + (n − 1) ( N − 1) ρ w S 2 ]
nN
N − n 2 N − 1 2 k (n − 1) 2
V ( y ) − V ( y sy ) = S − S + S wsy ( N − 1) S 2
nN N N = [1 + (n − 1) ρ w ] .
nN
k (n − 1) 2  N − n N −1 2 n − 1 2 n −1 2
= S wsy +  − S = S wsy − S The relative precision of systematic sample mean with simple random sample mean is given
N  nN N  n n by
n −1 2
= ( S wsy − S 2 ) = + ve quantity, only when S wsy
2
> S2. ( N − 1) S 2
n V ( y sy ) [1 + (n − 1) ρ w ]
nN N −1
This result shows that systematic sampling is more precise than simple random sampling if Relative Precision ( RP) = = = [1 + (n − 1) ρ w ] .
V ( y) N −n 2 N −n
the variance within systematic samples is larger than the population variance as a whole. S
nN
Comparison of systematic with simple random sampling It can be seen that the relative precision depends on the value of ρ w , if
For comparing systematic sampling with simple random sampling without replacement, it is 1
convenient to express V ( y sy ) in terms of intra-class correlation coefficient between the pairs i) ρw = − , the two methods give estimate of equal precision, i.e. as
N −1
of units of the same systematic sample. V ( y sy ) N − 1  n +1 
= 1 −  =1.
Consider the correlation coefficient between pairs of units that are in the same systematic V ( y) N −n N − 1
sample. It is defined as
1
k n ii) ρ w < − , the estimate based on systematic sample is more precise, i.e. as
1
E ( yij − Y ) ( yij′ − Y )
∑ ∑ ( y ij − Y ) ( yij′ − Y )
k n (n − 1) i =1 j ≠ j′=1 V ( y sy )
N −1

ρw = = < 1.
E ( y ij − Y ) 2
1 k n V ( y)
∑ ∑ ( yij − Y ) 2
nk i =1 j =1 1
iii) ρ w > − , Systematic sampling is less precise than simple random.
N −1
k n
∑ ∑ ( yij − Y ) ( yij′ − Y ) Comparison of systematic with stratified random sampling
i =1 j ≠ j′=1 1 k n
= , since S 2 = ∑ ∑ ( yij − Y ) 2 . Let us suppose that population of N = nk units is divided into n strata corresponding to the n
(n − 1) ( N − 1) S 2 N − 1 i =1 j =1
rows of schematic diagram of systematic sampling as follows:
Systematic sampling

k n
Sample numbers 1
1 2 K i K k E ( y ij − y. j ) ( yij′ − y. j′ )
∑ ∑ ( y ij − y. j ) ( yij′ − y. j′ )
k n (n − 1) i =1 j ≠ j′=1
ρ wst = =
1 2 K i K k E ( y ij − y. j ) 2 1 k n
1+ k 2+k K i+k K 2k ∑ ∑ ( yij − y. j ) 2
nk i =1 j =1
1 + 2k 2 + 2k K i + 2k K 3k
k n
M M M M M M
∑ ∑ ( yij − y. j ) ( yij′ − y. j′ ) n k
1 + ( j − 1)k 2 + ( j − 1)k K i + ( j − 1)k K jk i =1 j ≠ j′=1 1
M M K M K M
=
2
2
, as S wst = ∑∑
n (k − 1) j =1 i =1
( yij − y. j ) 2 (4.3)
(n − 1) n (k − 1) S wst
1 + (n − 1)k 2 + (n − 1)k K i + (n − 1)k K nk k n
⇒ ∑ ∑ ( yij − y. j ) ( yij′ − y. j′ ) = n (n − 1) (k − 1) ρ wst S wst
2
(4.4)
and one unit is drawn randomly from each stratum, thus giving a stratified sample of size n . i =1 j ≠ j′
1 k
Then the mean of the j − th stratum is y. j = ∑ yij , j = 1, 2, L , n , and the population
k i =1
By definition,
2 2
1 k n 1 n 1 k 1 k 1 k 1 n 1 n  1 k  n 
∑ ( yi. − Y ) 2 = ∑  ∑ y ij − ∑ y. j  = ∑  ∑ ( yij − y. j )
mean is Y = ∑ ∑ yij = n ∑ y. j = k ∑ yi. , and the variance of the mean of this stratified
nk i =1 j =1
V ( y sy ) =
k i =1 
k i =1 n j =1 n j =1  n k i =1  j =1
2 
j =1 j =1   
n 
1 1  2 2 1 k n 
random sample will be V ( y st ) = ∑ 
k
− W S .
2 ∑ ∑ ij
nj N j  j j =  ( y − y. j ) 2 + ∑ ∑ ( y ij − y. j ) ( y ij′ − y. j′ )
j =1  n k  i =1 j =1 
i =1 j′≠ j 
Here,
1 2 2
= [n (k − 1) S wst + n(n − 1) (k − 1) ρ wst S wst ] , by using eqn. (4.3), and (4.4).
N 2j k2 1 n2k
N j = k, n j = 1, for all j = 1, 2, L , n , and W j2 = = = .
2
N n2k 2 n2 2
(k − 1) S wst
= [1 + (n − 1) ρ wst ] .
Thus, nk
2 The relative precision of systematic sample mean with stratified sample mean is given by
 1 n Sj
V ( y st ) = 1 − ∑ 2 .
 k  j =1 n V ( y sy )
Relative precision ( RP) = = [1 + (n − 1) ρ wst ] .
V ( y st )
where, S 2j is the mean square between units of the j − th stratum and is defined as
It can be see that the relative precision depends on the depends on the values of ρ wst , if
k
1
S 2j = ∑ ( yij − y. j ) 2 .
k − 1 i =1
i) ρ wst = 0 , then V ( y sy ) = V ( y st ) , i.e. the two methods give estimate of equal precision.

Therefore, ii) ρ wst < 0 , then V ( y sy ) < V ( y st ) , i.e. the estimate based on systematic sample is more
precise than stratified sampling.
1  1 1 n k k −1 2
V ( y st ) = 1 −  ∑ ∑ ( yij − y. j ) 2 = nk S wst , iii) ρ wst > 0, then V ( y sy ) > V ( y st ), i.e. systematic sampling is less precise than simple
n2  k  k − 1 j =1 i =1
random sampling.
n k
1 1 n
2
where S wst = ∑ ∑
n(k − 1) j =1 i =1
( y ij − y. j ) 2 = ∑ S 2j is the mean of mean square between
n j =1
Comparison of systematic with simple and stratified random sampling in a population
with linear trend
units within stratum.
Suppose that the values of the successive units in the population follow a linear trend, so that
Further, we shall express the variance of the systematic sample in a suitable form for a
comparative study. Consider the correlation coefficient between the deviations from the yi = µ + iθ , i = 1, 2, L , N , where µ and θ are constants. In this case,
stratum means of the pairs of units that are in the same systematic sample. It is defined as
Systematic sampling

1 N 1 N 1 N  Hence,
Y = ∑ yi = N ∑ ( µ + iθ ) = N  Nµ + θ ∑ i  k  
N i =1 2
i =1  i =1  1 k 2 1
1 k 1 k
V ( y sy ) = ∑ i.
k i =1
( y − Y ) = ∑ i.  k ∑ i.   , since
k  i =1
y 2
− k  y  Y = ∑ yi.
k i =1
1  N ( N + 1)  N +1   i =1  
=  Nµ + θ =µ+ 2 θ.
N  2 
  k   1  2 k (k + 1) (2k + 1) 1  k (k + 1)  2 
2
and 1k 2 1
2 2
= ∑
k i =1
(i θ ) − ∑  = θ
i θ 
k  i =1   k  6
− θ
k 2
 
 
N N
 N +1  N
 N + 1   
( N − 1) S 2 = ∑ ( y i − Y ) 2 = ∑  iθ − θ  = θ 2 ∑ i − 
 2   2 
i =1 i =1 i =1  2k + 1 k + 1   k −1
= θ 2 (k + 1)  − 2
 = θ (k + 1)  
N N
N ( N + 1) 2
 2  6 4   12 
N +1 2 N ( N + 1) (2 N + 1) N ( N + 1)
= θ 2 ∑ i 2 − 2 ∑i +  =θ  − 
i =1 2 i =1 4   6 4  k 2 −1 2
= θ . (4.8)
12
 2 N ( N + 1) (2 N + 1) − 3N ( N + 1) 2  θ 2 N ( N + 1)
=θ 2  = ( N − 1) . Comparing the above three variances obtained in equations (4.6), (4.7) and (4.8), we see that
 12  12
k 2 −1 k 2 −1 (k − 1) (nk + 1)
Therefore, V ( y st ) = ≤ V ( y sy ) = ≤ V ( y sr ) = .
12 n 12 12
N ( N + 1) 2
S2 = θ . (4.5)
12
Suppose the population of size N is divided into n classes of k each, i.e. N = nk , then
nk (nk + 1) 2
S2 = θ , Accordingly V ( y sr ) is reduces to
12
 N − n  2  nk − n  nk (nk + 1) 2 (k − 1) (nk + 1) 2
V ( y sr ) =   S =  2  θ = θ (4.6)
 nN   n k  12 12
For stratified random sampling, we know that in a population, for N =nk ,
n
k −1
V ( y st ) = ∑ S 2j , where S 2j is the mean square between units of the j − th stratum.
n 2 k j =1

Here, S 2j is the mean square of yi = µ + iθ , for i = 1, 2, L , k , and is given by

1 k 1  k 2  1 k
S 2j = ∑ ( yi − Yk ) 2 = ∑ y i − k Yk2  , where Yk = ∑ yi
k − 1 i =1 
k − 1  i =1  k i =1

From equation (4.5), after replacing N by k , we get,
k (k + 1) 2
S 2j = θ , and hence,
12

k −1 k (k + 1) 2 (k − 1) (k + 1) 2 k 2 − 1 2
V ( y st ) = n θ = θ = θ . (4.7)
2 12 12 n 12 n
n k
For systematic sampling, the mean of the second sample exceeds that of the first by θ ; the
mean of the third exceeds that of the second by θ , and so on. Thus the means yi. may replace
by the numbers θ , 2θ , 3θ , L , kθ .