1

Technology Oriented Business Driven Sustainable Development Environmental Friendly

Definition:
A sample obtained by randomly selecting one element from the first k elements in the
frame and every k element thereafter is called a one-in-k systematic sample.

Advantages

1. Systematic sampling is easier to perform and hence is less subject to interviewer errors
than SRS.

2. Systematic sampling often provides greater information per unit cost than does SRS. A
systematic sample is frequently spread more uniformly over the entire POP and thus can
provide more info. about the POP than an equivalent amount of data contained in SRS.


Illustration
We wish to select a 1 in 5 systematic sample of travel vouchers from a stack of N = 1,000
to determine the proportion of vouchers filed incorrectly.




N = 1,000 k = 5
1, 000
200
5
N
n
k
= = =
2
Technology Oriented Business Driven Sustainable Development Environmental Friendly
A voucher is drawn from the first five vouchers (e.g. #3) and every fifth voucher thereafter is
included in the sample.

Voucher Voucher sampled


1
st
zone
1
2
3
4
5



3




2
nd
zone
6
7
8
9
10
.
.
.



8


n
th
zone
996
997
998
999
1000


998
3
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Divide sampling frame into zones of equal size of k units each



Note
Systematic sampling can also be employed to select sample from each of the
stratum in stratified sampling.

How to draw a systematic sampling

Let say we have a hypothetical POP with N=12:



A sample is to be drawn using 1 in 3 systematic sample.


( )
1 1... ... 1... 1...2
... 1 2
n k nk k k k
n
+ +
1 5 7 2 3 4 6 8 9 10 11 12
x x x x x x x x x x x x
12, 3
12
4
3
N k
n
= =
= =
4
Technology Oriented Business Driven Sustainable Development Environmental Friendly
5 7 1 2 3 4 6 8 9 10 11 12


1 2 3 4
2 3
1
x x x x x x x x x x x x
zon zon zon zon
j j k j k j k
j k
+ + + +
+ + +
s s
In general, j, j+k, j+2k,,j + (n-1)k

The no. of possible systematic samples = k

Example: N = 12, K= 3, n=4

Samples

#1 #2 #3

1
4
7
10
x
x
x
x
2
5
8
11
x
x
x
x
3
6
9
12
x
x
x
x
5
Technology Oriented Business Driven Sustainable Development Environmental Friendly
The probability of getting any systematic sample =

Case when Nnk

Systematic sample 1 in 5



The sample size can be 2 or 3

# of possible samples = 5
1
k
12, 5
12 2
2
5 5
N k
n
= =
= =
12 11 10 9 8 7 6 5 4 3 2 1
x x x x x x x x x x x x
12 11
10 9 8 7 6
5 4 3 2 1
5 # 4 # 3 # 2 # 1 #
x x
x x x x x
x x x x x
Samples
6
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Example

1. N=37
Draw a sample using systematic sample 1 in 8



2. N = 666, Systematic sample 1 in 20

37
8, 4.6
8
4 or 5
k n
n
= = =
=
666
20, 33.3
20
33 34
k n
n or
= = =
=
Statistical inference

1. Mean : or

Estimate with
Y
1
n
y
i
i
sys
y
n
=

=
( )
2
1
s
n
f
y e s
sys

=
7
Technology Oriented Business Driven Sustainable Development Environmental Friendly
2. Total Y or T
Estimate with





3. Proportion, P

Estimate with




( ) ( )

sys
sys sys
Y Ny
se Ny Nse y
=
=
1

n
i
i
sys
y
P
n
=
=

( ) ( )
1
1
1
sys sys
f
se p p p
n

8
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Example

An investigator wishes to determine the quality of maple syrup contained in the
sap of trees on a Vermont farm. The total number of trees, N, is unknown; hence
it is impossible to conduct an SRS of trees. As an alternative procedure, the
investigator decides to use 1 in 7 systematic sample. The data from this survey
are listed below. Entries are the percentage of sugar content (in the sap) for the
trees sampled. Use these data to estimate , the average content of maple
trees on the farm.

Calculate the standard error of estimator.

9
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Tree sampled Sugar content of the sap,y
1
2
3
.
.
.
210
211
212
82
76
83
.
.
.
84
80
79
6724
5776
6889
.
.
.
7056
6400
6241

1,486,800
2
y
212
1
17, 066
i
i
y
=
=

10
Technology Oriented Business Driven Sustainable Development Environmental Friendly
An estimate of is given by















95% CI for : 80.5 1.96 (1.47) 80.5 2.88

( )
( )
( )
1
2
2
2
2
17, 066
80.5
212
1
17, 066
1, 486,800
212
212
535.483
212
1 535.483
1484

212
2.16 1.47
n
i
i
sys
i
i
sys
y
y
n
y
y
n
s
n
se y
=
= = =

=
= =

11
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Example
Let us take a systematic sample of one in six physicians. We first take a random number
between 1 and 6; let say number 5. The physicians selected in the sample, then, are 5,11,17
and 23







i. The estimated mean number of household visits per physician is


ii. The estimated total number of visits made by all physicians in the POP, is




iii. The estimated proportion of physicians making one or more household visits is
Physician Number Number of visitors, y
5
11
17
23
7
0
7
0
5 . 3
4
14
= =
sys
y
( )

25(3.5) 87.5
Y or T Ny =
= =
50 . 0
4
2
= = p
12
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Sampling Distribution of Estimates

Suppose that a sample of one in five physicians is desired.

N = 25, k = 5
n = 5

No. of possible samples = 5
Random no.
chosen
Physicians in
sample


(one or more
visits)

1 1,6,11,16,21 2.6 64 0.4
2 2,7,12,17,22 4.8 120 0.6
3 3,8,13,18,23 1.4 35 0.4
4 4,9,14,19,24 9.2 230 0.8
5 5,10,15,20,25 7.4 185 0.6
y

Y
13
Technology Oriented Business Driven Sustainable Development Environmental Friendly
( ) ( )
( )
1
2.6 4.8 1.4 9.2 7.4
5
5.08
E y
or Y
= + + + +
= =
y is an unbiased estimate of
Similarly,
( )
( )
1

65 120 35 230 185

5
127
1
.4 .6 .4 .8 .6
5
0.56
E Y
Y
E p
p
(

(

= + + + +
= =
= + + + +
= =
14
Technology Oriented Business Driven Sustainable Development Environmental Friendly
The estimated mean, total, and proportion are unbiased estimates of the corresponding
POP. parameters.
( ) ( )
1
2
2
2 1
2.6 5.08 ...... (7.4 5.08
5
2.90
se y

(
= + +
`

)
=
( )
( ) ( )
1
2
1
2
2 2
2
2
1

(65 127) ...... (185 127)

5
72.57
1
0.4 0.56 ...... (0.6 0.56)
5
0.15
se Y
se p

( = + +
`

)
=

(
= + +
`

)
=
15
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Mean Standard Error
Estimate SRS Systematic Sampling SRS Systematic
Sampling
5.08 5.08 3.36 2.90
127 127 84.11 72.57
0.56 0.56 0.20 0.15
For both sampling schemes the estimates are unbiased. However, the standard errors are not
the same. In this example, systematic standard errors are smaller than those of SRS.
16
y
Y
P
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Example

Suppose we specify a 1 in 6 systematic sample of physicians in the list.

# of possible samples = 6
Random no.
chosen
Physicians in
sample
1 1,7,13,19,25 12 300 .80
2 2,8,14,20 1 25 .25
3 3,9,15,21 4.25 106.25 .75
4 4,10,16,22 6.5 162.5 .50
5 5,11,17,23 3.5 87.5 .50
6 6,12,18,24 1.5 37.5 .50
y
Y

17
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Probability of each sample being chosen = 1/6







Thus we see that in this instance systematic sampling does not lead to unbiased
estimates the impact made on the estimates is not the same for each element.

Illustrative example

Suppose that a list of appointments for a nurse practitioner is available to us and
that we will take a sample of 1 in 4 of the patients seen by this nurse on a given day
for purposes of estimating the average time spent per patient. Suppose that on the
day in which the sample was to be taken, the nurse saw a total of 12 patients in the
order shown in the table.
( ) ( )
( )
( )
( ) ( )
1
12 ...... 1.5 4.79
6
1

300 ..... 37.5 119.79

6
1
.80 ..... .50 0.55
6
E y
E Y Y
E p P
= + + = =
= + + = =
= + + = =
18
Technology Oriented Business Driven Sustainable Development Environmental Friendly
Data for Nurse Practitioners Visits (Unordered List)
Order of
visit
Time spent with
patient (min)
Order of
visit
Time spent with
patient (min)
1 15 7 49
2 34 8 40
3 35 9 25
4 36 10 46
5 11 11 33
6 17 12 14
The four possible samples are shown as follows
S1 S2 S3 S4
P T P T P T P T
1 15 2 34 3 35 4 36
5 11 6 17 7 49 8 40
9 25 10 46 11 33 12 14
Total 51 97 117 90
Mean 17 32.33 39 30
19
Technology Oriented Business Driven Sustainable Development Environmental Friendly
( )
( )
29.583 minutes =
63.6
E y
V y
=
=
Suppose now that the nurse scheduled appointments in such a way that the difficult patients
requiring the most time would be seen first and the easier ones requiring less time would be seen
later in the day. The lists of appointments for the same patients discussed above is as follows:-
Data for Nurse Practitioners Visits (Monotonically Ordered List)
Order of
Visit
Time spent
(min)
Order of
Visit
Time spent
(min)
1 49 7 33
2 46 8 25
3 40 9 17
4 36 10 15
5 35 11 14
6 34 12 11
20
Technology Oriented Business Driven Sustainable Development Environmental Friendly
( )
( )
29.583
13.13
E y
V y
= =
=
21
The four possible samples are shown as follows
S1 S2 S3 S4
P T P T P T P T
1 49 2 46 3 40 4 36
5 35 6 34 7 33 8 25
9 17 10 15 11 14 12 11
Total 101 95 87 72
Mean 33.67 31.67 29 24
Environmental Friendly Sustainable Development Business Driven Technology Oriented
Let us now suppose that the nurse scheduled an easy patient followed by two moderately
difficult patients followed by a very difficult (in terms of time required). The appointment list
containing the same 12 patients might then appear as shown below.
Data for Nurse Practitioners Visit
(Periodicity in List)
Order of
visit
Time spent with
patient (min)
Order of
visit
Time spent with
patient (min)
1 11 7 35
2 17 8 46
3 36 9 15
4 49 10 25
5 14 11 33
6 34 12 40
22
Environmental Friendly Sustainable Development Business Driven Technology Oriented
The four possible samples are as follows:
S1 S2 S3 S4
P T P T P T P T
1 11 2 17 3 36 4 49
5 14 6 34 7 35 8 46
9 15 10 25 11 33 12 40
Total 40 76 135
Mean 13.33 25.33 45
( )
( )
29.583
136.41
E y
V y
= =
=
Summary of Results for Four Types of Sampling
Sample Design
Systematic sampling (unordered list) 63.6 .124
Systematic sampling
(Monotonically order list)
13.1 -.371
Systematic sampling (periodically In list) 136.4 0.837
Simple random sampling 41.7 -
( ) y V
23
Environmental Friendly Sustainable Development Business Driven Technology Oriented
( )( )
( )( ) '
2
1 '
2
1 1
k
ir
ir
i r r
y Y y Y
n N

o
= =
=


= average correlation of all pairwise units in the same systematic sample


Random POP
Sampling units (su) are heterogeneous, ~ 0

As N

Periodic POP
su are homogenous,

As N

Ordered POP
su are heterogeneous, As N

2
n | |
|
\ .
( ) ( )
,
srs
sys
V y V y
o s
( ) ( )
,
sys srs
V y V y s
24
o >
( )
,
sys
srs
V y V y
| |
|
\ .
>
Environmental Friendly Sustainable Development Business Driven Technology Oriented
Repeated Systematic Sampling

Systematic sampling (sys):
i. Frame is random and either
i) (+ve integer)

ii) N is large
(sys) method (SRS) method

ii. Frame is not always random
i) monotonic relationship
ii) periodic relationship

Remedies

1. Rearrange randomly the units in the frame
2. Paired selections
3. Choose several random starts in the course of collecting the 1 in k sample.

N
k
n
=
...
i
j k j
i
+ k j
i
2 +
i i
j n k +
( )
2 1
1 j n k + +
25
Environmental Friendly Sustainable Development Business Driven Technology Oriented
4. m repeated systematic samples

a) original zone size

b) mk = new zone size

c) sample size


sample size for each 1 in mk systematic sample

d) obtain m independent systematic 1 in mk systematic samples as follows:

i) obtain m random starts

ii) obtain

N
k
n
=
N
N
m n
mk
n
| |
= =
|
\ .
1
,...
1
m
i
j j where
j mk i s s
1
... y y
m
26
Environmental Friendly Sustainable Development Business Driven Technology Oriented
e) Statistical inference

i) estimate



ii)



iii) 100 (1- )% CI for is:-



f) Statistical Inference (Y)

i) Estimate Y with

ii)

( )
Y
1
m
i
i
y
Y with y
m
=
=

( )
( )
1
2
1
;
1
m
i
i
y y
f n
se y f
m m N
=

= =

Y
( )
2
y Z se y
o

y N
( ) ( )
se Ny Nse y =
27
Environmental Friendly Sustainable Development Business Driven Technology Oriented
Example

Let us suppose that we wish to take a systematic sample of 18 workers from the list of 162
workers for purposes of estimating the mean number of work days lost per worker from acute
illnesses

N = 162 n = 18
k = 9

Notice that 18 = 6 x 3, therefore we can take a sample of 18 workers by taking six systematic
samples each containing 3 workers.

m = 6, k = 9


Repeated systematic samples 1in 54


Choose 6 random nos:-
Say 2, 31, 46, 13, 34, 53

54 mk =
1 54; 1,...,
i
j i m s s =
28
Environmental Friendly Sustainable Development Business Driven Technology Oriented
The six samples are :-
Random
number
Elements of
sample
Days Lost mean
2 2,56,110 6,2,7 5.00
13 13,67,121 4,4,5 4.33
31 31,85,139 6,4,2 4.00
34 34,88,142 5,3,2 3.33
46 46,100,154 6,12,3 7.00
53 53,107,161 7,3,0 3.33
29
Environmental Friendly Sustainable Development Business Driven Technology Oriented
( )
( )
( )
1
6
1
2
5.00 ... 3.33
6 6
4.5
1

1
18
1
162 1.9033
6
0.28197
0.53
m
i
i
i
i
y
y
y y
f
se y
m m
=
=
+ +
= =
=

=
=
=

30
Technology Oriented Business Driven Sustainable Development Environmental Friendly
31