Systematic Sampling

1. Definition
Systematic Sampling
Systematic sampling is a sampling method in which:

• The population units are arranged in some order.

• One unit is selected randomly from the first k units.

• Thereafter, every k th unit is selected.

If:

N = nk
then:

• N = population size

• n = sample size

• k = sampling interval

where
N
k=
n

1
2. Population Units
Suppose the population consists of:

U1 , U2 , U3 , . . . , UN
where:

• Ui = ith population unit

Associated study variable values:

Y1 , Y2 , Y3 , . . . , YN
where:

• Yi = value of study variable for unit Ui

3. Variable Descriptions

Symbol Meaning
N Population size
n Sample size
k Sampling interval
Ui ith population unit
Yi Characteristic value for unit Ui
Ȳ Population mean
ȳsys Systematic sample mean
S2 Population variance
V (·) Variance
SE(·) Standard error
T Population total
r Random start

2
4. Sampling Interval
Sampling interval:
N
k=
n
If N is not exactly divisible by n, approximate interval is used.

5. Selection Procedure
Choose a random number:

r ∈ {1, 2, . . . , k}
Then selected units are:

Ur , Ur+k , Ur+2k , . . . , Ur+(n−1)k

6. Possible Systematic Samples

Total possible systematic samples:

k
because only the random start determines the sample.

7. Probability of Selecting a Particular Sample

Each systematic sample has probability:
1
k

8. Population Mean
N
1 X
Ȳ = Yi
N i=1

3
9. Population Total
N
X
T = Yi
i=1

10. Population Variance

N
1 X
S2 = (Yi − Ȳ )2
N − 1 i=1
Alternative form:
N
!
1 X
S2 = Yi2 − N Ȳ 2
N −1 i=1

11. Systematic Sample Mean

Suppose selected sample is:

Yr , Yr+k , Yr+2k , . . . , Yr+(n−1)k

Then systematic sample mean:
n−1
1X
ȳsys = Yr+jk
n j=0

12. Expectation of Systematic Sample Mean

E(ȳsys ) = Ȳ
Thus systematic sample mean is unbiased under random start.

13. Estimator of Population Total

T̂ = N ȳsys

4
Expectation
E(T̂ ) = T
Thus unbiased.

14. Variance of Systematic Sample Mean

Variance under linear systematic sampling:
k
1X
V (ȳsys ) = (ȳr − Ȳ )2
k r=1
where:
n−1
1X
ȳr = Yr+jk
n j=0

Alternative form:
k−1 2
V (ȳsys ) = Sc
k
where:
k
1 X
Sc2 = (ȳr − Ȳ )2
k − 1 r=1

15. Variance of Population Total Estimator

V (T̂ ) = N 2 V (ȳsys )

16. Intra-class Correlation Form

Variance may also be expressed as:

S2
V (ȳsys ) = [1 + (n − 1)ρ]
n
where:

5
 • ρ = intra-class correlation coefficient

17. Important Cases of Variance

Case 1: ρ = 0
S2
V (ȳsys ) =
n
Equivalent to SRS.

Case 2: ρ > 0
Variance increases.

Case 3: ρ < 0
Variance decreases.
Thus systematic sampling becomes highly efficient.

18. Standard Error

q
SE(ȳsys ) = V (ȳsys )

19. Estimation of Population Mean

Estimator:

Ȳˆ = ȳsys

Expectation
E(Ȳˆ ) = Ȳ

6
20. Estimation of Population Proportion
Suppose:
(
1, if attribute present
Yi =
0, otherwise
Population proportion:
A
P =
N
Sample proportion:
a
P̂ =
n
where:

• A = total population units possessing attribute

• a = sampled units possessing attribute

Expectation
E(P̂ ) = P

21. Circular Systematic Sampling

When:

N ̸= nk
circular systematic sampling is used.
Selection proceeds circularly after reaching end of population list.
Selected units:

r, r + k, r + 2k, . . . (mod N )

7
22. Linear Systematic Sampling
If:

N = nk
ordinary systematic sampling is called linear systematic sampling.

23. Properties of Systematic Sampling

1. Simple to Implement
Only one random start required.

2. Uniform Coverage
Sample spreads uniformly over population.

3. Economical
Less time and cost.

4. Easy Fieldwork
Convenient for practical surveys.

5. Unbiasedness
Estimator unbiased under randomization.

6. Variance Depends on Population Arrangement

Efficiency depends strongly on ordering.

7. Can Be More Efficient Than SRS

If nearby units are similar.

8
8. Can Be Less Efficient Than SRS
If periodicity exists.

24. Advantages of Systematic Sampling

1. Very Easy to Use
Simple selection procedure.

2. Saves Time
Fast sampling process.

3. Low Cost
Economical compared to SRS.

4. Good Population Coverage

Entire population gets represented.

5. Useful for Large Populations

Suitable for industrial and field surveys.

6. Requires Only One Random Number

Unlike SRS.

25. Disadvantages of Systematic Sampling

1. Problem of Periodicity
If hidden periodic pattern exists:

Bias may occur

9
2. Variance Estimation Difficult
Exact variance estimation often difficult.

3. Population Ordering Affects Efficiency

Bad ordering may reduce efficiency.

4. Not Completely Random

After first selection, all units predetermined.

26. Uses of Systematic Sampling

• Industrial quality control

• Agricultural surveys

• Household surveys

• Forest surveys

• Production line inspection

• Population surveys

• Medical surveys

• Educational surveys

27. Comparison with SRS

Feature Systematic Sampling SRS

Random numbers needed One Many
Ease of implementation Easy Moderate
Uniform spread Yes Not guaranteed
Effect of ordering High Low
Risk of periodicity Yes No

10
 Field cost Lower Higher

28. Important Results Summary

Sampling Interval
N
k=
n

Selected Units
Ur , Ur+k , Ur+2k , . . . , Ur+(n−1)k

Probability of Sample
1
P (s) =
k

Sample Mean
n−1
1X
ȳsys = Yr+jk
n j=0

Expectation
E(ȳsys ) = Ȳ

Variance
k
1X
V (ȳsys ) = (ȳr − Ȳ )2
k r=1

Intra-class Correlation Form

S2
V (ȳsys ) = [1 + (n − 1)ρ]
n

11
Population Total Estimator
T̂ = N ȳsys

29. Most Important Exam Results

For exams like CSIR NET, GATE, IIT JAM, ISI, etc., the most important
formulas are:
N
k=
n

Ur , Ur+k , Ur+2k , . . . , Ur+(n−1)k

1
P (s) =
k
n−1
1X
ȳsys = Yr+jk
n j=0

E(ȳsys ) = Ȳ

S2
V (ȳsys ) = [1 + (n − 1)ρ]
n

T̂ = N ȳsys

V (T̂ ) = N 2 V (ȳsys )

12