Sampling theory

Suprakas Hazra
KPC Medical College
Asst . Professor cum Statistician
KPC Medical College, Kolkata
Complete enumeration and sample survey
Two ways to conduct a study
1. Complete enumeration or census method
-Information is collected from each and every member of the
population:
- 2. Sample survey method
– Involves only a small part of the population which represents
the population. The selection is done in such a way that every
individual in the population has the chance to be selected in
the sample. The size of the sample is estimated using suitable
formula, depending upon the objective of the survey.

11/10/2024 S. Hazra 2
 Complete enumeration( contd..)
• Merits
If the population size is small and the proper
care is taken to collect the data , it gives the
best estimate than the sampling as it involves
no error.
• Demerits
- It takes more time, money and manpower

11/10/2024 S. Hazra 3
 Advantages of sampling over complete
enumeration
- Takes less time labour and money
- Has greater scope , since only a small group of skilled
investigators is employed for collection of information
- Yields more precise results. Because of the deployment of
specially trained investigators there is possibility of better
control and supervision
- Magnitude of the error can be estimated
- It sometimes become inevitable when the population size is
too large . If any test is destructive e.g. testing of longevity
of bulbs , testing of crackers etc. then sampling is the only
way
11/10/2024 S. Hazra 4
Mostly used Sampling in test of significance

-Simple random sampling

-Stratified random sampling
-Systematic sampling
-Multistage sampling
-Cluster sampling
- Quota Sampling
- Snowball sampling

11/10/2024 S. Hazra 5
 Parameter and statistic
Parameter Statistic
• characteristic of a population • characteristic of a sample
• characteristics e.g. mean, • characteristics e.g. mean,
standard deviation , standard deviation , proportions
proportions etc based on etc based on sample
population observations observations
• Mean is denoted by µ •Mean is denoted by x̅
• Standard deviation is denoted • Standard deviation is denoted
by σ by s
• Proportion of any attribute is • Proportion of any attribute is
denoted by P denoted by p
The objective of the sampling is to estimate µ by x̅ , σ by s and P by
p respectively

11/10/2024 S.Hazra 6
 Simple random sampling(SRS)

• Sampling frame : List of all population members

• Sampling Unit: Members of the population who are
selected from ‘sampling frame’ by any sampling
technique
• Definition of SRS - Each & every member of the
population has equal probability to occur in the
sample
-If there are ‘N ‘members in the population & the
sample size is ‘n’, the probability that any member of
the population will be included in the sample is 1/N.
11/10/2024 S.Hazra 7
 Simple random sampling
- Let a class consists of 50 (N) students and we
will take a random sample of size 10 (n) to
estimate the average height ( say) then each
student has the chance 1/50 to be included in
the sample.
sampled members are selected by random
number table . Random numbers are numbers
where each number has equal probability to
occur.
11/10/2024 S. Hazra 8
 Random Numbers

11/10/2024 S. Hazra 9
 Drawing sampling units by SRS using random nos.

Example: In a class there are 70 students .Draw a

random sample of size 10 by SRSWOR
Solution: Since 70 is a two digit number , two digit
Identification numbers are given to all students
as 01, 02, 03, ........, 70 to prepare sampling frame.
- 2 digit nos. are drawn from random number
table.
- If the number drawn is 00 or > 70 or repeats,
reject it (since SRSWOR).

11/10/2024 S.Hazra 10
 Drawing sampling units by SRS using random nos.

To draw first 2 digit no. from R.N table, close your eyes
& put a pencil on one of the random no. page. Digit
on which the pencil is pointed , is selected as 1st
member of the sample.
- Next nos. are chosen either from right or from left or
up or down of the no. drawn .
- If the nos. drawn are according to row , next row is
chosen when reached at the end of the row .

11/10/2024 S.Hazra 11
 Drawing sampling units by SRS using random nos.

Random numbers drawn are shown below:

Random 37 76 03 16 22 98 21 06 38 64 91 61 11
Number
Identification 37 03 16 22 21 06 38 64 61 11
No.
Rejected . 76 98 91
Number
Sampled 1 2 3 4 5 6 7 8 9 10
member
Since there were 70 students three numbers
76, 98 and 91 rejected.
11/10/2024 S.Hazra 12
 Merits & Demerits of SRS
Merits
• Every sampling unit has equal chance to be included in the
sample.
• An efficient sampling technique when applied in a more or
less homogenous population.
• Gives better accuracy as it is totally free from any bias and
prejudice.
• Widely used by the researchers if the sampling frame is
available
• Convenient for smaller population and in a small area.
• Very easy to assess the sampling error.
11/10/2024 S.Hazra 13
 Merits & Demerits of SRS contd..
• Demerits:
- It incurs more time, money and manpower since the
sampled units lie scatter .
- For same sample size, it carries larger errors than
stratified random sampling.
- It may not give the true picture if applied on
heterogeneous population.
- SRS cannot be applied if sampling frame is not obtained.
- Since the sampling units are scattered it is very difficult
to contact them.
11/10/2024 S. Hazra 14
 Standard Error
- If same sample size is drawn from the
population, then the statistic has the tendency
to differ from the parameter and not only
that it has the tendency to differ among
themselves also. From these various values of
the statistics, we can get its probability
distribution. The standard deviation obtained
from this probability distribution is standard
error (S.E.).
11/10/2024 S. Hazra 15
 Standard Error contd…
• Standard error of some statistics are given
below:
1. S.E(x̅) = σ/ √n (for SRSWR)
= σ √ (N-n) / √n(N-1) ( for SRSWOR),
σ : population s.d. , N : Population size, n : Sample
size
2. S.E(p) = √PQ/n for SRSWR,
= √ PQ(N-n) / √n(N-1), ( for SRSWOR)
P : Population proportion, Q = 1-P
11/10/2024 S. Hazra 16
Thank You