SAMPLING Dr.

Bindu Kumari
SAMPLING

•A research process involving special methods to select a group of subjects of

the research study from its parent population enabling the researcher through

its study to draw inferences about the characteristics of the population.

•It is a process of selecting a few from a bigger group to become the basis for

estimating or predicting the prevalence of an unknown piece of information ,

situation, or outcome regarding the bigger group (Kumar, 2005)

SAMPLING TERMINOLOGY
The population of
study

The inferences
Sampling
drawn from the
process using
study of sample
sampling
are generalized
methods
to population

The
SAMPLE
FUNDAMENTALS OF
SAMPLING
1. statistic (s) and Parameter (s)

2. sampling frame

3. confidence level and significance interval

4. sampling design
PRINCIPLES OF SAMPLING
In most of cases of sampling, there will be difference between sample statistics
and the true population parameter and this attributable to the selection of units
in the sample

The second principle is greater the size of sample, the more accurate will be the
estimate of the population value or parameter.

The third principle is that the greater the difference in the variable under study in
a population for a given sample size, the greater will be the sampling error, that
is, greater will be the difference between sample statistics and true population
parameter.
ADVANTAGES OF SAMPLING
1. Economy in terms of cost
2. Economy in terms of time
3. Economy in terms of labour and efforts
4. Providing last resort or best alternative in many research
situations
5. Accuracy and quality control
FACTORS INFLUENCING
DECISION TO SAMPLE
1. size of the population

2. cost involved in obtaining the elements

3. convenience and accessibility of the elements

HOW LARGE SHOULD A
SAMPLE BE?
It depend on three factors-

A) degree of accuracy required

B) degree of variability in the population

C) number of different variables examined simultaneously in data

analysis
HOW LARGE SHOULD A
SAMPLE BE?
-For example, for small populations (under 1000), a researcher
must have large sampling ratio (30% i.e. 300)
- for moderately large populations (10,000), a sampling ratio of
10% is needed)
- for larger populations, (over 160, 000) , a smaller sampling ration,
1% can yield the results

Till second decade of 20th century, statisticians believed sample

should be large, (central limit theorem)
In 1915, Sealy Gosset (student’s t distribution) led to effective use
of small samples
BASES OF SAMPLING
-Probability is the base for sampling

-the meaning of probability is less than certain and for which there
exists some evidence

-in SAMPLING theory, it is used as equivalent to the relative

frequency

-Blalock (1960), categorized sampling methods into probability and

non probability sampling methods
TYPES OF SAMPLING
STRATEGIES

Types

Non- probability or
Probability or
non-random
random sampling Mixed sampling
sampling
strategies
strategies
PROBABILITY SAMPLING
oThe methods that clearly specifies the probability or likelihood of inclusion of
each element or individual in the sample.

oIt must specify two conditions-

oA) the size of the parent population or universe from which sample is taken ,
must be known to the investigator

oB) each element must have an equal chance of being included in subsequent
sample

oC) the desired sample must be clearly specified

PROBABILITY SAMPLING
Probability
sampling

1. Simple 2. Systematic 3. Stratified 5. Cluster

random random random random
sampling sampling sampling sampling

Proportionate
stratified
random
sampling
Disproportion
ate stratified
random
sampling
SIMPLE RANDOM SAMPLING
1. use of lottery system or a container draw

2. use of a table of random numbers

3. use of computer assisted technology

SIMPLE RANDOM SAMPLING

Sample with
Simple/ replacemen
Unrestricted t
random Sample
sample without
replacemen
t
 STRATIFIED RANDOM
SAMPLING
Population is divided into two or more strata based on a single
criterion such as gender , education level.

These divisions are called subpopulations (non-overlapping)

-stratification tends to increase the precision in estimating the

attributes of the whole population

-it gives some convenience in sampling

PROPORTIONATE STRATIFIED
RANDOM SAMPLING
Class Composition of Proportion of Proportionate
population each class breakdown
12th N1= 3000 0.30 1000*.30= 300
class
BA N2=4000 0.40 1000*.40= 400

MA N3=2000 0.20 1000*.20= 200

M.Phil. N4=1000 0.10 1000*.10= 100

N= 10,000 1.00 Sample = 1000

DISPROPORTIONATE
STRATIFIED
RANDOM SAMPLING
500
6000
10,000 randoml
males
individu y
500
als 4000
randoml
females
y
AREA/CLUSTER SAMPLING
•Origin in field of agriculture, farming experiments

•Geographical divisions of territory, community, neighborhood etc.

are made on map and a certain number of them is drawn at
random and called sample. The investigator proceeds to interview
all elements of the randomly drawn areas or clusters.

•Also called multi stage sampling. E.g. To assess the attitude of

people of Haryana towards family planning
NON- PROBABILITY
SAMPLING

Non- probability

Accidental/ Purposive /
Snowball
Incidental judgemental Quota sampling
sampling
sampling sampling
QUOTA SAMPLING
Sometimes referred as ‘the poor man’s proportionate stratified
sample’
In quota sampling, the investigator recognizes the different strata
of population and from each stratum, he selects the number of
individuals arbitrarily.
SES No. of individuals Selection (according
to convenience, not
randomly)
High 1000 100

Middle 7000 700

Low 2000 200

PURPOSIVE / JUDGEMENTAL
SAMPLING
•Based on typicality of the cases included in the sample
•Based on investigator belief that handpicked sample is very good
representative of the population
•Does not involve any random selection process
•Very convenient method
ACCIDENTAL/INCIDENTAL
SAMPLING
Investigator selects the sample according to his convenience

The sample selected consists of people who are willingly available

and suitable to take part and are the members of target population

Also called opportunity sampling (Pennington, 2002)

Heiman (1995) named it convenience sampling

SNOWBALL SAMPLING
Basically indirectly sociometric
Process of selecting sample using friends and knowns
When the researcher seeks to study a hidden population which is
not easily identifiable
E.g. drug addicts, hardened criminals, prostitutes etc.
It has research applications in relatively small business and
industrial organizations where n is not expected to exceed 100.
OTHER KINDS/
TERMINOLOGY
Saturation or dense sampling
Double sampling
Mixed sampling

Sampling distribution
Sampling error