You are on page 1of 28

SAMPLING DESIGN

(Types of Sampling)
WHAT IS SAMPLING?
• A sample design is a definite plan for obtaining a
sample from a given population.
• If one selects a small subset of the population for the
study and then generalizes the results to the entire
population, then it is referred to as sampling.
• Sampling is an act, process, or technique of selecting a
representative part of a population.
• The main aim is to draw inferences or determining the
characteristics of the whole population.
• Second aim is to test a statistical hypothesis / research
hypothesis relating to population.
SAMPLING TECHNIQUES OR
METHODS
• Sampling techniques are classified into two generic types:
Probability or random sampling and Non-probability or
Non- random sampling.

• Probability sampling is of following types:


• Simple random sampling.
• Stratified random sampling.
• Systematic random sampling.
• Multistage cluster sampling.
• Non -Probability sampling is of following types:
• Convenience sampling (accidental).
• Purposive sampling ( also similar is judgmental).
• Quota sampling.
• Snow-ball sampling.
PROBABILITY SAMPLING
METHODS-
SIMPLE RANDOM SAMPLING:
• This sampling technique gives each element an
equal and independent chance of being selected. A
equal chance means equal probability of selection.

• That is all elements should be included in the


sample frame to draw a random sample.

• If the purpose of research is do arrive at


conclusions/predictions affecting the population of
the whole , then choice of this method is useful.
• Procedure:
• Enumeration of all elements in the
population.
• Preparation of a list of all elements, giving
them numbers in a serial order, 1, 2, 3, ----
so on.
• Drawing sample number by using lottery
method. (computer facility- random
numbers)
• Two alternatives: with replacement /
without replacement.
• SIMPLE RANDOM SAMPLING:

1. Suitable only when population is a


homogeneous group with reference to
characteristics.
2. Possible only when population is
relatively small.
3. Possible only when complete list of all
elements are present.
SIMPLE RANDOM SAMPLING
Disadvantages:
Advantages:
1. Equal chance. 1. Impractical – non availability
of population list –
2. Easiest to apply. migrants/nomadic life, etc.
3. Does not require 2. Wasteful – fail to get full
prior knowledge of information about the
population. population.
4. Sampling error can 3. Techniques does not ensure
be easily be proportionate representation
computed. to various groups.
4. Less precise method –
sampling error is large.
5. Expensive in time and
money.
STRATIFIED RANDOM
SAMPLING
• This is an improved type of random or probability
sampling.
• The problems of simple random sampling gave
rise to this new form.
• In this method, the population is sub-divided into
homogeneous groups or strata, and from each
stratum, random sample is drawn.
• Example: University students based on
discipline OR employees of business
undertaking may be divided into managers and
non-managers based on salary/grade, etc.
• Need for stratification:

• To increase a sample’s statistical efficiency.


• Provides adequate data for analyzing the various
sub-populations and applying different methods to
different strata.
• Ensures representation to all relevant sub-groups of
the population. More efficient than simple random
sampling.
• Essential when researcher wants to study the
characteristics of population sub-groups, e.g., male
and female employees, etc.

• Stratification is useful to apply different methods of


data collection, example interviews for workers and
self-administered questionnaire for executives, etc.
• Suitable for large heterogeneous groups.
• Suitable for large heterogeneous groups.

• Stratification Process: Three major decisions:

• Stratification base to be decided for study – size of


firm, or block, etc.

• Number of strata – larger the representative ness,


the better – need go for sub-population groups –
cost of stratification to be considered.

• Strata sample sizes may be proportionate to


strata’s share of total population – proportionate /
disproportionate.
PROPORTIONATE STRATIFIED
SAMPLE
Employees Number of Elements Proportional Samples
(20%)
Clerical Workers 500 100

Computer 400 80
Administrators
Supervisors 200 40

Low level Managers 150 30

Middle Level Managers 40 8

Top Level Managers 10 2

TOTAL 1300 260


STRATIFIED RANDOM SAMPLING
(Proportionate)
Disadvantages:
Advantages:
1. Prior knowledge and
• Proportionate sampling characteristics is essential
gives proper to adopt this method about
representation to each the population.
stratum and its statistical
efficiency is generally 2. Expensive in time and
higher. money. Greater efficiency
may require additional
• Very popular method. cost.
• Easy to carry the method. 3. Identification of the strata
• Population mean can be might lead to classification
estimated simply by errors. Some may be
calculating the sample included in wrong strata.
mean. 4. Wrong interpretation.
DISPROPORTIONATE STRATIFIED
RANDOM SAMPLE
Employees Number of DIS-Proportional
Elements (samples)
Clerical Workers 500 70
14 %
Computer 400 60
Administrators 15 %
Supervisors 200 50
25 %
Low level 150 50
Managers 33 %
Middle Level 40 24
Managers 60 %
Top Level 10 6
Managers 60 %
TOTAL 1300 260
Disproportionate Stratified Random
sampling:
• This method does not give proportionate representation to strata.
• Involves over representation of some strata and
under representation to others.
• This type of technique is desired when there is more
mixed elements of a stratum – if need for larger sample -
cost per sampling is expected to be more in some strata
selections.

• Usage:
• Appropriate when population contains some small but important
sub-groups.
• When certain sub-groups are heterogeneous, while others are
homogeneous.
• When expected that there is appreciable differences in the response
rates of sub-groups., etc.
• It cannot be used if there is unknown characteristics of the
population or sub-groups.
STRATIFIED RANDOM SAMPLING
(Disproportionate)
Advantages: Disadvantages:
1. Less time consuming 1. Does not give each
compared with stratum the equal
proportionate sampling. importance or
2. Gives importance to representation.
particular groups which 2. Requires prior
needs more important. knowledge of the
composition of the
population.
3. Subject to
classification errors.
SYSTEMATIC SAMPLING
• Meaning and Process:
• Systematic sampling is also called Fixed Interval Method.
• This is an alternative to random sampling.
• It consists of taking every kth item in the population after a
random start with an item from 1 to k.
• Example:
• Suppose to select the sample of 20 students from the list of
300 students – divide the sample number with total - find
quotient (15).
• Select a number at random between 1 and 15 using lottery
method/random number tables – if suppose it is 9, add
9+15, - 24 is the second sample, 24+15 is third sample and
so on. Interval between sample is fixed.
WHERE YOU CAN APPLY THE SYSTEMATIC
SAMPLING ?
• Pseudo-random (randomness cum non-
probability sampling) characteristics.
• Applicable to students in a class,
• Houses in a street,
• Telephone directory,
• Customer of a bank,
• Members of a association, etc.
• Statistically more efficient than simple random
sample. Better representative.
SYSTEMATIC SAMPLING
Disadvantages
Ad van tag es

1. Much simpler than random 1. Ignores all elements


sampling. Easy to use. between two kth elements
2. Easy to instruct the field selected. Cannot be
investigators to use this considered as probability
method. sampling.
3. Requires less time and less 2. Cannot be said as random.
money.. 3. Generalizations will be
4. Limited time schedule will inaccurate.
prefer this method. 4. Biased sample - no
5. Easy to check the kth representative of the group.
sample.
6. Spread evenly over the
population
MULTISTAGE CLUSTER SAMPLING
• When the population elements are scattered over a
wider area and list of population elements is not
available - simple or stratified random sampling method
not possible. Cluster sampling is usually adopted.
• Meaning:
• Cluster sampling means random selection of sampling
units consisting of population elements. (Cluster
happens to be some geographic subdivisions. Clusters
need not be homogeneous in characteristics).
• Each such sampling unit is a cluster of population
elements.
• From each sampling unit, a sample of population
elements is drawn by either simple random selection or
stratified random selection.
• Determining the number of stages in cluster sampling
depends upon the geographical area of study, size of
population and the consideration of costs.
• If number of stages are more than two, it is called as
multi stage sampling.

• Example: To study / Investigating the working


efficiency of nationalized banks in India. Only few
banks to be taken for study.
• Single stage – select a large sampling unit – states in
a country (random selection of few banks).
• Two stage – certain districts (two stage sampling-
census of all banks within the districts);
• Three stage - certain towns and interview all banks in
the chosen town. If we select randomly at all stages, it
is called in multi stage random sampling design.
• Applications of multi – stage cluster sampling:
• The applications of cluster samplings are extensive,
particularly in farm management surveys, socio-economic
surveys, demographic studies, ecological studies, public
opinion polls, attitude surveys, and so on.
• Advantages:
• Easier and more convenient; cost is less; convenience for
field work; units can be substituted for other units within
the same random; flexible.
• Disadvantages:
• Bias in cluster size; sampling error is greater; statistically
less efficient; wrong representativeness will led to wrong
inferences.
NON-PROBABILITY SAMPLING
METHODS
CONVENIENCE or ACCIDENTAL SAMPLING:
• Non-probability sampling – selecting sample units in a just
‘hit and miss’ fashion – interviewing people whom we
happen to meet – conveniently – accidental meetings.
• Usefulness:
• Testing ideas or rough impression on the subject. Use
when data collection is not possible.
• Advantages:
• Cheapest and simplest; Does not require list of population;
Does not require any statistical expertise.
• Disadvantages:
• Highly biased; less reliable; findings cannot be generalized.
PURPOSIVE or JUDGEMENT SAMPLING:
• This method means deliberate selection of sample units that
conform to some pre-determined criteria. This is also known as
judgment sampling.
• This involved selection of cases which we judge as the most
appropriate ones for the given study.
• Does not aim at securing a cross section of a population.
• Subjective knowledge of the researcher helps in judgment.

• Usefulness:
• This method is appropriate when there is a typical and specific
relevance of the sampling units to be studied.

• Advantages:
• Less costly and more convenient; guarantees inclusion of relevant
elements in the sample, where as probability does not.
• Disadvantages:
• Does not ensure representative ness; less efficient in generalization
of statements; requires prior extensive information about the
population to be studied; is not useful for inferential statistics, only
useful for descriptive statistics.
QUOTA SAMPLING
• A convenient sampling technique involving selection of
quota groups (OR) a method of stratified sampling in
which selection within strata is non-random.
• Example:
• The number of respondents (quota) that are to be drawn
from each several categories is specified in advance and
the final selection of respondents is left to the interview
who proceeds until the quota for each category is filled.
• Similar like stratified sampling, but the judgment of
selection is left to the interviewer.
• Applications:
• Used in studies like marketing surveys, opinion polls,
leadership surveys which do not aim at precision, but quick
results.
QUOTA SAMPLING
• Advantages:
• Less costly and less time; no list of population – no
sampling frame; field work can be organized, no
strict supervision required.

• Disadvantages: No precise representative


sample; impossible to estimate sampling error,;
findings cannot be generalized; interviewers may
ignore the core respondents in slums for example;
difficult to look if more variables has to be
considered; high degree of classification error
(heterogeneous samples).
SNOWBALL SAMPLING
• It is a sample collected by building up a list of special
population by using an initial set of its members as
informants (chain system of referrals) .
• Example:
• Problems faced by Indians – Collect some name from Indian
embassy – ask for some more to supply names and the
procedure leds to exhaustive list of samples or to make a
census survey. OR feedback on the quality products, etc.
• Advantages:
• Useful to study social groups, informal groups in an formal
organization; useful for smaller populations where no frame
are determined, etc.
• Disadvantages:
• Does not allow the use of probability statistical methods;
difficult to apply when population is large; does not ensure
the inclusion of all elements in the list.
Probability Vs Non – Probability
Sampling
• Probability Characteristics Non – Probability Characteristics

1. Every population has a 1. Does not ensure a selection


chance of being selected. chance to each population unit.
2. Probability sampling yields a 2. Selection probability is unknown.
representative sample. 3. A non-probability sample may not
3. Findings are generalized. be a representative one.
4. Cost and time required are 4. Does not led to inferences some
large. times.
5. Benefits derived should justify 5. Distort results.
the cost. 6. It is done when data collection is
not feasible. (large and cost). And
not aim at generalizing the
findings.
Part – 2 Continued /-

Thank You

You might also like