Chapter 6

SAMPLING FUNDAMENTALS
NTRODUCTION
Ihe primary purpose of
problem with the intention anyof research effort is to look for solutions to
to
entre target
application. However to establish discovering
universality
principles having universal
it is neither
operationally feasible tO study the entire target economical nor
is to develop a population.
sample out of the universe in such a Hence an alternative
represents the characteristics of the universe. manner that it truly
researcher to draw valid inferences or Such a sample allows a
observation of variables within a relativelygeneralization
small
on the basis of careful
It is therefore necessary that
care should be proportion of the population.
various techniques of drawing a sample thattaken while drawing a sample. The
sample is drawn are discussed in the nextwould ensure that a trustworthy
chapter. The current chapter
introduces us to some of the basic concepts pertinent to the sampling method
and aiso explains the underlying
theoretical
of sample survey to the entire population. assumption behind the applicability

0BJECTIVES OF A SAMPLE
The sample survey can be broadly said to involve two
() estimation and (ii) testing of hypothesis. Both these objectives viz,
The asc
coectives of objectives involve
OE Sney are a)a making inferences about a population on the basis of information
gleaned from
2n and b) testing of a sample. According to Frank Yates', "The process of estimation
involves
making inferences about the population from knowledge of the sample, while
in hypothesis testing we take a deductive approach proceeding from a
hypothesis about the population to the behaviour of samples."
I. Estimation. In asample survey, the value of any unknown quantity is
estimated trom the information provided by sampie. It allows the researcher to
move from specificity to generality.
As can be seen in Figure 6.1, the sample provides to us the specitic
Situation which is measured through any statistical value like X or S.D and Is
popuiaion
«Ppied to the general situation i.e. the popuiation to estimate
 6.2
Sampling Fundamentals
parameters like l or o. The reason for using summary descriptors i.e. sample
statistics is that it saves time and money but moreimportantly the researcher is
interested in some characteristic of the population and not the individua! per se.
parameter.
sample statistic is a good estimator of the population
Hence

Sample

(Specific)
Statistic
( Population

(General)
Parameter

POPULATION
ESTIMATION FROM SAMPLE TO
FIG.II.6.1:Testing of Hypothesis. It is based on deductive thinking. Using
statistical tests of significance we can establish the validity of sample based
inferences about the population. It helps us to determine whether certain
proportion of a population differ from one another with respe.
population or hypothesis testing would b
certain characteristics. A detailed discussion on
done later in the book.

SAMPLING DETINITONS
terms relevant to the concent of
In this seetion, a description of various
sampling has been covered.
aggregate of all the units
I. Population/Universe. Population reters to the samnle
All units in any field of
which possess a certain set of characteristics on the basis of which the
nquiry constitute universe be conducted on 'Awareness of
and all units from whom seeks to draw inferences e.g. a study is to
nformation is SOugnt AIDS among rural wómen'. This study is conducted in a village which
to 0-18 a0e
onstitute population.
comprises of 2000 females out of which 35% i.e. 700 belong
group, 45% i.e. 900 belong to 18-50 age group and 20% i.e. 400 defined belong to
has been as
50+ age group. Since our study is on rural women which
'females within 18-50 age then the population is said to be of 900 females.
Universe refers to the total of all the units in the field of inquiry, which in this
case is 2000. Thus all units in any field of inquiry (females) constitute universe
and all units from whom information is sought, on the basis of one or more
characteristics (females within 18-50 age) e.g., age, constitutes population.
A universe or a population can be finite or infinite. If it is possible to
enumerate all the possible units in its totality then it is a finite population or
universe. Our earlier example of 2000 females is an example of finite universe.
The size of a finite population is generally denoted by the symbol 'N'. Ih
certain situation it may not be possible theoretically to enumerate or observeal
the items in the fieldof enquiry. Aclassic example is the number of stars in tie
 Chapter 7
SAMPLING DESIGNS
INTRODUCTION
The researcher collects data as per the guidelines laid
design. An essential component of the down in the research
which is concerned with the research design is the sampling design
selection of asample. We
encounter
a carton of fruits on sampl1ng
Our day to day lives e.g. when we in
purchase
inspectiOn of afew fruit pieces. These few pieces are our the basis of
of which a decision about the entire carton of fruits has sample,
been
on the basis
general statement about the entire population has been taken. Thus a
sample results. In our research efforts, we use the made on the basis of
in a more methodical and same concept of sample but
organised mnanner. The current chapter discusses
various aspects of sampling in detail.

CENSUS SURVEY AND SAMPLE SURVEY

All the individuats, things, events, documents or
observations which
belong to a group and possess specific attributes and from which a researcher
in census Survey all the eXiracts the desired intormation constitute population. For example, in a
units of a population are study on "job satisfaction among college teachers with respect to a
included in the study. A
city X, the population comprises of all the teachers who are workingparticular
Dortion of the population is in the
called a sample. colleges located in city X. A portion of this population is referred to as a
sample.
Goode and Hatt define sample as 'a smaller representation of a whole.
A researcher has the option of carrying out a census
survey or a saiie
survey. Under the census or complete enumeration survey ali tne uuts
population are included and the data is collected from each andevery unit e.g.
11 the response of all the teachers working in the colleges of city 'X IS aken
for analysis purposes then it becomes a
census survey.
A census survey is situations wherethe population size
especially suited in
is very small or the objectives of complete ennumeration e.g.
researchspecity few can be
customers are
demand forecasting of industrial products, whose
done using census survey.
 Sampling Desgns
8urvey
1. Advantages of Census
Acensus survey may be a preferred method of
(a) the subjevt uxter study oxcurs only rarely collecing data it
A certLR SAKvey i Statse
(b) the size of population is very small
(c) for building up records about an area, place,
very s r a whhn r e t s governnent udetakessa census survey to obtain thing
the deographie profile of the country.
Iespite these reasons, it has been seen that
meth of data collecion In a sanmple survey, the
from only some of the elements of population. The
inf
sampling the
is or
information
mation
is pret
a

sampe chosen should be representative of the true

Although there does not exist any fool proof method of
endeavour
character
of
he
is lhal
Sample Is a true refection of the population, unless at
comparative study of the population has been done. In assumi
the saTe thanpopulg ay
ite
practice,
devise sampling plans based on sound principles to ensure hoever
the sample. Asample survey offers many advantages to therepresentati1veness
I. Advantages of Sample Survey researcher
(a) Speed and Timeliness. A major advYntage of
rescarcher spends less time in collecting the data than ifusing
asample is
that the
used. Further, the data gathered using a sample methodthe census method ax
relevant unlike a census survey where a complete count is remains curo
take a long time and because of change in
conditions, required which ma
it would her
redundant and historical.
(6) Economy of Cost and Effort.
covered for getting responses, hence the Sampling require fewer people to
field staff required would also he lexu
than a census study. Although as Frank
Yates commented 'the amount t
effort and expenses required to collect
for a sample than for a complete information is always greater per u
census....if the size of the sample needet n
give the required accuracy represents only a
population, the total effort and expense required small fraction of the whoe
to collect the informain
sampling methods will be very much less than that required for a census o2
whole population."
Thus the criteria for choosing between a
comparison of cost vs. the value. sample and census shoulu t*
(c) Accuracy of Results. In the sampling method, it becomes possible
give more attention to every aspect of our study. It becomes possible
conduct a detailed and intensive study, hence the results are generally accurat
and reliable. The extent of non-sampling errors is less and if the sample
been chosen without a bias then
the results are
() Scope of
Further
where measuring or observingStudies
trustworthy. especially
Retained. Sampling
an element in the populationisdestroys Orinpat
its future usefulness. This is especially true for social research wherehurm
 Sampling Designs 7,3

beings are a part of more than onc study. The sampling technique involves only
asmall proportion of the total ppulation and the chances of one individual
repeatedly becoming a part of the study are relatively less
(e) Feasibility. In some situations the entire population may not be
available for measurement, or may be very large and geographically
widespread In such a situation there is no alternative available except to collect
the data from a sample
The above advantages do not mean that the sanpling technique is withou
limitation Sanpling suffers from the following drawbacks like:
() The sample needs to be drawn through a well planned methodical and
scientitic approach. Sometimes these designs may become more
complicated than conducting acomplete population survey.
(i) Sampling is always prone to errors of bias and prejudices. These
problems come to the forefront more strongly if the sarmple lacks
representative character.
(i) Sampling method presupposes that the researcher possesses specific
and specialised knowledge to develop a sample.
Despite these drawbacks, sampling continues to be the preferred method of
study.

CHARACTERISTICS OF A GOOD SAMPLE

A good sample is that which fulfils the objective of research. It should
possess the following characteristics
Representativeness. A representative sample alone is useful for
research. It means that it should possess the characteristics of the population
with the greatest possible accuracy.
Agood sampie is repre IL Adequacy. Since the results of the entire study are governed by the
sertatve. AOequate, representative units, it is therefore necessary that their number i.e. the size of
roased does not g0 for
SuOSttuton onginal the sample should be
adequate. The sample can be small or large, but it should
population.
seleted unts and has a be adequate enough to properly represent the characteristics of the
hgh degree of precision
ieve IL tnbiasedness. The selection of sample units should be done
objectively. It should be free from all bias and prejudices ; only then
dependable results can be obtained. The higher the degree of absence of bias
the more accurate is the sample.
TV. No Substitution. While conducting a sample survey the there chances of
care that is no
non-response are very high. The researcher should take
substitution of the original unit by some convenient method.
degree of precision.
V. High Precision. A good sample will have a high
estimate. The smaller the
Precision is measured by the standard error of
standard error, the higher is the precision. when described in
which
These features make up a good sample,
measurement terms produces valid results.
 Sampling Designs
7.4 certain points
adherence to) shoWn in
SAMPLING DESIGN PROCESS more of an
process is steps, as
umarised in six
saHpting design I can be
The detinitive process.
them a
Figure 7. 1.
Defining the Population

Sarmple Unit
Defining the

Sampling Frame
Determining the

Technique
Selecting aSampling

Size
Determiningthe Sample

Process
Execution of Sampling
PROCESS
SAMPLING DESIGN thus
FIG. 7.1 THE
flow out of each other and are
the diagram in a
The steps shown in aspects of each of these steps are discussed
design
The sampling
is a six-step interrelated. The important
procesS
little more detail.
the most critical step wherein
process starting with
defining the population, Population. The first step is
defining the sample unit,
Defining the definition for his relevant population.
operational
determining the sampling a researcher has to give an total number of individuals in
frame, selecting a sampling Population in the most common form implies the
technique, determining the
the area of study from whom information is being sought, but depending on the
sample size and executing etc. As stated earlier,
workplaces
the sampling process. nature of study the population can be events, required by the
the members who possess the characteristics or information
researcher constitute target population.
T1. Defining the Sample Unit. A corollary to the above point is defining
the sample unit. The sampling unit is the person, place or object about which
or from which the information is required for research e.g. in case of study
conducted on newspaper readership an entire household could be a sampling
unit, however, in case of study on brand preference of college students in
jeanswear every college going student is a sampling unit.
IlI. Determining the Sampling Frame. Also referred to as 'Source List'
it is a comprehensive listing of all the nembers of a population e.g. a telephone
directory is a good sampling frame, as association directory e.g. NASSCOM
will have a listing of all the firms in the industry. If the sampling
frame S not
available, then the researcher needs to prepare the list. It may be possible tha
the source list does not cover all the elements of population. In that case
 sampling trame Sampling Designs 7.5
error enters our
further analysis. rescarch which has to be
accounted for in
A.Selecting the
the type of sample Sampling Technique. The researcher must
ie. a probability or decide on
tcchniques. The decision as regard non-probability
based
objectives of study, the level of tothe tecthnique to be uscd is affectedsampling
hy thc
accuracy
ditferent sampling techniques available to a desired, time and cost etc. The
the nexI seCtion. rescarcher are discussed in detail in
¥.
Determining the
the researcher is What Sample Size. The million dollar question that
sample size is needed ?" A faces
sample SIZe is a morerepresentative sample. misconception is
that a large
However there is no rigid rule
that is used, although a formula doesexist. A sannple size is
optimunm if it fulfills the considered
requirements
reliability and flexibility. Number of of efficiency,
representativeness,
population, nature of respondents, timesubjective factors like nature of
and funds available, sampl1ng
techniue used etc. influence the decision on sample size.
IV. Execution of Sampling
Process. The last step involves working out
the details of drawing a sample.
Procedures for
worked out. In case of non-response, the procedureselecting cach unit must be
to be adopted for filling the
vacant unit should be stated clearly.

TYPES OF SAMPLING TECHNIQUES

Sampling technique or sampling design can be broadly grouped in two
distinct categories : probability and non-probability. In probability sampling.
allthe elements in the population have a known chance or probability of being
included in the sample. In non-probability sampling the elements do not have a
known or pre-determined chance of being selected as subject. Figure 7.2 makes
a comparison of probability and Non probability sampling.
PROBABILITY SAMPLING NON-PROBABILITY SÁMPLING
I. Every individual has a known and equal chance 1. No probability is associated wih an individual being
of being selected. selected

L. We refer to both sample as well as population.2. Sampling frame is not developed.

Sampling frame is used.
3. Paramteric tests are mostly 3 Non-paramteric tests are preferred.
used. about the
4. A Nothing definite be said
sample is more representative of population 4 Can
representativeness of non probability sample.
characteristics.
FIG. 7.2: PROBABILITY AND NON-PROBABILITY SAMPLING
is looking for
Probability samples are used in studies where the ressearcher
about the sample
generalisations
high degree of representativeness so that cost, time, convenience
results can be made. However when other factors like
 7.6 Sampling Designs

two tvpos of become

important rather than generalizability,then non
probabil,
There are
sampling techniques (a) used.
Probablity sampling where The different sampling designs can be shown using
every otbyect or ndividual classifies
has a known and equal representation basis, which
selection
design as
basis. probability
chance of being selected (b) and(ü) on the element When an element
Non-probability sampling then it is known as a unrestricted is
which has
associated
no probability
with
without any restriction
the element is drawn, subject to some
conditions
then sample
it is
Oraw a
samplingtermed
selection of an object or
7.3 shows the different
sample. Figure types ot
designs ca
OVIdual
the basis of these two elements.

Representation
Bass
PROBABILITY
Element
Selection NON-PROBABILITY
Technique
UNRESTRICTED Simple Random Sampling
Convenience Sampling
RESTRICTED Complex Probability Sampling Purposive Sampling
" Stratified Sampling " Judgement Sampling
" Systematic Sampling " Quota Sampling
"Cluster Sampling " Snow Ball
" Sequential Sampling
Sampling

FIG. 7.3 DIFFERENT TYPES OF SAMPLING DESIGNS

PROBABILITY SAMPLING DESIGNS

1. Simple Random Sampling
A simple random sample is defined as one in which each element uf
population has an equal and independent chance of being selected. In cas
Simple random samping is Population with Nunits, the probability of choosing nsample units, sit
one in which each element
of the popuiation has an possible combinations of Nc. samples is given by 1 -. e.g. If we hae
equal and independent
chance of selection. It can Nc,
be drawn using random Populaion of five elements (A, B,C, D, E) i.e. N= 5. and we want as
number tables or lottery of size n = 3, then there are 5c, =
method. 10 possible samples and the probabiliy
any single unit being a member of the sample is 1
given by 10
Simple random sampling can be done in two different ways l.e.
replacemtent' or without replacement'. When the units are selectedinto
sarmple successively after replacing the selected unit before the next draw,i
asimple random sample with replacement. If the units selected are notrepl3
before the next draw, and drawing of successive units are made onlyfrom
remaining units of the population, then it is termed as simple randomsary
without replacement. Thus in the former method a unit once selected may
repeated, whereas in the latter a unit once selected is not repeated, Dueto
 Sampling Designs 7.7

statistical efficiency associated with a simple random

sample without
replacement it is the preferred method.
A simple random sample can be drawn
through either of the
procedures i.e. through lotery method or through random number tables. two
(a) Lottery Method. Under this method units are selected on
the basis of
randent draws. Firstly each member or element of the population is assigned a
unique number. In the next step these numbers are written on
separate cards
which are physically similar in shape, size, colour etc. Then they are placed in
abasket andthoroughly mixed. In the last step the slips are taken out
randomly
without looking at them. The number of slips drawn is equal to the sample size
required.
Lottery method suffers from few drawbacks. The process of writing N
number of slips is cumberSome and shuffling alarge number of slips, where
population size is very large, is difficult. Also human bias may enter while
choosing the slips. Hence the other alternative i.e. random numbers can be
used.

(b Random Number TabBes. These consist of columns of numbers which

bave been randomly prepared. Number of random tables are available e.g.
Fisher and Yates Tables, Tippets random number etc. Listed below is a
sequence of twodigited random numbers from Fisher & Yates table :
61,44, 65, 22, 01,67, 76,23, 57, 58, 54, 11, 33,86, 07, 26, 75, 76, 64,
22, 19. 35, 74, 49, 86, 58, 69,52, 27, 34, 91, 25, 34, 67, 76,73, 27, 16, 53,
18, 19,69, 32, 52, 38, 72, 38, 64, 81, 79 and 38.
The first step involves assigning a unique number to each member of the
population e.g. if the population comprises of 20 people then all individuals are
numbered from 01 to 20. If we are to collecta sample of 5units then referring
to the random number tables 5 double digit numbers are chosen. e.g.using the
above table the units having the folowing five numbers will form asample:
01, 11,07, 19 and 16. If the sampling is without replacement and a particular
random number repeats itself then it will not be taken again and the next
number that fits our criteria will be chosen.
Thus a simple random sample can be drawn using either of the two
procedures. However in practice, it has been seen that simple random sample
involves lots of time and effort and is impractical. Zina0'Leary comments that
"Simple random sampling is rarely used in practice because the process
Sanpling involves demands identification of all the elements of the population, lists of all those
Ding strata which are
iferent from each elements, and a way of randomly selecting from this list-a tall order" Hence
Therogeneous other methods of probability sampling, also called as complex probabuty
Thereafter units are
Trom each strata Sampling, are used.
ty. it is of two
types II. Stratífied Samnpling
ortionate stratified population can be
and where the
onionate sarnpling This method of sampling is used in situation different from each
distinctly
easily divided into groups or strata which are with respect to sOme
OHher, but the elements within a group are homogeneous
 Sampling Designscharacteristics e. g. students of college can be divided into strata on the basis of
7.8
gender, courses offered, age etc. In this the population is first divided into
strata and then a simple randomsample is taken from each stratum. Stratified
stratified
sampling is
disproportionate
of two ypes
stratified sampling.
proportionate
and sampling
(a) Proportionate Stratified Sampling. In this the number of units
selected from cach stratum is proportionate to the share of stratum in the
population e.g. in a college there are total 2500 students out of which 1500
students are enrolled in graduate courses and 1000 are enrolled in post graduate
courses. If a sample of 100 is to be chosen using proportionate stratified
sampling then the number of undergraduate students in sample would be 60 and
two strata are represented in
40 would be post graduate students. Thus the
representationin the population
same proportion in the sample as is their is to estimat
This method is most suitable when the purpose of sampling
is no difference in with:.
the population value of some characteristic and there
Stratun variances.
(b) Disproportionate Stratified Sampling. When the purpose of study is
to compare the differences among strata then it become necessary to drau
equal units from all strata irrespective of their share in population. Sometimes
some strata are more variable with respect to some characteristic than other
strata, in such a case a larger number of units may be drawn from the more
variable strata. In both the situations the sample drawn is a disproportionate
stratified sample.
The difference in stratum size and stratum variability can be optimally
allocated using the following formula for determining the sample size from
different strata

I1; =
n. N, o; for i= 1,2 ....k
N,o, + N5, t .....+ No
Where n, is the sarmple size of istrata, N; the size of strata and o; is the
standard deviation of i strata. In addition to it, there might be a situation where
coSt of collecting asample might be more in one strata than in other. The
optimal disproportionate sampling should be done in a manner that

N,o, VC N,¡,/C, N,¡kVC

Where CI, C, ....C; refer to the cost of sampling in k strata. The sampt
size fromn different strata can be determined using the
following formula :
1, =
N¡, Ct N,¡,! C, +.... N,oj/EC for i = 1,2 ...k
Illustration 1. An organisation has 5000 employees vho have beu
stratified into three levels
Stratum A: 50 executives with standard
deviation = 9
 Sampling Desiys
Sratum B: 1250 non-numal workers with stuward devition - 4
Stratum C: 3700 anual workers with standard deviation -I
How will a sample of 300 emnployces be drawn on a disproportionate hasis
Iuving optimum allocation.
Solution : Using the formula of disproportionate aipliny fon optimum
allocation.
n. N, O
No, +N,d, + N,o,
For Stratum A :
300 (50)(9)
(50)(9) +(1250)(4) +(37000))
135000
14,75 or say 15
9150
For Stratum B :
300 (1250)(4)
(50) (9) +(1250)(4) + (3700)1)
1500000
163.93 or say 164
9150
For Stratum C
300 (3700) (1)
(50) (9)+ (1250)(4) +(3700)(1)
300 (37O00) (1) 121.3 or say 121
(50) (9) + (1250) (4)+ (3700O)(1)
the sample of 300 cmployccs should have 15 executives, 164 non
Thus
manual workers.
manual workers and 121
H1. 8ystematicSampling
versatile and simple form of probability sampling. In this
It is a very population is selected in the sarmple. It
In systematic sample every every nth item within adefined
nt Mem within a detned method random start of an element in the range of
with a
1 o n.
Opuiaion is selected. It involves beginning entries is worked out and after a
iOves beginning with a The sample raio vis-a-vis the total number of
random start of an element
every nh item is chosen e.g. every 20th person in the list or every
In the range of 1 to n. random start
10th house could be chosen as a part of the sample.
major advatage of systematic sampling is that there is no nced to
The is
unique mumber to each element or use random number tables. It
assign a elements have simiiar
statistically more efficient if the population drawbacks
from tWo major
characteristics. The systematic sampling suffers is 'periodicity' in the
there
viz., periodicity and monotonic trend. In case ratio, then the randomness
is lost
population that coincides with the sampling weekly store
sales from a
e.g. if systematic sampling is used to generate years and the sanple
sampling frame containing weekly sales of past two
7.10 Sampling Designs
interval is 7, then the sample would not reflect the weekly variation il saley
This is bccausethe sanple interval coincides wiith the period of data collection
The second problem arises if there is a'monotonic trend' in population i.c. tle
been arranged in some order like a chronological
order
from smallest
sanpling has
frameto largest etc. e.g. the population of India's software industry
companies which have been arranged from hbe smalles
comprises of 3000
largest order and a sample of 100 is drawn with a sanple interval of 15 Le
the draw start randomly from a company listed on number 12. Then drawng
every 15U company would exclude the top or largest compan1es from the

sampBe
WCluster Sampling
In cluster sampling, groups of elements that ideally speaking,
heterogeneous in nature within group, are chosen randomly. Unlike straiied
groups are homogeneous and few elements are rond
sampling where intra oroe
ciuster sampling groupsshosen from each group, in cluster sampling the group withbecome
tn elements that are group
of
hetrogeneity are developed and allthe elements within the
heterogenous in naure fon
the sample. Whereas stratified sampling has intra group homogeneiy am
a group and then the ups of intra group heterogeneity e.g
are chosen randomly. inter group heterogeniety, cluster sampling hasfrom different departments hasa
committee comprising of number of members
of such conmittees, few are
high degree of hetrogeneity. When from number
samnpling.
chosen randomly, then it is a case of cluster
sampling. If from each cluster
This is an example of one stage cluster
chosen randomly using
which has been randomly chosen, few elements are
method then it is a two stage
simple random sampling or any other probability
stage sampling, when te
cluster sampling. A cluster sample can be a multiplemultiple stages e.g. if in a
choice of element in a sample involves selection at
insurance companies is to be
national survey on insurance products a sample of
stages. In the first stage
drawa, then it requires developing clusters at multiple
private companies. At tie
the clusters are formed on the basis of public and
randomly from each cluster
next stage a group of companies is chosenlocation of each chosen conmpai
developed earlier. In the third stage the office
randomly. Thus in multista_
from where data is to be collected is chosen e
done, then from
Sampling. probability sampling of primary units isdrawn and then the t
primary unit a sample of secondary sampling units is
level, till we reach the final stage of breakdown for the sample units.
M. Area Sampling Broupsa
AvarianN of cluster sampling is area sampling. In area sanpling, Inthisa
etc.
clusters are formed on geographic basis such as sectors, blocks
households
particuBar block is chosen randomly and then all the units or areasanpling
the biuck are inciuded in the sample. This is a single stage Samplingliatk
Cluster sampling or area Saunpling is particularly suitable when
Iists of clusters
does not include every member of the population, but
drawback is
geograpl:ical areas are available. However, a Imajor
 Sampling Designs 7.11
clusters are rarely heterogeneous e.g. houses
than dissimilr. within a block are more similar
Fouenta/
sampling V. Sequential Sampling
is a
ofsampling. ItA complex form of
1oTm
Sotexdrawingsamplesin
and ita but data collection and
sampling, it involves drawing samples in a sequence,
analysis is not fixed in
analysis is done at each stage. The size of the sample is
and
stage. The each
advance but a decision rule is stated before the
acto
at each determined
stage after analysis has been done, the sampling begins. At
o size e, stated further sampling is to be decision rule is checked to see if
continued or
eorefhee
samping
begins.to study their proclivity for domesticnot,or e.g. a sample of consumers is drawn
foreign
sampling is continued till the timne sufficient data isbrands
in AC's. In this the
preference. collected to establish their
The probability sampling tech.niques have been
represented in figure 7.4 diagrammatically

(A) Simple Random (B) Proportionate Statified (C) Disproportionate Stratified

Sampling9 Sampling Sampling

(E) Cluster Sampling (E) Systematic Sampling

FIG. 7.4: EXAMPLES OF PROBABILITY SAMPLING TECHNIQUES

ON PROBABÍLITY SAMPLING
LConvenience Sampling
The most common type of non probability sampling done freedom of without any
convernsence samplingthe restrictions
te rosearcher
is convenience sampling. In this the researcher has the
has
Irecdom of choosing any choOosing any respondent based on his convenience. Respondents become apart
Tespondent based on his
CuVenieHce. of the sample because they happen to be at the right place and at the right time,
people
C.g. in surveys conducted at a retail outlet or shopping mall interviews,
 7.12 Sampling Designs
the timne the response was
who happen to bc at these places at
become a part of the sample. Convenience
saimpling is an betng take
economical
phase of a rescarch project. T
and is generally used in exploratory fhe
precisiOn andsampling
and reachable.
units are cooperative, casily accessible However
sampling suffers from many drawbacks like bias, less
proper representation of the population.
conveeT
.Judgement Sampling
I is a kind of purposive sampling where those respondents are
Judgernent Sampling
purposivo sarnpling where
those respondents are
delibaretely made a part ot
a
of their position, knowledge or
made a part of a sample, by virtue
criteria, which meet research purpose. In this case the
researcher uses his
deliberately
any olher
meet judgement or expertise to decide who would be a part of the
sanple, which
research purpose based orn survey is being conducted on finding out what it takes to be a saimple
e.g. if a
researcher's
judgement.
OWn
leader, then it is the people who are in this position i.e. student union
student union
or those who frequently interact with such people, who can give firstleader,
information. .This method is particularly suitable when only a hand
of people have access to the knowledge or information requiredlimited
for number
rescarch.
This sampling does not allow for generalizability to aspecific population
it does not represent a population explicitly.
II, Quota Sampling
This method is used when a researcher needs certain group to be
In quota sampling. group adequately represented. n this groups are created which are homogeneous i
created which
are
are respect to certain characteristics within group. A quota is fixed for each proun
homogenous within and a
quota is fixed tor cach and sample units are drawn from the group. The quota sampling is simibr
group. stratified sampling except for the fact that no sampling frame is used and
san1ple units are drawn on convenience basis instead of random basis. The
researcher while drawing the sample, using this method, tries to ensure that the
composition of sample is the same as the conmposition of population with
respect to the characteristics of interest. It is also possible to assign quotas to a
group which is greater than the proportion of that group in the population e.
while studying consumer behaviour it may be desirable to oversample heay)
users of a product so that their behaviour can be studied in detail. Quoa
samplingies to make the sample representative at a low cost.
IV`now Ball Sampling
In this a set of respondents are selected initially and interviewed. Afie
In snow ball sampling a set us, hese Tespondents arc asked to list the names of other people who
of respondents are selected opiiion are a part of the target population. Thus it is like setting the ballit
initially and interviewed and motion whereby referrals are obtained from referrals, thus creating asnow al
they turther recominend the
names of otlier people. effect wlhicli keeps on growing in size as it Seentha
rolls down. It has been
people relerred by the respondents have
greater demog«bychance
psvchographic characteristics similar to them thanright
This tecnique has the advantage of locating
would occur
they people with the desi
ciharacteristics ai a low cost.