Sampling

Sampling methods and Central Limit theorem: Sampling methods, Sampling error,
The Central Limit Theorem, Using the Sampling distribution of the Sample mean
Contents
 Definition
 Sampling Process
 Types of Sampling
 Sampling Methods
 Merits of Sampling Methods
 Limitations
 Sampling Error
 Central Limit Theorem

Ms. Shaheen, Assistant Professor, IP 2

E, OU Campus, Hyderabad
 DEFINITION

 Sampling - statistical practice concerned with

selection of individual observations intended to
yield some knowledge about a population of
concern

 Population refers not only to people but to all the

items that have been chosen for study

 Sample describes a portion chosen from the

population
Ms. Shaheen, Assistant Professor, IP 3
E, OU Campus, Hyderabad
 SAMPLING
 Statistic is a characteristic of a sample
 Parameter is a characteristic of a population
 Example
– The mean height in inches of all PGDM students in India is
60 inches. 60 inches is the characteristic of the population
“PGDBM students” & can be called population parameter.
– If we say that the mean height of PGDBIF students in IPE is
60 inches, then 60inches describes the characteristic of the
sample. In this case, 60 inches would become a sample
statistic.
 Sample statistics allow to make inference about the
population parameters, through estimation and hypothesis
testing.

POPULATION SAMPLE
Population Size = N Sample Size = n
Population Mean = µ Sample Mean = x
Population Standard Deviation= σ Sample SD = s

Ms. Shaheen, Assistant Professor, IP 4

E, OU Campus, Hyderabad
SAMPLING PROCESS
1. Defining the population of concern
2. Determine a sampling frame
3. Specifying a sampling method for
selecting items or events
4. Determining the sample size
5. Implementing the sampling plan
6. Sampling and data collecting
7. Reviewing the sampling process

Ms. Shaheen, Assistant Professor, IP 5

E, OU Campus, Hyderabad
TYPES OF SAMPLING
 There are 2 methods of selecting samples from
populations:
• Non-random/ Judgment Sampling
• Random/ Probability sampling

 Random/ Probability sampling: - All the items in the

population have a chance of being chosen in the
sample.

 Non-random/ Judgment Sampling: - Personal

knowledge & opinion are used to identify the items
from the population that are to be included in the
sample.
Ms. Shaheen, Assistant Professor, IP 6
E, OU Campus, Hyderabad
SAMPLING METHODS
1. Simple random sampling
Probabilistic/
2. Systematic sampling
Random
3. Stratified sampling
Sampling
4. Cluster sampling
5. Other sampling techniques

Non-probabilistic/ 1. Convenience sampling

Non-Random 2. Judgmental sampling
Sampling 3. Quota sampling
4. Snowball sampling

Ms. Shaheen, Assistant Professor, IP 7

E, OU Campus, Hyderabad
1. SIMPLE RANDOM SAMPLING

 Each sample has equal

chance of being
selected

 Every element is
selected independently
from other elements

Ms. Shaheen, Assistant Professor, IP 8

E, OU Campus, Hyderabad
 Methods followed when drawing a random
sample are:
– Lottery Method
– Use of Random Numbers

 Lottery Method: - Each unit of Population is

numbered & shown on a chit of paper or disc.
– Each chit is folded and put in a box from
which a sample of the requisite size is
drawn.
 Use of Random Numbers: - Tables of random
numbers are used.
– Units of population are numbered from 1 to
N from which n units are selected.
Ms. Shaheen, Assistant Professor, IP 9
E, OU Campus, Hyderabad
 Example
 Suppose a population consists of 845 employees of Nitra
Industries. A sample of 52 employees is to be selected
from that population.

 Write the name of employee on a small slip of paper and

deposit all of the slips in a box. After thoroughly mixed, first
selection is made by drawing a slip out of the box and is
repeated until 52 employees is chosen.

 A more convenient method is to use identification number of

each employee and a table of random numbers.
– First choose a starting point in the table. Any starting point will do.
– Suppose time is 3:04, you might look at third column and then
move down to fourth set of numbers.
– Number is 03759. since there are 845 employees, we will use first 3
digits of 5 digit random number.
– To continue selection, move in any direction. Suppose you move
right. First 3 digits of the number to the right of 03759 are 447
selected to be second member of the sample.
– You continue to the right and select employee 784, then 189 and so
on.

Ms. Shaheen, Assistant Professor, IP 10

E, OU Campus, Hyderabad
Portion of random numbers

50525 57454 28455 68226 34656 38884 39018

72507 53380 53827 42486 54465 71819 91199
34986 74297 00144 38676 89967 98869 39744
68851 27305 03759 44723 96108 78489 18910
06738 62879 03910 17350 49169 03850 18910
11448 10734 05837 24397 10420 16712 94496

Ms. Shaheen, Assistant Professor, IP 11

E, OU Campus, Hyderabad
 Advantages Disadvantages

oNo possibility of oLottery method is time-

personal bias, since consuming & expensive
selection depends on oEnsures reliability only
chance in case of large sample
oAs the size of the size
sample increases, oCases selected are
becomes increasingly widely dispersed; time
representative of the & cost of collecting data
population becomes too large
Ms. Shaheen, Assistant Professor, IP 12
E, OU Campus, Hyderabad
2. SYSTEMATIC SAMPLING
 Used in cases where complete list of the
population is available
 A list of N elements in the population is
compiled
 A sampling size n is chosen

 Then calculate
k=N/n
 Select any random number s from the
1st group

 Then the other elements selected are

s+k, s+2k, s+3k…

Ms. Shaheen, Assistant Professor, IP 13

E, OU Campus, Hyderabad
Example: - If a complete list of 1000 students of a
college is available & if we want to draw a sample of
200 students, this means we must take every 5th item
(k=5). First item between 1 & 5 shall be selected at
random. Second item would be 8th student; third
item, 13th and so on

Advantages Disadvantages
Convenient to adopt Sample does not have
Time & work involved equal chance of being
are relatively smaller selected
Problem of introducing a
sampling error

Ms. Shaheen, Assistant Professor, IP 14

E, OU Campus, Hyderabad
Example
 Suppose the sales division of Computer Graphic, Inc.,
needs to quickly estimate the mean dollar revenue per
sale during the past month. It finds that 2000 sales
invoices were recorded and stored in file drawers, &
decides to select 100 invoices to estimate the mean
dollar revenue.
 Numbering process would be time consuming task, so
we use systematic random sampling.
– First k is calculated. K=(2000/100)=20. so we would select
every 20th invoice from the file drawers.
– Random sampling is used in the selection of the first
choice. A number between 1 & 20 is selected for first
sample.
– Say random number 18 was selected.
– Starting with 18th invoice every 20th invoice (18, 38, 58 etc.)
would be selected as sample.
– If the invoices were filed in order of increasing sales, SRS
would not guarantee a random sample.

Ms. Shaheen, Assistant Professor, IP 15

E, OU Campus, Hyderabad
 3. STRATIFIED SAMPLING
 Also reffered to as restricted
random sampling
 Population is divided into
subgroups, called strata, & a
sample is randomly selected from
each stratum.
 Gurantees each group is
reperesented in the sample.

 Then we use one of the 2

approaches:
 Proportionate: - selecting number
of observations depending on
strata
 Disproportionate: - drawing an
equal number of observations
from each stratum
Note: With either approach, it guarantees that every
element in the population has a chance of being
selected.

Ms. Shaheen, Assistant Professor, IP 16

E, OU Campus, Hyderabad
Example: - Human population may be fivided into different
strata on the basis of sex, age groups, occupations, education or
regions.

Advantages Disadvantages

Ensures greater accuracy if Utmost care should be take

the stratum consists of while dividing the
uniform or homogeneous population into various
items strata.
Concentrated Time consuming &
geographically; time & expensive
expense reduced In the absence of skilled
labor it is difficult to select
items at random.

Ms. Shaheen, Assistant Professor, IP 17

E, OU Campus, Hyderabad
Example
 We might study the advertising expenditures for the 352 largest
companies in US. Suppose the objective of the study is to determine
whether firms with high returns on equity spent more of each sales
dollar on advertising than firms with a low return or deficit.
 To make sure that the sample is a fair representation of the 352
companies, the companies are grouped on percent return on equity.
Stratified sampling will guarantee that atleast one firm in strata 1 and
strata 5 are represented in the sample. Lets say 50 firms are selected
for intensive study. Then 1 (0.02*50) firm from stratum 1 is selected, 5
(0.10*50) firms from stratum2 would be selected and so on.

Stratum Profitability No. of Relative Number

(return on equity) firms Frequency Sampled
1 30 % and over 8 0.02 1
2 20 up to 30% 35 0.10 5
3 10 up to 20% 189 0.54 27
4 0 up to 10% 115 0.33 16
5 Deficit 5 0.01 1
Total 352 1.00 50

Ms. Shaheen, Assistant Professor, IP 18

E, OU Campus, Hyderabad
 4. CLUSTER SAMPLING
 Often employed to reduce the
cost of sampling a population
scattered over a large
geographical area.

 The population is partitioned

into clusters using naturally
occuring geographic
boundaries. Then clusters are
randomly selected & a sample
is collected by randomly
selecting from each cluster.

 Also called as Area Sampling

Ms. Shaheen, Assistant Professor, IP 19

E, OU Campus, Hyderabad
Example
 Suppose we want to determine the views of residents in
AP about State & Central environment protection
policies.
– Selecting a random sample residents in AP and personally
contacting each one would be time consuming and very
expensive
– Instead, we could employ cluster sampling by subdividing
the state into small units – either cities or regions. These
are called Primary Units.
– Suppose we divide the sate into 12 primary units, then
selected at random 4 regions – 2, 7, 4, and 12 – and
concentrated efforts in these primary units.
– We can take a random sample of the residents in each of
these regions and interview them.

Ms. Shaheen, Assistant Professor, IP 20

E, OU Campus, Hyderabad
 Advantages Disadvantages

Reduced costs Less precision

Used when there is Inference can be difficult

considerable variation within
each group but groups
similar to each other

Higher feasibility

Ms. Shaheen, Assistant Professor, IP 21

E, OU Campus, Hyderabad
 NON-PROBABLISTIC
SAMPLING

CONVINIENCE JUDGEMENTAL QUOTA SNOW BALL

SAMPLING SAMPLING SAMPLING SAMPLING

Ms. Shaheen, Assistant Professor, IP 22

E, OU Campus, Hyderabad
5. CONVINIENCE SAMPLING
 Sometimes called grab or opportunity sampling
 This is the method of choosing items arbitrarily and
in an unstructured manner from the frame
 The sample is selected according to the convenience
of the investigator

Advantages Disadvantages

Cheapest method Selection is biased &

Quickest method unsatisfactory
Non representative ness
Inference is not possible

Ms. Shaheen, Assistant Professor, IP 23

E, OU Campus, Hyderabad
 6. JUDGEMENTAL SAMPLING
 Selection is based exclusively on the judgment
of the researcher/ investigator
 Mailnly used in marketing for promotional
activities.

 Example: - If a sample of 10 students is to be

selected from a class of 60 for analysing the
spending habits of students, the investigator
would select 10 students who, in his opinion,
represent the class

 Though simple, it is not scientific because the

results may be considerably affected by
personal prejudice or bias of the investigator

Ms. Shaheen, Assistant Professor, IP 24

E, OU Campus, Hyderabad
 Advantages Disadvantages

Low cost Non representative ness

Inference is not possible
Subjective (because the
Quick observations are selected
depending upon the
statement considered for
investigation)

Ms. Shaheen, Assistant Professor, IP 25

E, OU Campus, Hyderabad
APPLICATIONS
 Used in solving many types of economic &
business problems such as:
Used when size of sample is small.
In solving everyday business problems & making
public policy decisions, executives or officials are
often pressed for time & cannot wait for
probability sample designs. Enables businessmen
& government officials to arrive at solutions to
their urgent problems.
Used to conduct Pilot studies.

Note: - Success of this method depends

upon the excellence in judgment.
Ms. Shaheen, Assistant Professor, IP 26
E, OU Campus, Hyderabad
 7. QUOTA SAMPLING
 Quotas are set up according to the given criteria but,
within the quotas the selection of sample items
depends on personal judgment.
 Applies a “restricted judgmental sampling”, so that
quotas in the sample are the same of those in the
population.
 Example: - In a radio listening survey, interviewers
may be told to interview 500 people living in a
certain area & that out of every 100 persons
interviewed 60 are to be housewifes, 25 farmers &
15 children under the age of 15. Within this quotas,
the interviewer is free to select the people
interviewed.
Ms. Shaheen, Assistant Professor, IP 27
E, OU Campus, Hyderabad
 Advantages Disadvantages

Cheapest method There is no guarantee

that the sample is
representative

Many sources of
Quickest method
selection bias

Ms. Shaheen, Assistant Professor, IP 28

E, OU Campus, Hyderabad
8. SNOWBALL SAMPLING
 Snowball sampling is a technique for
developing a research sample where
existing study subjects recruit future
subjects from among their acquaintances.
 Thus the sample group appears to grow
like a rolling snowball
 This sampling technique is often used in
hidden populations which are difficult for
researchers to access
 Because sample members are not selected
from a sampling frame, snowball samples
are subject to numerous biases

Ms. Shaheen, Assistant Professor, IP 29

E, OU Campus, Hyderabad
 Advantages Disadvantages

Lower costs
Inference is not possible
Low variability

Useful for “rare”

populations

Ms. Shaheen, Assistant Professor, IP 30

E, OU Campus, Hyderabad
MERITS OF SAMPLING METHODS
 Less time
 Less cost
 More reliable results
 More detailed information
 Destructive nature of certain tests
 Only a few chalks out of a certain lot are
tested for ascertaining the breaking strength
 Only few seeds are tested for germination
prior to planting season.

Ms. Shaheen, Assistant Professor, IP 31

E, OU Campus, Hyderabad
LIMITATIONS OF SAMPLING
 Must be carefully planned and executed;
otherwise results obtained may be inaccurate
& misleading.

 Shortage of experts in the sampling field is a

serious hurdle in the way of reliable statistics.

 If the information is required for each & every

unit, a detailed survey has to be done

Ms. Shaheen, Assistant Professor, IP 32

E, OU Campus, Hyderabad
 Sampling Error
 Def: - Difference between a sample statistic and its
corresponding population parameter.
 Example: - Jane and Joe operate the Tara Inn, a
bed & breakfast in Hyderabad. There are 8 rooms
available for rent. Listed is the no. of these 8 rooms
rented each day during June 2009. Calculate the
mean of each sample & compare it to the
population mean. What is the sampling error in
each case?
 Population is the no. of rooms rented each of the
30 days in June 2009.

Ms. Shaheen, Assistant Professor, IP 33

E, OU Campus, Hyderabad
Ju 1 2 3 4 5 5 7 8 9 10 11 12 13 14 15
ne
Re 0 2 3 2 3 4 2 3 4 7 3 4 4 4 7
nt
als
Ju 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
ne
Re 0 5 3 6 2 3 2 3 6 0 4 1 1 3 3
nt
als

During the month there were a total of 94 rentals. So mean

number of units rented per night is µ = (94/30) = 3.13
Random sample of 5 nights is x = (4+7+4+3+1/5) = 3.80
Sampling error is µ - x = 3.80 – 3.13 = 0.67
Therefore 0.67 is the sampling error made in estimating
the population mean.
Ms. Shaheen, Assistant Professor, IP 34
E, OU Campus, Hyderabad
 SAMPLING & NON-SAMPLING
ERROR

 Error: - Refers to the difference between the value of a

Statistic & that of corresponding Parameter

 Sampling Error:- Error arising due to drawing inferences

about the population on the basis of few observations
(sample)

Ms. Shaheen, Assistant Professor, IP 35

E, OU Campus, Hyderabad
SAMPLING ERRORS
 Even if utmost care has been taken in selecting a sample,
the results derived from the sample may not be
representative of the population from which it is drawn.

 Sampling errors are of 2 types – Biased and Unbiased

 Biased Errors: - These errors arise from any bias in
selection, estimation, etc.
 Unbiased Errors: - These errors arise due to chance
difference between the members of population
included in the sample &those not included.

Ms. Shaheen, Assistant Professor, IP 36

E, OU Campus, Hyderabad
 CAUSES OF BIAS
 Bias may arise due to:
– Faulty process of selection
– Faulty work during collection of information
– Faulty methods of analysis
1. Faulty process of selection: - Faulty selection of
sample may give rise to bias in a number of ways, such
as:
– Deliberate selection of ‘representative’ sample
– Conscious or Unconscious bias in the selection
of a random sample
– Substitution – An item is replaced in place of an
item chosen
– Non-response – If all the items to be included in
the sample are not covered, there will be bias
– An appeal to the vanity of the person questioned
may give rise to another kind of bias. Example: -
“Are [Link],
goodAssistant
student?
E, OU Campus, Hyderabad
Yes”
Professor, IP 37
 Conti..
2. Bias due to Faulty collection of Data: - Bias may arise
due to Improper formulation of the decision problems,
wrongly defining the population, specifying the wrong
decision, etc. Biased observations may result from a
poorly designed questionnaire, an ill-trained
interviewer, failure of a respondent’s memory etc.

3. Bias in Analysis: - Such type of bias can be avoided

by adopting the proper methods of analysis

Ms. Shaheen, Assistant Professor, IP 38

E, OU Campus, Hyderabad
Methods of reducing Sampling
Errors

Once the absence of bias has been

ensured, attention should be given
to the random sampling errors.
Apart from reducing errors of bias
the simplest way of increasing the
accuracy is to increase its size.

Ms. Shaheen, Assistant Professor, IP 39

E, OU Campus, Hyderabad
NON-SAMPLING ERRORS
 In the processing of data tabulation errors may be
committed affecting the final results. Errors arising
in this manner are termed as non-sampling errors.
 Non- sampling errors may arise from one or more of
the following factors:
1. Data being inadequate & inconsistent
2. Inaccurate/ inappropriate methods of interview,
observation or measurement.
3. Lack of trained & experienced investigators.
4. Lack of adequate inspection & supervision of
primary staff.
5. Errors due to non-response.
6. Errors in data processing operations – coding,
punching, verification, tabulation etc.
7. Inadequate scrutiny of the basic data.
Ms. Shaheen, Assistant Professor, IP 40
E, OU Campus, Hyderabad
Ms. Shaheen, Assistant Professor, 41
IPE, OU Campus, Hyderabad
Sampling
 Sampling
• A statistic is a characteristic of a sample; a parameter is a
characteristic of a population.

• Example: - The mean height in inches of all PGDBM students in

India is 60 inches. 60 inches is the characteristic of the population
“PGDBM students” & can be called population parameter.
– If we say that the mean height of Prof. Mishra’s PGDBM
students in IPE is 60 inches, then we are using 60inches to
describe the characteristic of the sample “Prof. Mishra’s
PGDBM students “. In this case, 60 inches would become a
sample statistic.

• Sample statistics allow to make inference about the population

parameters, through estimation and hypothesis testing.

Ms. Shaheen, Assistant Professor, 43

IPE, OU Campus, Hyderabad
 Types of Sampling
• There are 2 methods of selecting samples from populations:
– Nonrandom or Judgment Sampling
– Random or Probability sampling

• Random or Probability sampling: - All the items in the population

have a chance of being chosen in the sample.

• Nonrandom or Judgment Sampling: - Personal knowledge & opinion

are used to identify the items from the population that are to be
included in the sample.

• Note: - If a study uses judgmental sampling & loses a significant

degree of representativeness, it will have purchased convenience at
too high a price.

Ms. Shaheen, Assistant Professor, 44

IPE, OU Campus, Hyderabad
 Representativeness

• A sample can be considered as

“representative” when it is expected to
exhibit the average properties of the
population.

Ms. Shaheen, Assistant Professor, 45

IPE, OU Campus, Hyderabad
 The sampling design process
1. Define the target population, its elements and
the sampling units.

2. Determine the sampling frame (list).

3. Select a sampling technique

• Sampling with/without replacement.
• Probability/Nonprobability sampling.

4. Determine the sample size.

5. Execute the sampling process.

Ms. Shaheen, Assistant Professor, 46
IPE, OU Campus, Hyderabad
 The sampling techniques
• Probabilistic/ Random samples

– Simple random sampling

– Systematic sampling
– Stratified sampling
– Cluster sampling
– Other sampling techniques

• Nonprobabilistic samples

– Convenience sampling
– Judgmental sampling
– Quota sampling
– Snowball sampling
Ms. Shaheen, Assistant Professor, 47
IPE, OU Campus, Hyderabad
 Simple random sampling
• Selects samples by methods that allow each possible sample to
have an equal probability of being picked & each item in the
entire poulation to have an equal chance of being incuded in the
sample.
• Every element is selected independently from other elements.

•No possibility of personal bias, •Task of preparing slips is time-

since selection depends on consuming & expensive.
chance.
•Size of sample required to
•As the size of the sample ensure reliability is usually
increases, becomes increasingly large.
representative of the
•Cases selected are widely
population.
dispersed; time & cost of
collecting data becomes too
large.
Ms. Shaheen, Assistant Professor, 48
IPE, OU Campus, Hyderabad
• Methods of obtaining a Simple Random Sample: -

1. Lottery Method: - All items of the population are numbered/

named on separate slips of paper of identical size, color &
shape. Slips are folded & mixed up in a container. Blind-fold
selection is then made.
– Selection of items thus depends entirely on chance.
– As the size of the population increases, method becomes quite
cumbersome to use.

2. Table of Random Numbers: - Alternative method of random

selection is that of using the table of random numbers.
– A number of random number tables are available such as:
1. Tippets table of random numbers
2. Fisher & Yates numbers &
3. Kendall & Babington smith numbers.

3. These numbers can be generated by computer programmed

to scrambble numbers.
Ms. Shaheen, Assistant Professor, 49
IPE, OU Campus, Hyderabad
 Systematic sampling
• Used in cases where a complete list of the population is available.

• A list of N elements in the population is compiled, ordered according

to a specified variable.
• A sampling size n is chosen
• A systematic step of k=N/n is set
• A random number s between 1 and N is extracted and represents
the first element to be included.
• Then the other elements selected are s+k, s+2k, s+3k…

• Example: - If a complete list of 1000 students of a college is

available & if we want to draw a sample of 200 students, this means
we must take every 5th item (k=5). First item between 1 & 5 shall be
selected at random. Second item would be 8th student; third item,
13th and so on.

Ms. Shaheen, Assistant Professor, 50

IPE, OU Campus, Hyderabad
•More convenient to adopt. •Each sample does not have equal
chance of being selected.
•Time & work involved are
relatively smaller. •Problem of introducing an error
into the sample process.
•If population is sufficiently large,
results yielded is similar to those •Example: - We are sampling paper
obtained by Proportional stratified waste produced by households, &
sampling. we decided to sample 100
households every Monday.
•Chances are high that our sample
would not be representative,
because Monday’s thrash would
very likely include Sunday’s
newspaper.

Ms. Shaheen, Assistant Professor, 51

IPE, OU Campus, Hyderabad
 Stratified sampling
• Restricted random method.
• We divide the population into relatively homogeneous groups, called
strata.

• Then we use one of the 2 approaches:

– Proportionate: - Either we select at random from each stratum a
specified number of elements corresponding to the proportion of
that stratum in the population as a whole.
– Disproportionate: - We draw an equal number of elements from
each stratum & give weight to the results according to the
stratums proportion of total population.

• With either approach, it guarantees that every element in the

population has a chance of being selected.
• Appropriate when the population is already divided into groups of
different sizes.

Ms. Shaheen, Assistant Professor, 52

IPE, OU Campus, Hyderabad
• Example (Proportional sample): - If the population is divided into 4
strata, their respective sizes being 15%, 10%, 20%, 55% of the
population & a sample of 1000 is to be drawn.
From stratum 1 (15% of 1000 = 0.15) = 150 items
,, ,, 2 (10% of 1000 = 0.10) = 100 items
,, ,, 3 (20% of 1000 = 0.20) = 200 items
,, ,, 4 (55% of 1000 = 0.55) = 550 items
Note: - Not efficient when there is considerable variation in different
strata.

• Example (Disproportionate sample): - in conducting an all-India

survey of a new product, all states of India may be taken as strata. If
potential consumers of a given state are 10% of all the consumers,
we may take 1% or 2% of our sample from that state. However, the
outcome of another state is highly doubtful we may decide to give it
a much greater representation in the sample from its relative size.

Ms. Shaheen, Assistant Professor, 53

IPE, OU Campus, Hyderabad
•Ensures greater accuracy if the •Utmost care required in dividing
stratum consists of uniform or the population into various
homogeneous items. strata. Involves considerable
time & expense.
•Concentrated geographically;
time & expense reduced. •Items should be selected at
random from each stratum,
which may be difficult in the
absence of skilled supervisors.

Ms. Shaheen, Assistant Professor, 54

IPE, OU Campus, Hyderabad
 Cluster sampling
• The population is partitioned into groups or clusters, & then
select a random sample of these clusters.
• Elements within the cluster should be as heterogeneous as
possible with respect to the variable of interests (e.g. area
sampling)
1. A random sample of clusters is extracted through SRS (with
probability proportional to the cluster size)
– All the elements of the cluster are selected (one-stage)
– A probabilistic sample is extracted from the cluster (two-
stage cluster sampling)

–Reduced costs –Less precision

–Used when there is considerable –Inference can be difficult
variation within each group but groups
similar to each other.
–Higher feasibility Ms. Shaheen, Assistant Professor, 55
IPE, OU Campus, Hyderabad
• Example: - If a market research team is attempting to determine by
sampling the average number of television sets per household in a
large city, they could use the city map to divide the territory into
blocks & then choose a certain number of blocks for interviewing.
Every household in each of these blocks would be interviewed.

• Note: -
– We use Stratified Sampling when each group has small variation within
itself but there is a wide variation within the groups.
– Cluster Sampling is used when there is a considerable variation within
each group but the groups are essentially similar to each other.

Ms. Shaheen, Assistant Professor, 56

IPE, OU Campus, Hyderabad
Non probabilistic samples
 Convenience sampling
• Also called the chunk. A chunk is a fraction of population taken for
investigation because of its convenient availability.
• Thus, chunk is selected neither by probability nor by judgement but
by convenience.
• A sample is obtained from readily available lists such as telephone
directories or automobile registers is a convenience sample & not a
random sample.
• Only “convenient” elements enter the sample.

–Cheapest method –Selection is biased & unsatisfactory

–Quickest method –Non representativeness
–Inference is not possible

Ms. Shaheen, Assistant Professor, 58

IPE, OU Campus, Hyderabad
 Judgmental sampling
• Selection is based exclusively on the judgment of the researcher/
investigator.
• Example: - If a sample of 10 students is to be selected from a class
of 60 for analysing the spending habits of students, the investigator
would select 10 students who, in his opinion, represent the class.
• Though simple, is not scientific because the results may be
considerably affected by personal prejudice or bias of the
investigator.

–Low cost –Non representativeness

–Quick –Inference is not possible
–Subjective

Ms. Shaheen, Assistant Professor, 59

IPE, OU Campus, Hyderabad
• Used in solving many types of economic & business problems
such as:
1. Used when size of sample is small.
2. In solving everyday business problems & making public policy
decisions, executives or officials are often pressed for time & cannot
wait for probability sample designs. Enables businessmen &
government officials to arrive at solutions to their urgent problems.
3. Used to conduct Pilot studies.

Note: - Success of this method depends upon the excellence in

judgment.

Ms. Shaheen, Assistant Professor, 60

IPE, OU Campus, Hyderabad
 Quota sampling
• A type of judgement sampling.
• Quotas are set up according to the given criteria but, within
the quotas the selection of sample items depends on
personal judgment.
• Applies a “restricted judgmental sampling”, so that quotas in
the sample are the same of those in the population.
• Example: - In a radio listening survey, interviewers may be
told to interview 500 people living in a certain area & that out
of every 100 persons interviewed 60 are to be housewifes,
25 farmers & 15 children under the age of 15. Within this
quotas, the interviewer is free to select the people
interviewed.

–Cheapest method –There is no guarantee that the sample

–Quickest method is representative
–Many sources of selection bias
Ms. Shaheen, Assistant Professor, 61
IPE, OU Campus, Hyderabad
 Snowball sampling
• At first small sample is selected randomly.
• Respondents are asked to identify others who belong to the
population of interests.
• The referrals will have demographic and psychographic
characteristics similar to the referrers.

–Lower costs –Inference is not possible

–Low variability
–Useful for “rare” populations

Ms. Shaheen, Assistant Professor, 62

IPE, OU Campus, Hyderabad
 Merits of Sampling Method
1. Less time
2. Less cost
3. More reliable results
– Always possible to determine the extent of sampling errors.
– Other types of errors to which a survey is subject, such as inaccuracy
of information, incompleteness of returns. etc., are likely to be more
serious in a complete census.
4. More detailed information
5. Destructive nature of certain tests
• Only a few chalks out of a certain lot are tested for ascertaining the
breaking strength.
• Only few seeds are tested for germination prior to planting season.

Ms. Shaheen, Assistant Professor, 63

IPE, OU Campus, Hyderabad
 Limitations of Sampling
• Must be carefully planned and executed; otherwise results obtained
may be inaccurate & misleading.

• Shortage of experts in the sampling field is a serious hurdle in the

way of reliable statistics.

• If the information is required for each & every unit, a complete

enumeration survey is necessary.

Ms. Shaheen, Assistant Professor, 64

IPE, OU Campus, Hyderabad
 Sampling & Non-Sampling Errors
• Error: - Refers to the difference between the value of a Statistic &
that of corresponding Parameter.

• Sampling Error: - Error arising due to drawing inferences about the

population on the basis of few observations (sample).

• Note: - Sampling error is almost non-existent in a complete

enumeration survey, since the whole population is surveyed.

Ms. Shaheen, Assistant Professor, 65

IPE, OU Campus, Hyderabad
 Sampling Errors
• Even if utmost care has been taken in selecting a sample, the
results derived from the sample may not be representative of the
population from which it is drawn.

• Sampling errors are of 2 types – Biased and Unbiased

1. Biased Errors: - These errors arise from any bias in selection,

estimation, etc.
2. Unbiased Errors: - These errors arise due to chance difference
between the members of population included in the sample
&those not included.

Ms. Shaheen, Assistant Professor, 66

IPE, OU Campus, Hyderabad
 Selection bias
• Improper selection of sample units (ignoring a
relevant “control variable” that generate bias), so that
the values observed in the sample are biased and
the sample is not representative.

• Example: - A survey is conducted for measuring goat

milk consumption, but the interviewers just select
people in urban areas, that on average drink less
goat milk.

• Bias: - It exists when the value of a sample statistic

shows a persistent tendency to deviate in one
direction from the value of the parameter.

Ms. Shaheen, Assistant Professor, 67

IPE, OU Campus, Hyderabad
 Causes of Bias
• Bias may arise due to:
1. Faulty process of selection
2. Faulty work during collection of information
3. Faulty methods of analysis
1. Faulty process of selection: - Faulty selection of sample may give
rise to bias in a number of ways, such as:
• Deliberate selection of ‘representative’ sample.
• Conscious or Unconscious bias in the selection of a
random sample
• Substitution – An item is replaced in place of an item
chosen.
• Non-response – If all the items to be included in the
sample are not covered, there will be bias.
• An appeal to the vanity of the person questioned may give
rise to another kind of bias. Example: - “Are you a good
student? Yes”
Ms. Shaheen, Assistant Professor, 68
IPE, OU Campus, Hyderabad
2. Bias due to Faulty collection of Data: - Bias may arise due to Improper
formulation of the decision problems, wrongly defining the population,
specifying the wrong decision, etc. Biased observations may result from a
poorly designed questionnaire, an ill-trained interviewer, failure of a
respondent’s memory etc.

3. Bias in Analysis: - Such type of bias can be avoided by adopting the

proper methods of analysis.

Method of reducing Sampling Errors: -

– Once the absence of bias has been ensured, attention

should be given to the random sampling errors.
– Apart from reducing errors of bias the simplest way of
increasing the accuracy is to increase its size.

Ms. Shaheen, Assistant Professor, 69

IPE, OU Campus, Hyderabad
 Non-Sampling Errors
• In the processing of data tabulation errors may be committed
affecting the final results. Errors arising in this manner are termed
as non-sampling errors.

• Non- sampling errors may arise from one or more of the following
factors:
1. Data being inadequate & inconsistent
2. Inaccurate/ inappropriate methods of interview, observation or
measurement.
3. Lack of trained & experienced investigators.
4. Lack of adequate inspection & supervision of primary staff.
5. Errors due to non-response.
6. Errors in data processing operations – coding, punching, verification,
tabulation etc.
7. Inadequate scrutiny of the basic data.
Ms. Shaheen, Assistant Professor, 70
IPE, OU Campus, Hyderabad