Unit III: Sampling Techniques
SK Shrestha
Objectives: This unit intends to introduce the
concept of sampling, census, their differences,
types of sampling and their uses in different
scenario of real data problem.
Inferential Statistics
• The statistical device in which someone is
drawn conclusion about the population from
the sample is the inferential statistics.
• Main functions of Inferential statistics are:
• To estimate the value of population
parameter from the sample.
• To test the significance of the population
parameter by testing of hypotheses.
Population:
Data set which is under the investigation is
said to be population.
It is the collection of all the items or objects
which are considered under the investigation.
It is entire group of items or objects or people
or event from which sample is drawn.
Example: Visitors of the hospitals, customers of
the departmental store, trees in the jungle,
students of the university etc.
Sample:
The representative part of the population is
called sample.
It is the subset of the population under study.
Information collection or measurement is
done on the sample and generalize for the
population.
Uses of sample methods
To make a new policy or change the existing
policies by the government or organizations
for the betterment .
To improve the better quality of the product by
the industries.
To provide the better service for the customers.
To develop the plan for the betterment of the
organization.
To help for decision making process.
Sampling:
• The process of selecting representative
part of the study population is the
sampling.
• It is a tools for drawing part of the items as
sample from whole population.
Why Sample?
Get information about large populations
To minimize cost of investigation.
To reduce the time of investigation.
To get more accuracy i.e. Can do a better job
of data collection.
When it’s impossible to study the whole
population i.e. population is infinite.
When the investigation is destructive in
nature.
Census and Sample Surveys
The total number of households in Nepal is about 5.4 million
(population census of 2011). It is possible to enumerate from the
census conducted by government of Nepal.
Data Collection in Statistics
Any investigation is based upon the statistical
calculations or interpretation lies in the collection of
data.
Two methods of data collection are
(i) Census method and (ii) Sampling method.
Census Method
Census method: The enumeration of all the items or members of
the population under the studied is known as census method. It is
the method of statistical enumeration where all members of the
population are studied.
Every country wants to a more realistic level, a country wants to
keep maximum information and records about all households by
using census method. The government of Nepal conducts the
census of Nepal in every 10 years. From the census the
information are collected from the households related to their
incomes, the members of earners, the size of family, the
information of demographic variables like number of births, number
of death, age etc are collected. In this method no one can leave
from the enumeration and all the units are taking into consideration.
Sampling Method
Due to lack of enough time and money sampling method is
feasible of the investigation.
Information is collected from some representative part of the
whole population.
This representative part which contains some units of the
whole population is said to be the sample.
For example, if someone want to carry out a survey to find
out visitors feedback about the service of the Prime Bank,
then all the visitors of that bank would form the ‘population’
for your study and if only small number of visitors (say 30)
are observed or survey for that investigation then it is
sample method.
Parameter Census survey Sample survey
Definition A statistical method
A statistical method that studies only
that studies all a representative
the units or group or part of
members of a the population,
population. and not all its
members.
Calculation Total/Complete Partial
It is a time-
Time It is a quicker
consuming
involved process.
process.
It is a relatively
Cost It is a costly
inexpensive
involved method.
method.
The results are
The results
relatively
obtained are
inaccurate due to
accurate as each
leaving out of
Accuracy member is
items from the
surveyed. So,
sample. The
there is a
resulting error is
negligible error.
large.
Reliability Highly reliable Low reliability
The smaller the
Error Not present sample size, the
larger the error.
This method is This method is
suited for suited for
Relevance
heterogeneous homogeneous
data. data.
How to select a Sample?
* Determine the population or target population.
* Make a sampling frame or study population.
* Determine the sampling unit.
* Fix the sampling techniques.
* Determine the sample size.
* Execute the sampling process.
A good sample is one which is:
Small in size.
It provides adequate information about
the whole population.
It takes less time to collect and is less
costly.
Target Population: The population to be studied
or to which the investigator wants to generalize
his results
Sampling Unit: Smallest unit from which sample
can be selected
Sampling frame: List of all the sampling units
from which sample is drawn
Sampling scheme: Method of selecting sampling
units from sampling frame
Types of sampling:
Probability Sampling or Random sampling
(with and without replacement)
Non-Probability or Non-random sampling
Definition: The sampling techniques in which each sample
unit has included as sample with some law of probability. The
chances or probability of being included as sample unit has an
equal or unequal chances. When every element in the
population have the same probability of selection, this is
known as an 'equal probability of selection' (EPS) design. It is
also referred to as 'self-weighting' because all sampled units
are given the same weight.
Random selection means unbiased representation of the
whole population. If the sample does not represent the
population then there is variation between sample value
(statistic) and population value (parameter) which is said to be
sampling error.
The selection of units in the sample from a population is
governed by the laws of chance or probability.
There are verities of methods for the selection of random
sample.
(a) By using Lottery method
(b) By using random number table
[See appendix for random number table]
(c) By using computer software
By using excel: =RAND(), =RANDBETWEEN(1, 100),
Data analysis tools random number generator
By using R: sample(1:1000, 10)
[1] 587 215 396 457 702 438 498 676 368 82
Random or Probability Sampling:
Five types of random sampling:
1. Simple Random sampling
(a) Simple Random Sampling with Replacement(SRSWR):
Random sampling with replacement means once a unit is
selected as a sample then it is again included as a sample. It
is the case of independent.
(b) Simple Random Sampling without Replacement
(SRSWOR): Random sampling without replacement: Once a
unit is selected as a sample has no chance of selecting in the
second time. That is each unit of sample of the population
has only one chance to be chosen in the sample. It is the
case of dependent.
Random Sampling
The process of selecting sample from the pre
defined population such a way that each unit or
member of the population has equal chance and of
being selected as a sample.
All the members of the sampling frame have equal
chance.
Merits:
It is simple, easy mostly used sampling techniques.
t is useful for the small and homogeneous
population.
It is unbiased techniques.
Demerits:
If the exhaustive list of all members units are not
available then it is not feasible.
If the members of the population are heterogeneous
then it may not represent properly (that is biased
representation)
If the population size is large then it is most costly
and time consuming.
Process of selecting simple random
sample:
Identify the study population and make a
list of all units by assigning a number
individually.
Determine the sample size.
Select the sample either by lottery method
or by using random number or computer
software.
Simple Random Sampling
[20 samples are selected from 200 study population]
1 2 3 4 5 6 7 8 9 10
11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30
31 32 33 34 35 36 37 38 39 40
41 42 43 44 45 46 47 48 49 50
51 52 53 54 55 56 57 58 59 60
61 62 63 64 65 66 67 68 69 70
71 72 73 74 75 76 77 78 79 80
81 82 83 84 85 86 87 88 89 90
91 92 93 94 95 96 97 98 99 100
101 102 103 104 105 106 107 108 109 110
111 112 113 114 115 116 117 118 119 120
121 122 123 124 125 126 127 128 129 130
131 132 133 134 135 136 137 138 139 140
141 142 143 144 145 146 147 148 149 150
151 152 153 154 155 156 157 158 159 160
161 162 163 164 165 166 167 168 169 170
171 172 173 174 175 176 177 178 179 180
181 182 183 184 185 186 187 188 189 190
191 192 193 194 195 196 197 198 199 200
2. Stratified Sampling
The sampling technique in which entire population is divided into
homogenous subgroups called strata and then sample units are
drawn randomly from each stratum.
Two stages are used for this technique:
(i) In first stage divide the whole population into
different strata.
(ii) In second stage Select the sample from each
stratum randomly.
• Strata are divided into some specific criteria for
example: education level(undergraduate,
graduate), geographical location (Mountain, Hill,
Terai), Occupation(Technical, non technical),
Profession(Teachers, Businessmen, Doctors,
Engineers), age (child, young, adult) gender
(male, female) etc.
• Strata are mutually exclusive and exhaustive.
Merits:
It is useful for heterogeneous population.
It reflects the characteristics of the population
properly.
It has more scope (covers more area).
It provides the efficient and reliable information
without increasing cost.
Demerits:
It is not feasible for all investigations.
There may overlap of the strata if the objects
having multiple characteristics.
It requires the complete or exhaustive list of the
entire population.
Process for the stratified random sampling:
Identify the entire population and make a list.
Determine required sample size.
Identify the attribute for the stratification.
Divide all members of the population as members
of one identified subgroup or strata
Randomly select an appropriate number of
members or units as a sample from each of the
subgroups.
3. Systematic Sampling
Sampling techniques in which sample units from the larger
population are selected with some fixed or periodic interval or
equal space from the consecutive list of that population. The
first selected unit is said to be random start and others are
selected automatically in the certain interval. The interval or
space is said to be sampling interval. It is calculated by dividing
the size of population by the size of the sample required for the
study.
Process of Systematic Sampling:
Define the population and make a sampling frame.
Arrange the objects in the consecutive orders and
assign the number.
Determine the population size and desired sample
size.
Calculate the sampling interval for the selection.
Select a starting number randomly.
Collects the number corresponding to the sampling
interval.
Systematic Sampling
𝑁 200
[N= 200, n= 20, K = = = 10]
𝑛 20
1 2 3 4 5 6 7 8 9 10
11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30
31 32 33 34 35 36 37 38 39 40
41 42 43 44 45 46 47 48 49 50
51 52 53 54 55 56 57 58 59 60
61 62 63 64 65 66 67 68 69 70
71 72 73 74 75 76 77 78 79 80
81 82 83 84 85 86 87 88 89 90
91 92 93 94 95 96 97 98 99 100
101 102 103 104 105 106 107 108 109 110
111 112 113 114 115 116 117 118 119 120
121 122 123 124 125 126 127 128 129 130
131 132 133 134 135 136 137 138 139 140
141 142 143 144 145 146 147 148 149 150
151 152 153 154 155 156 157 158 159 160
161 162 163 164 165 166 167 168 169 170
171 172 173 174 175 176 177 178 179 180
181 182 183 184 185 186 187 188 189 190
191 192 193 194 195 196 197 198 199 200
Merits:
Easy to select the sample.
More representative of the entire population.
Cost effective
Save time and require less effort.
Demerits:
Exhaustive list of the population is required.
Each member does not get equal chance for the
selection.
The members between two selected number are
ignored.
There may be biased selection if hidden periodicity
is presented on the list of population.
4. Cluster Sampling
Sampling technique in which the entire population is divided into
separate homogeneous called clusters then the clusters selected
randomly then sample units are selected either by single stage or two
stages. In single stage, all the units within the selected cluster are
considered as the sample where as in two stages sampling the
samples are selected from the chosen clusters by simple random or
systematic random technique. That is the researcher investigates
only from the selected clusters.
Process of Cluster Sampling:
Identify the study population.
Determine the variable for making clusters for the
population and divide the population into clusters
such that the members within cluster is
heterogeneous and between cluster is
homogeneous.
Determine the desired sample size and sample
size of each cluster.
Calculate the average number of population
members per cluster
Determine the number of clusters
Process of Cluster Sampling:
Identify the study population.
Determine the variable for making clusters for the
population and divide the population into clusters such that
the members within cluster is heterogeneous and between
cluster is homogeneous.
Determine the desired sample size and sample size of
each cluster.
Calculate the average number of population members per
cluster
Determine the number of clusters
Select required clusters using simple random sampling
Include either all members in each selected cluster (single
stage) or samples are drawn randomly (or systematically)
(two stage)
Merits:
It is cost effective.
Feasible for large population.
Save time for sampling.
Sampling frame is not necessary.
Demerits:
Sampling error is larger than others.
The units can overlap in two or more clusters.
Biased information may collect due to wrong
clusters.
Thank you
