CHAPTER FOUR

Sampling
Design
1
Census and Sample Survey
Population: is the set of people or entities
to which findings are to be generalized.
Census: is collections of data from every
person or entity in the population
The study of all individual members of a
population.
Sample: is a group of units selected from a
population.
2
Cont…
Sampling Technique: is the process of
selecting samples from a population
Sample design: - is the technique or
procedure the researcher adopts for
selecting items for the sample from the
population.
How a sample should be selected?
What size such a sample would be?
3
The Need for Sampling
 There could be resource limitations to study the whole

population.
 Time

 Finance

 Manpower
 In some cases, tests may be destructive.
 For example, when we test the breaking strength of
materials, we must destroy them.
 A census would mean complete destruction of
materials.
 In such a case, we must sample.
4
Cont…
 Sampling provides much quicker results than
does a census.
 Is the only process if the population is infinite.
 There is also an argument that the quality of a
study with sampling is often better than with a
census.
 better interviewing;
 more thorough investigation of missing,
wrong, or suspicious information;
 better supervision; and 5
 Two major principles
underlie all sample design

 The desire to avoid bias in the selection

procedure.

 To achieve the maximum precision.

6
Bias in selection can arise

 If the selection of the sample is done by

some non-random method.
 If the sample frame doesn’t adequately
cover the target population.
 If some sections of the population are
impossible to find or refuse to cooperate.

7
Steps in Sampling Design
While developing a sample design, consider the following
points.
1. Type of universe/population
 Define the population explicitly.
 Realize whether the universe is finite or infinite.
 Finite: E.g. the population of a city, the number of workers in a
factory
 Infinite: E.g. listeners of a specific radio program, number of
stars in the sky

2. Sampling Units
 Is the unit of analysis or case in a population.
 It may be a geographical area such as district, kebele, village, or
a social unit such as family, school, or it may be an individual.
8
Steps in Sampling Design
3. Source list/sampling frame
 A list that contains the names of all elements in the population.
 should be complete, correct, reliable and appropriate
4. Sample size
 The number of items to be selected from the population to
constitute a sample.
 should neither be excessively large, nor too small
 It should be optimum.
 Should fulfills the requirements of efficiency, representativeness
& reliability
 Points to be considered in deciding sample size:
 The size of population

 The parameter of interest in the research study

9
 Budgetary constraint
Steps in Sampling Design
5. Parameters of interest
 Consider the question of the specific population
parameters which are of interest.

6. Budgetary constraint
 Have a major impact upon decisions relating to sample
size and sampling method.
 This fact can even lead to non-probability samples.

7. Sampling procedure
 Deciding on the type of sample design the researcher
will use.
 Decide about the technique to be used in selecting the
10
items for the sample
 Types and Causes of Incorrect Inferences
in Sampling

 There are two causes of incorrect inferences: -

i. Systematic bias (Non- sampling error):

 Results from errors in the sampling procedures.
 It cannot be reduced by increasing the sample
size.
 The causes responsible for these errors can be
detected and corrected
 It may include field work problems, interviewer
bias, clerical problems, etc.
11
Cont…
Systematic bias is the result of one or more of
the following factors:
 Inappropriate sampling frame
 Biased representation of the population.
 Non-coverage error or wrong population sampled.
 Defective measuring device
 If the measuring device is in error
 If the questionnaire or the interviewer is biased.
 If the physical measuring device is defective.
 Instrument error.
12
Cont..
 Non-respondents
 If we are unable to sample all the individuals
initially included in the sample.
 Low response rate.
 Indeterminacy principle
 Sometimes individuals act differently when
kept under observation than what they do
when kept in non-observed situations.

13
Cont…
 Natural bias in the reporting of data

 Downward bias in the income data collected by

government taxation department.

 Upward bias in the income data collected by some

social organization.

 Understate incomes if asked about it for tax purposes

 Overstate if asked for social status or their wealth.

14
Cont…
ii. Sampling errors
 Are random variations in the sample estimates
around the true population
 Are errors attributable to sampling not to
census.
 It decreases with the increase in the size of
the sample.

15
Characteristics of a good Sample Design
 Result in a truly representative sample
 Result in small sampling error
 Be viable in the context of funds available for
the research study
 Must enable to control the systematic bias in a
better way
 The results of the sample study can be applied
for the universe with a reasonable level of
confidence.
16
 Types of Sample Design

 Sample designs are basically of two

types:
A. Probability Sampling

B. Non-probability Sampling

17
 Probability Sampling
Designs
 Is a sampling technique in which each
member of a population has a known
non-zero probability of being chosen.
 Equal chance of inclusion in the sample.
 Is also known as random sampling or
chance sampling.

18
 The advantage of probability
sampling
 Sampling bias will be minimized.
 Relies on random process.
 Helps to get consistent and unbiased
estimates of the population parameters.

19
 Common types of Probability
Sampling
 Simple Random Sampling
 Systematic Random Sampling
 Stratified Random Sampling
 Cluster Random Sampling
 Multi-stage Random Sampling

20
Simple Random Sampling
Each unit in the population has equal chance to
be included in the sample.
Easy to accomplish and to explain to others.
Fair way to select a sample.
Reasonable to generalize the results from the
sample back to population.
Appropriate for homogenous population.
Is the best technique for selecting a
representative sample.
21
Simple Random Sampling
Uses lottery method and random number
table method to take a random sample.
For example, if researchers wished to draw a
sample of 50 individuals from a population of
600 customers of HARAR BRANCH OROMIA
INSURANCE COMPANY:
 they could place the 600 names in a container and
blindfolded.
 draw one name at a time until the sample of 50
was selected. 22
Simple Random Sampling
The above procedure is bulky and is
rarely used.
Rather, a random numbers table or
computer-generated list is more common.
The simple way of doing this also is to
use Microsoft Excel to draw the sample.

23
 Systematic Random
Sampling
 Is a technique of selecting sampling units
using a fixed interval on a randomly
ordered list of the sampling frame.
 It involves selection of every kth element
in the sampling frame.

24
 Steps to use Systematic
Random Sampling
i. Number the units in the population from 1 to
N sequentially.
ii. Decide on the sample size “n”.
iii. Determine interval size “k” using, k = N/n.
iv. Randomly select starting point (it must be
between 1 & k).
v. Take every kth element until the sample size
reach to n.
25
Systematic Random Sampling
 For example, if a sample of 200 were to be
selected from a directory of Savings Account
holders with 200,000 listings:
 One would select the first name by randomly
selecting a name from a randomly selected page.
 Then every thousandth name would be selected
until the sample of 200 names was complete.
 If the last page were reached before the desired
number had been selected, the count would
continue from the first page of the directory.
26
Stratified Random Sampling
 If a population from which a sample is to be

drawn does not constitute a homogeneous group

 Involves dividing the population into

homogenous subgroups/strata and then taking a

sample from each subgroup.

 Each of the homogenous subgroup is known as

strata.
 Used when the population is heterogeneous.
27
Stratified Random
Sampling
 Questions relevant in the context of
stratified sampling:
 How to form strata?
 How should items be selected from each
stratum?
 How many items be selected from each
stratum or how to allocate the sample size
of each stratum? 28
The basic steps for stratified sampling

i. Divide the population into subgroups or strata.

ii. Determine sample size of the whole population.
iii. Compute sample proportion for each stratum
based on their population proportion.
iv. Make a separate and complete list of each
stratum.
v. Select sample from each stratum using simple
random sampling or systematic sampling
technique to constitute the sample.
29
Stratified Random
Sampling
 The strata are formed on the basis of common
characteristics of the items.
 The number of items to be selected from each
stratum is done using proportional allocation
method.
 Strata are purposively formed usually based on past
experience and personal judgment of the researcher.
 At times, pilot study may be conducted for
determining a more appropriate and efficient
stratification plan.
30
Example
 In an income study of wage earners in a
community, a true sample would
approximate the same relative number
from each socio economic level of the
whole community.
 If, in the community, the proportion were
15% professional workers, 10% managers,
20% skilled workers, and 55% unskilled
workers. 31
Example
 The sample should include approximately
the same proportions to be considered
representative.
 Within each subgroup a random selection
should be used.
 For a sample of 100, the researcher
would randomly select as follows.

32
 Stratified sample of 100 from a population
of 1,000 employees
Employee Proportio % (Total) Sample
Type Populatio n (n)
(Strata) n (N)
Profession 150 150/1000 15% 15%*100
al workers =15
Managers 100 100/1000 10% 10%*100
=10
Skilled 200 200/1000 20% 20%*100
Workers =20
Unskilled 550 550/1000 55% 55%*100
Workers =55
Total 1,000 100% 100
33
Cluster (Area) Sampling
 If the total area of interest happens to be a big one
 Divide the area in to a number of smaller non-overlapping

areas
 Randomly select a number of these smaller areas (clusters)
 The elements within each cluster are heterogeneous.
 The area or cluster sample is a variation of the simple random

sample that is particularly appropriate:

 when the population of interest is infinite,

 when a list of the members of the population does not

exist, or

 when the geographic distribution of the individuals is

34
Steps in Cluster Sampling

i. Divide the population into clusters.

ii. Randomly select clusters for inclusion in
the overall sample.
iii. Select samples from sampled clusters
again randomly.

35
Example
 If researchers wanted to survey on real estate
investment in 20 African countries from total.
 The researcher may employ cluster sampling
so that the sample may represent all African
countries.
 Thus, from the five sub-regions (clusters), the
researchers randomly select three sub-regions
out of which 20 countries would be again
randomly selected.
36
37
 Multi-Stage Random Sampling

 It is a further development of the principle of

cluster sampling.
 Taking of samples in stages using smaller and

smaller sampling units

 Complex form of cluster sampling
 Involves dividing the population into groups or

clusters
 Then, one or more clusters are chosen at random

and everyone within the chosen cluster are

38
 Multi-Stage Random Sampling
 E.G. we want to investigate the working efficiency of

health stations in Ethiopia, and we want to take a sample

of few health stations for this purpose.

 The first stage is to select large primary sampling unit

such as regional states in country.

 Then we may select certain zones in the second stage.
 Then we may again select districts (woredas).
 Taking a census of all health stations within the selected
districts.
 Randomly sample health stations from each selected
districts. 39
 Non-probability Sampling Designs
 Doesn’t involve random selection of sampling
units.
 Members being selected for sample cannot be
calculated.
 Items for the sample are selected deliberately by
the researcher.
 Relies on judgment of the researcher.
 Conclusions drawn from a study of a non-
probability sample are limited to that sample and
cannot be used for further generalization. 40
Major Non-probability Sampling
Techniques

i. Judgment (purposive) sampling

ii. Convenience (accidental) sampling

iii. Quota Sampling

iv. Referral Sampling

41
 Judgment (Purposive) Sampling

 Takes sample with a purpose in mind.

 A researcher uses his/her judgment to
select people that he feels are representative
of the population.
 Is frequently used in qualitative research
studies.
 In case studies the participants are often
selected because they are typical or particularly
interesting to the researcher.
42
Cont…

 Purposeful sampling allows the

researcher to select those participants:
 who will provide the richest information,
 who are the most interesting, and
 who manifest the characteristics of most
interest to the researcher.

43
Cont…
Example:
 TV reporters stopping certain individuals
on the street in order to ask their
opinions about certain political changes
 The TV reporter has to apply certain
judgment when deciding who to stop on
street to ask questions

44
 Convenience (Accidental)
Sampling
 Is where the researcher selects subjects on the basis of

availability.
 Also called availability/opportunity sampling
 The most convenience population is chosen, which may

be the researchers friends, work colleagues, etc.

 Sample is taken from a group of people easy to contact or

to reach.
 In certain types of descriptive studies, the use of

convenience samples may restrict generalizations to

similar populations. 45
Cont…
 For example, when a Marketing
researcher uses customers who visited a
supermarket on Friday afternoon cannot
generalize to all customers of that Store.
 Instead, he can make generalization for
only those who visited the store on that
day.

46
Quota Sampling
 Is selecting a predetermined number of individuals from

different group (i.e. based on age, gender, etc).

 By setting a quota, and selecting people until the quota

is filled the investigator ensures that the group or

characteristics he/she wants in the sample are there.

 Divide the population into mutually exclusive, sub-

groups
 Select sample items from exclusive sub-groups on the

basis of knowledge and professional judgment.

47
Quota Sampling
 Example:
 Suppose an interviewer is told to interview
250 peoples living in certain geographical
areas.
 Out of which 100 males, 100 females and 50
children are to be interviewed.
 With in this quotas, the interviewer can select
any person on the basis of his personal
judgment. 48
Referral Sampling
 Respondents who are initially contacted are asked

to recommend others.
 Also called snowball sampling/chain

sampling/respondent-driven sampling.
 Often used in hidden populations like drug users

which are difficult for researchers to access.

 Also used when there is no list or other obvious

sources for locating members of the population

(Homeless)
49
THE END!

50