Part 4 Sampling Design

• Quantitative research intends to generalize the findings

to a population under study;

Uses survey design and experimental design

• The research data under this approach need to be
representative of the population in question

• Qualitative research intends to explore and gain depth

understanding about the phenomenon under
investigation;
• Generalization of study findings is not the concern of this
approach;
Part 4 Sampling Design

•Components of survey design (quantitative research

approach)
• Sampling design ,
• Instrumentation (instrument design), and
• Data collection (conduct of the survey)
•Sampling also deals with experimental design;

•Sample and instrument design and mode of collecting

data are crucial to generate representative statistical
data about a population under study.
Sampling Design
• Sampling design:
• Population and sampling frame
• Sample size
• Sample selection techniques
• A population (universe) is any complete group—for example,
of people, companies, stores, or college students—that
shares some common set of characteristics.

• Population element refers to an individual member of the

population.

• When measuring every item in a population is impossible,

inconvenient, or too expensive, we take a sample and
undertake a survey.
• Typically, business researchers do not conduct a census
of the population, but rather select a smaller number of
population elements, a sample.

• A sample is a subset, or some part, of a larger population.

• Sampling is the process of selecting a number of study

units from a defined population.

• The purpose of sampling is to estimate an unknown

characteristic of a population .
• Researchers should try to draw a sample that is
representative of the population.

• A representative sample has all the important

characteristics of the population from which it is drawn.

• In sample design, how well a sample represents a

population depends on the sample frame, the sample
size and the specific procedures of selecting potential
respondents.

• Statistically, a sample can only be representative of the

population included in the sample frame.
• 1. Sampling frame
• Sampling frame is the list of study units that have a
chance to be selected, given the sampling approach
adopted

• Is a list of study units (along with their key characteristics)

from which the sample is drawn

• each study population consists of study units (may

consist of persons, institutions/ companies , etc.)

• sampling requires a listing of all study units that exist or

can be compiled – sampling frame;
• Example: if the study is about manufacturing companies
in Ethiopia all manufacturing Cos. make up the population

• For sampling purpose, you may not get data on the list of
each and every manufacturing Co in Ethiopia and you may
manage to compile the sampling frame based on records
maintained by the Ethiopian Chamber of Commerce
(Sampling frame),

• The research plan (proposal) needs to clearly define the

study population `– sampling frame

• State population size, if size can be determined, and the

means of identifying study units in the population.
• In a study with the intent of statistically generalising about
a population, the sample frame from which sample to be
drawn ought to be representative of the population (i.e.,
the sample frame should be comprehensive enough).

• In designing a sample, basing the sample selection on a

comprehensive list of potential respondents who have a
chance of selection is vital to increasing the
representativeness of the sample.
• 2. Sample size
• refers to the number of items (study units) to be selected
from the universe (or the sampling frame) to constitute a
sample.

• The research plan (proposal) has to provide information

and justification about sample size;

• The choice of a sample size has a bearing on the

reliability of a study.

• However, this does not mean that large sample size

always leads to high level of accuracy in the estimates;

• It is not necessarily true that the bigger the sample, the

better the study;
• 2. Sample size (Cont’d)
• Sarantakos (2005, p. 171) noted that large samples do not
always guarantee a high degree of precision and success
in general.

• Beyond a certain point (level), an increase in sample size

will not improve the study much;

• Fowler (1984, p. 43) indicated that for a population, the

level of precision increases steadily up to sample sizes of
150 to 200. After that point there is much more modest
gain to increasing sample size.

• Increasing sample size may do the opposite; if the quality

of the measurement or data collection is adversely
affected by the large size of the study;
• 2. Sample size (Cont’d)

• After a certain sample size, in general, it is much better to

increase the accuracy and richness of data collection (for
example by improving the training of interviewers, by pre-
testing of the data collection tools) than to increase
sample size.

• Also, it is better to make extra effort to get a

representative sample rather than to get a very large
sample.
• 2. Sample size (Cont’d)
• Sample size determination
• No a single right way for the determination of sample size,
there are three approaches that are inadequate (Fowler,
1984);

• 1. Taking a certain percentage of the population

• Like taking 5% or 10% of the population as a sample size;

• According to Fowler (1984), specifying the fraction of the

population to be included in the sample is never the right
way to decide on the sample size.

• This is because the population from which a sample of a

particular size is drawn has no impact on how well that
sample is likely to describe the population (Fowler 1984).
• Sample size determination (Cont’d)
• 2.Considering the usual/ typical sample size in past
studies

• Consider the usual sample size in past studies and adopt

a similar sample size (Managerial judgment)

• Saying that a particular sample is the usual or typical

approach to studying a population is also always the
wrong answer;

• Although it is good to see the sample size used by other

competent researchers, the sample size should be
decided on a case by case basis, considering the different
goals to be achieved by a particular study (Fowler 1984).
• Sample size determination (Cont’d)
• 3. Using statistical formula.
• The statistical determination of sample size requires
knowledge of
• (1) the variance (the heterogeneity) of the population,
• (2) the confidence level and
• (3) the magnitude of acceptable error.

• Different computational formulas are available for

determining sample size.
• Sample size determination (Cont’d)
• Variability (or Heterogeneity)
• The determination of sample size depends on the
variability (or heterogeneity) of the population;
• In statistical terms, this refers to the standard deviation of
the population.

• If the population is homogeneous, only a small sample is

required;

• For example, predicting the average age of college

students requires a smaller sample than predicting the
average age of people who visit Merkato on a given
business day.

• As heterogeneity increases, so must sample size.

• Sample size determination (Cont’d)
• For statistical sample size determination, population
standard deviation needs to be estimated;
• Ideally, similar studies conducted in the past will give a
basis for judging the standard deviation.

• In practice, researchers who lack prior information may

conduct a pilot study to estimate the population parameters
so that another, larger sample of the appropriate sample
size may be drawn.

• A rule of thumb for estimating the value of the standard

deviation is to expect it to be about one-sixth of the range.

• If a researcher conducting a study on average age of

college students expects the age to range from 20-26, a rule
of thumb estimate of standard deviation would be 1 year.
• Sample size determination (Cont’d)
• The confidence level
• This value is used to measure the degree of certainty
about how well a sample actually represents the entire
population within the margin of error chosen for the study.
• The confidence interval estimate is a range of numbers
within which the researcher is confident that the
population value will lie.
• The confidence level is a percentage or decimal that
indicates the chance that the results will be correct.
• In most business research, we typically use the 95 percent
confidence level (Zikmund et al 2012).
• When a larger confidence level is chosen, it shows a
greater degree of accuracy provided that the sample size
is larger.
• Sample size determination (Cont’d)
• The magnitude of acceptable error,
• The magnitude of error, in statistical terms E, indicates
how precise the estimates must be.
• The maximum allowable difference between the sample
estimate and the true population value.
• Sampling errors are the random variations in the sample
estimates around the true population parameters.
• Sampling error decreases with the increase in the size of
the sample, and the sampling error happens to be of a
smaller magnitude in case of homogeneous population.
• Sample size determination (Cont’d)
• Make a judgment about the allowable magnitude of error
and confidence level;
• The judgment about the allowable error and the
confidence level may be the manager’s decision to make.
n=(( Zσ)/ E) 2

• Where n = estimated sample size

• Z = standardized value that corresponds to the
confidence level
• σ= sample standard deviation or estimate of the
population standard deviation
• E = acceptable magnitude of error,
• Sample size determination (Cont’d)
• Example: Suppose a survey researcher studying annual
expenditures on cosmetics wishes to have a 95 percent
confidence level (Z = 1.96) and a range of error (E) of less
than $2. If the estimate of the standard deviation is $29,
the sample size can be calculated as follows

• n= ((1.96*29)/2)2
• Estimated sample size would be = 808

• If the margin of error is doubled to 4 the

estimated sample size would be less 202
• Sample size determination (Cont’d)
• To help researcher calculate sample size using statistical
methods, a number of easy-to-use tables have been
compiled (check books in statistics where sampling is
discussed);

• Fowler (1984) argued that in some theoretical sense

there is nothing wrong with this third approach; however,
in practice this approach has little help to most
researchers-
• Sample size determination (Cont’d)
• The discussions so far show the need for considering
different factors in deciding on the desired sample size.

• These factors include the availability of time and

resources, homogeneity of the target population, the
accuracy required and the aim of the research
(Sarantakos 2005, p. 171).
Procedures of sample selection
• Two types of sample selection: probability sampling
and non-probability sampling;
• Probability sampling is also known as ‘random
sampling’ or ‘chance sampling’.
• Under probability sampling every item of the universe/
population has an equal (or known) chance of inclusion
in the sample.
• Random sampling ensures the law of Statistical
Regularity which states that if on average the sample
chosen is a random one, the sample will have the same
composition and characteristics as the population.
• This is the reason why random sampling is considered
as the technique of selecting a representative sample.
Probability sampling (cont’d)
• Random sampling ensures that the sample is
representative of the population, reducing bias and
allowing for generalization of the results to the
population

• The random selection can be done with tools like:

• Lottery methods
• Software like SPSS
Probability sampling (cont’d)
•Simple Random Sampling
• Each individual has an equal chance of being selected.
• Example: Randomly picking 100 names from a list of 1,
000 people.
•Systematic Sampling
• Select every Kth individual from a list after randomly
choosing starting point.
• Example: Surveying every 10th person on an
alphabetically ordered list.
•Stratified Sampling
• The population is divided into subgroups (strata) based
on a specific characteristic (e.g., age, gender); (company
size, business activity type etc.)
Probability sampling (cont’d)
• A random sample is taken from each stratum
proportionally

• Ensuring equal representation of men and women in a

study; ensuring representation of small, medium and
large businesses in the sample

•Cluster Sampling
• The population is divided into clusters (e.g., neighborhoods,
schools).
• Entire clusters are randomly selected, and all or some members
within chosen clusters are surveyed.
• Example: Surveying students in randomly selected classrooms
instead of all students in a school.
•Multistage Sampling
• Combines multiple sampling methods (e.g., selecting clusters
first, then applying simple random sampling within those clusters)
.
Probability sampling (cont’d)
• Advantages of Probability Sampling
• Representative samples and reduce bias : Reflects the
characteristics of the population.
• Supports Statistical Analysis: Enables valid generalizations and
hypothesis testing.

• Issues under probability sampling

• Resource intensive: requires a complete and accurate sampling
frame;

• Complexity – some methods like stratified or multistage

sampling can be challenging to implement;

• Non-response – even with random sampling non-response can

affect representativeness
Non-probability sampling
• Non-probability sampling methods involve selecting
samples using factors other than randomization;

• It does not give all study units some chance of selection;

there may not be even a complete list of the population
under investigation;

• Non-probability sampling is typically used when it is

impractical or unnecessary to obtain a random sample;
Non-Probability sampling
• Convenience Sampling:
• Selecting samples that are easier to access or readily
available.

• Samples are selected based on their availability, proximity,

or ease of access to the researcher, rather than through a
randomized process.

• Participants are selected simply because they are easily

accessible or readily available, regardless of their
relevance to the research topic;

• The goal is to save time, effort, and resources by

collecting data from accessible participants, regardless of
their relevance.
Non-Probability sampling

•Strengths
• Time, and cost saving- Saves time as participants are
readily available; minimal resources are needed to
locate and recruit participants.
• Ease of Implementation: Requires less effort and
expertise compared to probability sampling.
• Useful for Exploratory Research: pilot studies or for
generating initial insights.
Non-Probability sampling
• Limitations
• High Risk of Bias: samples may not represent the
population accurately;

• Lack of generalizability: findings cannot be reliably

applied to the broader population;

• Overrepresentation of certain groups: only those who

are easily accessible or willing to participate are
included;

• Sampling error: the absence of random selection

increases the likelihood of errors
Non-Probability sampling
• Judgmental or purposive sampling
• Selects participants based on their knowledge, experience,
or specific characteristics relevant to the study;

• Participants are selected intentionally based on specific

characteristics or criteria that align with the study’s
objectives;

• The goal is to ensure the sample is relevant and aligned

with the research questions or objectives;
Non-Probability sampling
• The researcher decides which individuals or groups are
most appropriate for the study (based on researchers’
judgement);

• Ensures that the sample includes participants who can

provide insights into the research topic.

• For example, for a study on capital markets in Ethiopia,

experts working in this field (like capital markets authority,
Ethiopia stock exchange etc.) may be selected
intentionally;
Non-Probability sampling
• Purposive sampling is used in qualitative research: to gain
in-depth insights rather than generalizable results.

• It may be used when the population of interest is small in

size and hard to access;

• Strength: information relevant for the research objective

and questions can be obtained;

• Limitation: results may be biased to certain groups and

hence difficult to generalize to the population of interest
Non-Probability sampling
• Snowball sampling:
• Existing participants recruit other participants from their
networks;

• It is often used to study hidden, hard-to-reach, or

specialized populations where a sampling frame is not
readily available.

• Helps in studying marginalized groups, those living with

some disease, criminals etc.

• Example: asking current study participants to refer others

they know;
Non-Probability sampling
• Strengths: useful for hard-to-reach populations;

• Limitations: Can lead to a homogeneous sample; relies

heavily on participant networks;
• When to use snowball sampling:
• When the population is difficult to access due to stigma,
legality, or safety concerns;

• When there is no sampling frame or existing list of the

population;

• When participants are more likely to respond due to peer

referral;
Sampling design (qualitative research)
• Qualitative research is not about generalization from
sample to population; it rather tries to have depth
knowledge about the phenomenon under investigation.
• The sample design considers the intention of the research
approach (Qualitative);

• Sample size under qualitative research is estimated when

the research proposal is developed; the actual sample size
may be different from what is initially specified in the
research proposal;

• The researcher may start the study with some estimated

sample size and if after a certain interview, the
information becomes redundant, the researcher stops
conducting the interviews there.
Sampling design (qualitative research)
• For example, the estimated sample size for indepth
interviews may be 10; when interviews are conducted if
the information obtained with the 7th interview is
redundant, the researcher stops the interviews, and the
sample size would be 7.

• Under qualitative research approach, since statistical

generalization is not the intent, data representativeness,
randomization etc., in sample selection are not the focus;

• Sample selection for qualitative research is hence based

on non-probability sampling such as purposive sampling,
snowballing, convenience etc.
 Steps in developing sampling design
• While developing a sampling design, researchers need to
pay attention to the following points:
• Type of universe: The first step in developing any sample
design is to define clearly the set of study subjects,
technically called the Universe/ Population, to be studied.
• The universe/ population can be finite or infinite.
• In finite universe the number of items to be studied is
certain(can be determined/ estimated);
• Infinite universe the number of items is unlimited, i.e., we
may not have idea about the total number of items in the
population.
• The population of a city, the number of workers in a factory
etc. are examples of finite universes,
• Number of stars in the sky, listeners of a specific radio
programme etc. are examples of infinite population
Steps in sampling design
• Sampling unit: decide on the sampling unit before
selecting the sample;
• Sampling units may be:
• a geographical one such as subcities, woredas, villages
etc; or
• a construction unit such as house, flat, etc., or
• it may be a social unit such as family, club, school, etc.,
or
• it may be an individual.

• The researcher needs to decide which sampling unit to us;

one or more of the above sampling units may be used in a
study;
Steps in sampling design (Cont’d)
• Source list: it is also known as ‘sampling frame’ from
which sample is to be drawn;

• It contains the names of all items of a universe (in case of

finite universe);

• If source list is not available, researcher has to prepare it;

such a list should be comprehensive;

• It is important for the source list (sampling frame) to be as

representative of the population as possible.
Steps in sampling design
• Size of sample: This refers to the number of items to be
selected from the universe to constitute a sample.

• Sampling procedure: the researcher must decide on the

technique to be used in selecting the items for the sample.
Sampling procedure selection criteria
• Selecting an appropriate sampling procedure is critical
to ensuring that the study’s results are valid, reliable,
and applicable to the research objectives.

• There are some criteria that researchers may use in

selecting the appropriate sampling procedure
 Sampling procedure selection criteria
(cont’d)
• 1.Nature of the Research
• Exploratory research: non-probability methods (e.g.,
convenience or purposive sampling) are often suitable to
explore (initial) insights.
• Descriptive, Causal Research: Probability methods (e.g.,
simple random or stratified sampling) are preferred for
generalization and hypothesis testing.
• 2.Population Characteristics
• Population Size:
• Large, diverse populations may require stratified or cluster
sampling for representativeness.
• Small or specialized populations may rely on purposive or
snowball sampling.
 Sampling procedure selection criteria
(cont’d)

• Accessibility: The feasibility of reaching the population (e.g.,

remote areas, hidden populations) influences the sampling
method (purposive sampling may be used);

• Homogeneity vs. Heterogeneity:

• Homogeneous populations might require simple random
or convenience sampling.
• Heterogeneous populations might need quota or
stratified sampling to account for variability.
Sampling procedure selection criteria
(cont’d)
• 3.Sampling Objectives
• Representativeness (generalizability):
If the goal is to generalize findings to the entire
population, use probability sampling.
• Specificity (specific group- depth understanding):
If targeting a subgroup or specialized population,
purposive or quota sampling may be more appropriate.

• 4.Sample size requirements

• Larger sample sizes often demand probability sampling to
reduce biases.
• Smaller studies or pilot projects may employ non-probability
methods due to lower cost and ease of execution.
Sampling procedure selection criteria
(cont’d)
• 5.Resource Availability
• Budget: Probability sampling methods, such as stratified
or systematic sampling, can be more expensive than non-
probability methods.
• Time Constraints: Tight deadlines may necessitate
methods like convenience or voluntary response sampling.
• Manpower and Expertise: Complex sampling techniques
may require experienced personnel and advanced tools.
• 6.Availability of Sampling Frame
• A sampling frame (list of the population) is essential for
methods like random sampling or systematic sampling.
• If no sampling frame is available, methods like snowball or
convenience sampling might be necessary.
Sampling procedure selection criteria
(cont’d)
• Statistical Rigor
• Probability sampling ensures that every individual in the
population has a known and non-zero chance of selection,
which supports statistical analysis and hypothesis testing.
• Non-probability sampling lacks statistical rigor but is
useful for qualitative insights or initial exploration.
• Ethical Considerations
• The sampling procedure must respect participant rights,
including informed consent and confidentiality.
• Vulnerable populations may require specific sampling
considerations (e.g., purposive sampling with ethical
oversight).
Sampling procedure selection criteria
(cont’d)
• Degree of Accuracy Needed
• Studies requiring high accuracy or generalizability often
rely on probability sampling.
• Exploratory or hypothesis-generating studies can tolerate
less precise sampling methods.