Definition

Sampling is the statistical process of selecting a subset (called a sample”) of a

population of interest for purposes of making observations and statistical inferences
about that population (Otaha, 2015).

Overview on Sampling
One of the Important aspect of research is the stage where the researcher makes
decision on whom to interview or include in the study. Every researcher want to be able
to generalise his/her findings to the wider population I.e to be able to say that his/her
findings are valid or applicable to more people than were actually covered by the study.
The ideal is to study the whole population (if Possible) such a study would give more
confidence in the findings than from findings from study of say 200people (sample).
However, it is often impracticable, tedious, expensive and time consuming to study the
whole population. For example, if a researcher was to study the entire Nigerian
population (with over 200million people spread all over the country) . Such a study will
require so much time, money, personnel and effort. Sometimes it may even be
unnecessary and wasteful to interview everybody when a fraction would suffice.for
instance, it will be unnecessary perhaps murderous, to draw out the entire blood in a
person in order to study its content, just as it will be unnecessary to to drain all the water
in a river for study when a small sample will do.

For this reason the researcher often select a fraction of the people to be studied on the
assumption that they are representative of the whole population and that we can
generalize our findings from the sample for the population. Hence, a sample can be
seen as a smaller representation of a larger whole. In other words, a sample is a
fraction or part of the whole. In simple statistical sense, a 1% sample of the population
means picking one person out of every 100 elements, while a 50% sample means
selecting half of the population for study.
Whatever proportion of the population the sample is, it is desirable that it be
representative of the whole population.

Features of a Good sample

Representativeness: A good sample design should ensure that the sample is

representative of the population. The sample should accurately reflect the
characteristics of the population, such as age, gender, income, and education. For
example, in a study of the adequacy of hostel accommodation at ABU Zaria, the sample
would be unrepresentative if it has only male students as subjects. A representative
sample should include both males and females to reflect the true characteristics of the
Population.

Proportional: A good sample should be proportionately large enough to represent the

population properly to provide statistical stability or reliability by giving the accuracy
required for the purpose of the particular study.

Random Sampling: Random sampling is a crucial aspect of a good sample design.

Random sampling ensures that every member of the population has an equal chance of
being selected for the sample. Random sampling reduces the risk of sampling bias and
ensures that the research findings are generalizable to the population.

Sampling Frame: A good sample design should have a well-designed sampling frame.
The sampling frame is a list of individuals or units in the population.

Appropriate Sampling Technique: A good sample design should use appropriate

sampling techniques that are relevant to the research question and objectives. The
sampling technique should be appropriate for the population, sample size, and research
methodology. Appropriate sampling techniques, such as stratified sampling, cluster
sampling, or simple random sampling, ensure that the research findings are reliable and
accurate.

Choosing a sample size

A good researcher should start from the knowledge of the whole population in order to
reach a decision on a desirable sample size. The researcher should be guided by
various factors such as;
a) Population size. Usually, the larger the population size, the larger the sample
should be
b) Nature of the population. The more homogeneous the population, the smaller the
sampling error and the smaller the sample size could be. On the other hand, the
more heterogeneous the population is, the larger the sample size should be to
reflect all important parts of the population.
c) Purpose of the study. The higher the level of precision that is required the larger the
should be.
d) Types of statistical analysis. Is statistical analysis is to be made from the data, it is
advisable that the sample be large enough to have sufficient cases in all cells of the
table
e) Allowance for non response. The number of respondents who may not be located
or who may refuse to be studied or who may not return the questionnaire as well as
those who may provide illegible, irrelevant or useless responses or those who
return incomplete questionnaire need to be taken into account. For example, if the
desired number of of successful interviews is 500 and non response is estimated to
be 10%, a sample of 550 should be issued with an expectation that if about 50 do
not respond. The number of successful interview will still be close to the desired
sample size.

POPULATION
population refers to the sum total of all sampling units or, according to Bailey (1982) the
sum total of all the units of analysis. Seltiz and Jahoda (1965) also defined the
population as the aggregate of all cases that conform to some designated set of
specifications. The population for a study could be all the citizens of Nigeria or all
University students in Nigeria, or all students of the ABU Zaria, or all residents of a
students hostel. It is the entire universe of interest for which we want to make inferences
or generalizations. Whatever it is, the study population should be defined in terms of:
a) Content
b) Extent
c) Time
For example
a. Content definition, we may say students;
b. Extents definition: ABU Zaria
c. Time definition: 1994/95 session

SAMPLING FRAME
This is a complete list of all units from which the sample is drawn. A good sample frame
is important for a good sample and to be considered good, a sampling frame should
meet the following condition;
I. It should be comprehensive and up to date. In other word, it should cover the whole
population to be studied. If some elements of the population are not included, the
sample is likely to be biased. Ideally, the sampling frame should be obtained from a
 central and reliable source. For instance a study for the attitude of all Nigerians for
instance a telephone directory would not be reliable because Nigerians who do not
use telephone will be excluded. In this case the National census will be more
reliable.whatever the source of sampling frame, the researcher should cross-check
for omissions.
II. Names should not be duplicated on the sampling frame. No person should be listed
twice otherwise, such people will have a higher probability of being included.
In the absence of a comprehensive list, the researcher may also draw up a sampling
frame himself/herself, making sure that the list is as comprehensive as possible. For
instance, in a small community study which is based on households, in the absence of a
centrally collated sampling frame, the researcher may go round to list all houses,
making sure that no street or house is omitted.

ELEMENT
An element is a specific unit or individual that makes up the population being studied.
Elements can be people, objects, events, or any other unit of interest. In quantitative
research, elements are typically represented by numbers or measurements, while in
qualitative research, elements are represented by descriptions or narratives.
For example, in a study on the effectiveness of a new weight loss program, the
elements would be the individuals who participate in the program. In a study on the
impact of social media on academic performance, the elements would be the students
who use social media while studying.
In a study on the prevalence of a particular disease in a population, the elements would
be the individuals who have the disease. In research, elements are typically selected
through a sampling process to form a sample that represents the population. The
sample size and sampling method should be chosen carefully to ensure that the sample
is representative of the population and provides accurate estimates of the population
parameters.

PROBABILITY AND NON- PROBABILITY SAMPLING TECHNIQUES

There are two main methods of sampling namely;

Probability sampling: according to Goode and Hatt (1952) probability, express the
frequency of the nonoccurrence of that event in any series which could produce either
occurrence or nonoccurence.
For example, if an unbiased coin is tossed in the air, it has an equal chance of turning
up a head or tail. The probability of a head is 0.5 similarly while the probability of turning
up tail is also 0.5. similarly, a die has an equal chance of turning up 1,2,3,4,5 or 6. The
probability sampling technique gives the benefit of being able to specify the chance of
each of the sampling units being included in the sample.

Types of probability Sample

1. Simple random sampling: Each member of the population has an equal chance of
being selected for the sample.the idea of randomness implies that the selection is
independent of human judgement. For example, it may happen that if interviewers were
asked to interview people they may interview only people they believe or know will
respond, and avoid those they believe may not respond (eg. People with frown on their
face or those who seem to be in a hurry to get to somewhere). in this case the sample
will be biased because there is not equal probability of inclusion for each sample unit
and it is possible that the omitted person differ significantly from those interviewed.
Therefore, there are two main methods often used for random sample namely;
a. Lottery Method; here each unit of the population is represented by a disc or folded
paper. They are put in a container or bag and shuffled and the required sample is
drawn. Example of this is the raffle draw where every one has an equal chance of
being selected.
b. Use of random Numbers: here members of the population are listed and numbered
in a serial order. Numbers are then selected from the list in a desired but systematic
manner using figures from a table of random numbers. For example, for two figure
populations (up to 99) the researcher may use either the first two digits or the last
two digit in the numbers. Whichever one is chosen should be used consistently for
all the selections. For three figure populations, use the first or the last three digits.
The selection can start at any point on the table of numbers and should proceed
either downwards or sideways but in a consistent manner.

Disadvantages of simple random sampling include

i. It cannot be used without a comprehensive sampling frame
ii. Units selected may be so widely dispersed that it may be difficult to reach them
iii. The sample may not be representative.

2. Stratified Random Sample: with this technique, the population is divided into strata
that the researcher considers important for inclusion in the study or which he/she
believe may generate differences in opinion. Foe example it may be believed that men
and women will differ in their attitude towards family planning or preference for the sex
of a child. A sample is then selected from each stratum using the simple random
technique. This ensures that no important stratum of the population may, by chance be
omitted from the sample.

3. Systematic Sample: with this, selection is done by picking every nth name where n
could be any number. To use this method the researcher needs to first calculate what
fraction of the population the sample is and pick every nth name. For example if the
population is 100 and the desired sample size is 20, this means that the sample is one-
fifth of the population the researcher should number the sampling units serially from 1
to 100 and then pick every 5th name, starting with any number from 1 to 5. if the
researcher decides to start with number 2 for instance, he/she will select number 2 and
every fifth name on the list I.e. 2,7,12,17,22,27 etc.

4. Cluster or Area Sample: a cluster sample may be defined as a simple random sample
in which each sampling unit is a collection or cluster of element. This method is specially
useful for the study of rural communities and also for the study of populations for which
there are no comprehensive sampling frames.
On the other hand Area sample entails the use of maps when there is no reliable
sampling frame and where population can be grouped e.g. into neighbourhood, districts
or cities.

5. Multistage sampling: this is an extension of cluster sampling. With this method,

sampling goes through many stages. After clustering the population, further divide the
clusters into groups and divide the groups into classes. For example, to study Nigerian
population for instance, the researcher may go through the following stage
Stage 1: Divide the country into states
Stage 2: Divide the states into local governments areas
Stage 3: Divide the local government areas into wards
Stage 4: divide wards into streets
Stage 5: divide streets into houses.

6. Multistage sampling: This is a type of sampling design in which required information is

collected from a large sample of units, and additional information is collected from the
sub-samples of the whole sample either at the same time or at a later stage.
Multi-phase sampling should not be confused with multi-stage sampling, although the
two appear to be the same. In multi-stage sampling, different sampling units are
sampled at different stages, while in multiphase sampling, we are concerned with similar
sampling units at each phase.

Non probability sampling: Non probability sampling on the other hand cannot specify
the chances of inclusion of a sampling unit in the sample. This means that the sample is
not chosen randomly, and there is no guarantee that the sample will accurately
represent the population being studied. Instead, the researcher selects participants
based on their convenience, availability, or judgment. Non probability sampling is often
used in exploratory research or when it is difficult or expensive to obtain a random
sample.

Non probability sampling comprises of:

a) Accidental (also called the convenience) sampling
b) Quota sampling
c) Purposive sampling

Accidental sampling: is the one which the researcher simply chooses available people.
A researcher who wishes to sample the opinion of 100 students may simply go to a
faculty and interview any 100 students he/she/ sees and who arfe willing to be
interviewed. Persons who are not available or are unwilling to be interviewed are simply
ignored, and the researcher moves on to the next person.
The accidental sample is perhaps most liable to error and the degree of
representativeness is usually low.

Quota sampling: as the name implies, certain proportions of the sample are reserved for
certain groups or categories of respondents. The researcher first of all decides on how
many of the elements in different classes or categories will be included in the sample.
Example, How many males, and how many females, if sex is used as a quota factor.
Quota sampling is similar to stratified random sampling in that in the two methods,
specified proportions of sample are reserved for certain groups or categories however
the two methods differ in that under the stratified random sampling, respondents are
selected on random basis while under quota sampling, respondents are selected on non
random basis.
Purposive Sampling: with this technique the sample is selected on the basis of the
researchers knowledge of the population. In other words, the researcher chooses only
those who meet the purposes of the study as may be determined by him/her.

Overview
In spite of the limitations, the non probability sampling techniques are also widely used.
Their main advantages include convenience and economy. These two advantages may
sometimes outweigh the risk involved in not using the probability sampling.
The non probability sampling may also be used when the researcher merely wants to
find out about something, but is not so much interested in generating his/her findings to
the whole population, or wen the study is intended to be a pilot study for a larger study.
Probability sampling, involves selecting a sample in such a way that every element in
the population has a known chance of being selected. It ensures that each element has
an equal opportunity of being selected and that the sample is representative of the
population. Common types of probability sampling include simple random sampling,
stratified sampling, cluster sampling, and systematic sampling. On the other hand, Non
probability sampling, also known as judgment sampling or convenience sampling,
involves selecting a sample based on the researcher's judgment or convenience rather
than using a random process. This method is often used when it is difficult or expensive
to obtain a complete list of the population or when time is a limiting factor. Common
types of non probability sampling include Accidentalsampling, quota sampling, and
purposive sampling. The main difference between probability and non probability
sampling is the level of confidence that can be placed in the results. Probability
sampling provides a higher level of confidence because it ensures that every element
has an equal chance of being selected and that the sample is representative of the
population. This allows researchers to make statistical inferences about the population
based on the sample data. Non probability sampling, on the other hand, provides a
lower level of confidence because it does not ensure representativeness and may result
in selection bias.

Reference:
Babbie E. (1979) the practice of social reaesch (Belmont CA: Wardsworth Publishing
Co.
Bailey K. (1982) Methods of Social Research (New York Free Press)
Babatunde Ahonsi and Omololu Soyombo (1996) - Readings in Research Methods and
Applications.
Ruth Macklin (2014)Research Ethics: Protecting Participants and Promoting Integrity
Chinyere Okoro (2017) - Research Methods for Development: A Nigerian Perspective
Olufunmilayo Olopade (2018)
Tshuma (2015) -African Rese