In November 1988, George Bush was elected president
of the US with 54% of the popular vote as against 46%
for Michael Dukahis.
Prior to the election, a number of political polls had
predicted the Bush victory.
CBS/New York Times 57 43
ABC/Washintong Post 56 44
Gallop Poll 56 44
Gordon Black/USA Today/CNN 55 45
KRC/APN 55 45
Harris Poll 53 47
NBC/Wall Street Journal 53 47
Although the poll estimates were varied, you
can see how they were clustered around the
actual election-day results (54%, 46%).
Ques. Now how many interviews do you suppose
it took each of these pollsters to come with a
few % points in estimating the behaviour of
about a 100 million voters?
Ans. Fewer than 2000
The level of accuracy in due to the fact that, the
sample population provided useful
descriptions of the total population and did
contain essentially, the same variations that
exist in the population.......This is what
sampling is all about.
Sampling is the process of selecting
observations. In order words, it is selecting a
part to represent a whole.
• Sampling allows researchers to draw precise inferences
on all the units (a set) base on relatively small number
of units (a subset) when the subsets accurately
represent the relevant attributes of the whole set.
• Is associated with coverage and scale of a survey.
i. Saves money – although unit cost of total coverage will
ii. Saves labour
iii. Saves time.
Element – an element is that unit about which
information is collected and that provides the basis
of analysis. It can be people or certain types of
people, families, social clubs or company.
NB. Elements and units of analysis are often the same
in a given study, though the former refers to
sample selection and the latter refers to data
Population – It is aggregation of study elements. In
other words it is the entire set of relevant units or
cases or individuals that fit a certain specification
e.g. hhs,, hs, teachers, students etc.
Study population – Is that aggregation of elements
from which the sample is actually selected.
Sampling Unit – Is that element or set of elements
considered for selection in some stage of sampling.
In a simple-stage sample, the sampling unit are the
same as the elements. In a more complex samples,
different levels of sampling units may be
employed e.g Census blocks in a city – selection of
sample households in the selected blocks –
selection of sample of adults from the selected
Sample Frame – Is the actual list of sampling
units from which the sample is selected. In a
single-stage sampling design, the sample frame
is simply a list of the study population e.g.
Student roster could be the sample frame when
students are selected from it. If available, its
reliability has to be assessed – updating.
Limitations of Sample Frame:
• Incomplete frames - when some units are missing
• Cluster elements - when samples are listed in
clusters rather than individuals. Individual
clusters must be listed.
• Blank foreign elements – when some of the units
are not included in the research population e.g.
Non-students in a youth list – new settlers, Error in
sampling Frame – 1936 Us Election.
Observation Unit – An observation unit or unit of
data collection is an element or aggregation of
elements from which information is collected.
The unit of analysis and unit of observation are
often the same, but they need not be the same
always. For example, a researcher may interview
heads of households (observation units) to collect
information about all members of the households
(the unit of analysis).
Sample size - the selected number of the
population to be surveyed. How is this
Sample fraction - Sample size expressed as a
ratio of the frame.
Sample randomness - distribution – spread of
There are two types of sampling methods –
Probability and Non-probability sampling methods.
The ultimate purpose of sampling is to select a set of
elements from a population in such a way that
descriptions of these elements accurately portray the
parameters of the total population from which the
elements are selected.
Probability sampling enhances the likelihood of
accomplishing this aim and also provides methods for
estimating the degree of probable success.
• Random selection is the key to this process.
• In a random selection, each element has an
equal chance of selection independent of any
other event in the selection process.
• Mean of the sample population will be close to
the mean of the total population.
• The bigger the sample, the closer the sample
mean to the total population mean.
Types of Probability Sampling
1. Simple random – it is the basic probability design and is
incorporated in almost all elaborate sample design. It is a process
by which each member of the sample population has an equal
chance or known non-zero chance of being selected.
Identification number must be given to every unit on member of
a) Lottery Approach
b) Use of random tables
Is applicable when a sample frame can be secured or compiled.
2. Systematic sampling – every Kth
element in the total list is chosen
systematically for inclusion in the sample. The first element is
selected by simple random process.
3. Stratified sampling – This method is not an alternative to the
simple random and systematic sampling methods but it represents
a possible modification in their use. Stratified sampling ensures a
greater degree of representativeness – thereby reducing the
probable sampling error.
- disaggregating of the population into coherent sub-groups so
that the sample becomes more representative.
- ensures that different groups of population are adequately
represented in the sample.
- it combines homogeneity and heterogeneity at different levels
4. Cluster sampling - It is often used in large surveys
because it is not very expensive. It involves selecting
larger grouping called clusters from which the
sampling units are chosen.
Different categories or levels
• Major – states, regions, districts, cities
• Medium – village, town, NB. Etc
• Compound housed, classes
- Used in situations where no reliable sample frame
- Can be used to estimate population when census
data are not available
5. Multi-stage Sampling - Sub – Sampling in
stages – combination of sampling techniques
E.g. choosing about 1000 farmers for a study
Region, Districts then to individuals.
Even though no probability sample will be perfectly
representative in all respect, controlled selection
methods permit the researcher to estimate the degree
of expected error in that regard.
In spite of the above comment, it is sometimes not
possible to use standard probability sample methods or
sometimes, it may be even appropriate to use non-
probability sampling methods.
1. Purposive or Judgemental Sampling
• The use of purposive sampling is heavily dependent on the
subjective decision of the researcher.
• Samples are selected because they satisfy certain criteria of
interest. choice of informants/people with necessary information
or knowledge or experience e.g.
- Mining & AIDS
- Ethnic conflict
• This method is very cheap.
• In spite of the subjectiveness of this approach in the selection of
units, social scientist have used it with very good success.
2. Accidental Sampling
• This method is also known as “Convenience
• It includes selecting anyone who is handy – i.e.
anyone the interviewer meets on the street.
- Interviewers are influenced by a lot of factors in
choosing the sample.
- Researcher must be aware of biases
3. Quota sampling
• This method is an improvement over Accidental
sampling. The main objective is to select a sample that
is similar as much as possible to the sampling
• Certain parameters and characteristics are defined for
interviewers - Sex, age, ethnicity, education, etc.
- Interviewers are however, not told precisely who to
- Although certain categories are defined, they are not
Problems associated with Quota Sampling
• Lack of information on which to base the quota
• Small number of variables can be used in quota - not all
characteristics are visible e.g. social classes in Africa.
• Unavailability of certain segments of the population -
e.g. sick and confined.
• Difficult to supervise
- More useful in a situation where a small sample is
taken from a large population.
- Be careful to generalise - check the representativeness.