RESEARCH METHODS

Lesson 5

RESEARCH DESIGN AND

SAMPLING
 Session Objectives
• Participants should understand different Research
Strategies/Approaches/Types.

• Should be able to identify and apply different

research designs.

• Be able to select and apply appropriate sampling

techniques.

2
 Research Methodology
 Research methodology focuses on the manner in which
the research is planned, structured and executed in order
to comply with scientific criteria (Mouton and
Marais,1996).
 Leedy (1993), argues that research methodology forms an
integral part of any research that is undertaken.
 It assists in explaining the nature of the data, and
 highlights the methods employed that will lead to the
generation of appropriate conclusions through applicable
data processing.
 The methodology chapter is an important component that
maps out the methods to be utilized during the research.
 This chapter addresses the question; “How did you
conduct the research?’’
 It helps your audience understand the various steps of
your research process.
3
 Research Methodology
This chapter should explain the following:
 Your methodology; why you chose it and why it is
appropriate for your research.
 Your methods of data collection.
 How you expect to show that your data is reliable.
 Your methods of data analysis.
When writing a methodology chapter, you will need to
include:
1. An introductory paragraph that describes the problem
that you addressed through your methodology.
2. Include a general overview of the approach you used.
3. Provide a thorough description of how you collected the
necessary data as well as the analytical procedure used
to draw conclusions.
4. Include a rationale for the choice of methods for data
4
gathering and for data analysis.
 Deciding on Research Methodology
A proper research methodology should answer the questions:
1. What data will be collected- quantitative, qualitative or
both?
2. How will the data be collected-Interviews,
questionnaires, observation, FGD etc.
3. Is the research method both suitable and informative
enough?
4. What are the advantages as well as limitations of a
particular method selected?
5. Should one method be applied or perhaps using a variety of
them will add to the novelty of my thesis?
6. Are the research methods you have chosen effective and
fitting in the spectrum of possible approaches?

NB: Your methodology chapter should have all the necessary

parts like participants, instruments, procedure and data
5
analysis.
 Research Strategy

 It outlines the logistics of your research.

 A research strategy is a general orientation of the

conduct of a research which can either be
quantitative, qualitative, pragmatic or
advocacy/participatory.

 The researcher should be very clear about the

research strategy because it informs/influences the
research design.

6
 Fundamental differences between research
strategies
Type Process/Main Principle Epistemolo Ontological
emphasis gical orientation
orientation
Quantitative Measurement Deductive Post- Objectivism
(P-Value) reasoning; testing Positivism
of theory
Qualitative Words Inductive; Interpretivis Constructio
generation of m nism
theory
Pragmatic/ Measurement and Inductive and Interpretive Constructiv
Mixed Words deductive and e and
(Triangulation) positivism objective
Advocacy/ Involve Bring positive Interactivis Personalise
participatory marginalized or change m d reporting
vulnerable groups (emancipate) 7
 Research Design
 Guy et al.(1987:92) define research design as “a plan of
procedures for data collection and analysis that are
undertaken to evaluate a particular theoretical
perspective.”

 Bryman and Bell (2003) define a research design as a

structure that guides the execution of a research
method and the analysis of the subsequent data.

 Kothari (2007:31) defines a research design as “a

conceptual structure for the collection, measurement
and analysis of data which combines relevance and
economy”.

 Thus a research design is a plan that shows how all the

major parts of the research project come together.

8
 Research Design
 Generally, there are two parts of the research design;
1. Sampling design which deals with the methods of
selecting subunits to be observed for the given study.
2. Statistical design which consists of the appropriate
methods of statistically analyzing data.

 The choice of each reflects decisions about the priority

given to a range of dimensions of the research process
which include the importance attached to:

 expressing causal connections between variables;

 generalizing to the population.
 understand behaviour and meaning of that behaviour in
it’s specific context.
 Having a temporal appreciation of phenomena and their
interconnections.
9
 Types of Research Designs
 They include, exploratory, descriptive, explanatory, case
study, comparative and hypothesis-testing.
 The researcher should decide in advance of the collection
and analysis of data as to which design would be most
appropriate.
Exploratory design
 Also known as formulative because the main aim is
formulating a problem for more precise investigation or of
developing a working hypothesis from an operation point
of view.
 The major emphasis is on discovery of ideas and
insights.
 Uncovers basic facts about the topic and allows
consideration of different aspects.
 Usually focused on the “what,” not the “why”.
 Suggests directions and feasibility of future research
through descriptive research. 10
 Types of Research Designs
Descriptive design
 Concerned with describing the characteristics of a
particular individual or of a group.
 Gathers data that describe events or situations and
then organizes, tabulates, depicts, and describes the
data.
 Uses description as a tool to organize data into patterns
that emerge during analysis.
 Answers the “what?” questions i.e. “What is the best
way to provide access to computer equipment in
schools?”
 Often provide leads in identifying needed changes –
good for action research e.g. responding to a crisis.
11
 Good for testing theory but will not find causes.
 Types of Research Designs
Explanatory/Causal Design
• Focuses on “why questions”, or the reason a situation
or behavior occurs.
• Builds on exploratory and descriptive research, and
other explanatory research.
• The emphasis is to explain why variables change – find
causes.
• Uses theory.

• Much of the research published in journals is

explanatory.
12
 Types of Research Designs
Case study design
• This is a detailed and intensive analysis of a single case.
• A case is an object of interest to the researcher and can be a
single organization, location, person or a single event.
• A case study cannot be generalized to the population.
• The case study research design is also useful for testing whether
scientific theories and models actually work in the real world.
• The actual research should be thorough, meticulous and
systematic.
• Analyzing results for a case study tends to be more opinion
based than statistical methods.
• It is best for situational analysis to address local and imminent
problems. 13
 Types of Research Designs
Comparative design

 A study that compares two or more groups on one/set of

variables.
 This design entails using identical methods of contrasting
identical cases.
 The contrasts can either be similarity or difference based.
 It is a way of understanding social phenomena better by
comparison.
 Internal validity is weak but external validity is strong when
random sampling is used.
 Cost is a limiting factor but provides a deeper
understanding of the phenomena of interest to the
researcher.
14
 Types of Research Designs
Hypothesis-testing
 Generally known as experimental studies.
 The researcher tests the hypotheses of causal
relationships between variables.
 Enables drawing of inferences about causality because
uses procedures which reduce bias and increase
reliability.
 It is probably the strongest design with respect to internal
validity.
 Enables a deductive approach to the relationships
between theory and data.
The process of deduction; theory hypotheses data collection
15
hypotheses testing revision of theory
 Time Dimensions
Some research neglects the element of time, other research
focuses heavily on time.
Cross-Sectional Research
• Observations at a limited point in time.
• Frequently used by social scientists.
• Although it does not capture time, change, or process, it is
often used in explanatory research.
Longitudinal Research
• Much more expensive and time-consuming than cross-
sectional.
1. Time-series: multiple observations over time of the several
units.
2. Panel Study: multiple observations over time of the same
units.
3. Cohort Study: multiple observations over time of similar units.
16
 Terms used in Sampling
 Population – a set of units having common observable
characteristics.
Also known as target population.
 Accessible population – this is the portion of the target
population that is accessible to the researcher for a particular
study.
 It represents the sampling frame from which the sample is then
selected.
 An element is an individual unit of a population.
 Sample - a finite subset of the population that is selected for
research.
 The goal is to be able to find out true facts about the sample
that will also be true of the population.
 Sampling is the process by which we select our sub-set of the
population.
 Sampling frame - a list of all elements from which the sample is
actually drawn.
 Sampling unit – the individual members of the sample or cases.
 A parameter is a numerical characteristic of the population while
that of a sample is a statistic.
17
 A unit of analysis is the type of element or variable of interest.
 Why do we sample?
 Economy of expenditure - Data can easily be collected - e.g. to
establish brand preference for a soap, it costs too much money and
consumes too much time to ask every person in the country.
 Greater speed - Data can be collected and summarized more
readily especially when information is required urgently.
 Greater scope - When complete destruction of a product is
involved, a cent per cent test of production is out of the question.
– e.g. to ascertain if glass containers can withstand pressure, we
put a sample of containers under pressure until they break.
Therefore we do not break every item produced.
 Greater accuracy - Sample results are more accurate than census
results.
 Quality data come from samples.
 Smaller numbers allow quality field study, checks and tests are
affordable at all stages, editing and analysis of results can be
carefully done.
 Practicability – It is practical to interview a sample than the whole
population.
 You don’t really have to study all birds in order to know how they
live.
18
Stages in the Selection of a Sample
Define the target population

Select a sampling frame

Determine if a probability or non-probability

sampling method will be used

Plan procedure
for selecting sampling units

Determine sample size

Select actual sampling units

Conduct fieldwork
19
 Determining a Proportional Sample Size
If N>10 000, then n=
z 2 pq
ME 2
If N<10 000, then nf = n
n
1
N
OR n = z 2 pq  ME 2
2
2 z pq
ME 
N = population size N
n = desired sample size
z = standard normal variable at the required confidence level (z
statistic).
p = estimated characteristic of target population
q =1–p
ME = level of statistical significance set/margin of error
If the proportion of a target population with a certain
characteristic is 0.50, the z-statistic is 1.96, and we desire
accuracy at the 5% level of significance, population is 4000
20
then the sample size can be calculated.
 How to Determine Sample Size

Rule of Thumb
 If the population is less than or equal to 1000 take 30% of
the population.
 If more than 1000 but less than or equal to 10 000 take
23% of the population.
 If more than 10 000 but less than or equal to 100 000 take
10% of the population.
 If you have a population of more than 100 000 take 1% of
the population.

21
 Assessing the Adequacy of a Sample
• There are three key aspects of sample quality namely
sample size, representativeness and implementation.

1. Size--The size should be adequate for the purpose of the

study, especially if there were many sub-groups included
in the analysis, or many variables used simultaneously.

2. Representativeness--The sample should be selected

randomly from the population, using probability theory.
The sampling frame should be adequate.

3. Implementation--The sampling plan must be carried

out carefully, and adequately supervised. Was there
some quality control plan? The result should generate a
good response rate.

22
 Sampling Methods
 Are clearly distinguished between probability and non-
probability.

Probability Sampling
 Each member of the population has an equal chance of being
selected.
 The chance of a unit/element being chosen into the sample is
‘known’ by its exact probability or by statistical estimates.
 Key to a representative sample.
 It is a mechanism for reducing bias in the selection of the
sample.
 Provides accurate data about too large groups.
 Efficient system to capture the variation in a small group. Every
element of the population has equal chance of being selected in
the sample.
 Allows researcher to estimate the accuracy of estimated
23
parameter of population by sampling error.
 Probability Sampling
 Allows to estimate confidence interval for parameter
thus allows generalization of sample results to a larger
population.
 Allows measurement of errors of estimation or the
significance of results obtained.
 Can provide accurate information about groups that are
too large to study in their entirety.
 Purely quantitative studies typically use larger samples
selected through probability techniques.
 There are five common methods of random sampling
namely;
i. simple random sampling,
ii. systematic random sampling,
iii. stratified random sampling,
iv. cluster sampling and 24
v. probability proportionate to size sampling.
 Non-Probability Sampling
 Probability of selecting any particular member is
unknown.
 Also known as ‘deliberate’ sampling because items
are selected deliberately by the researcher.
 Used when probability sampling would be
prohibitively expensive and/or when precise
representativeness is not necessary.
 Most qualitative studies use non-probability
samples because the focus is on in-depth
information and not making inferences or
generalizations.
 The four common types are;
i. Convenience
ii. Judgement
iii. Quota
iv. Snowball 25
 Error in Research

 All research will have some form of error.

 Keep the error component at a minimum because it has adverse
effects on the validity of a measure.

Error

Sampling Sampling-related Data collection Data

error error error processing
error 26
 Error in Research
Sampling Error
 This is the difference between a random sample and the
population from which it was selected.
 This kind of error arises because it is extremely unlikely
that one would end up with a truly representative sample
even when probability sampling is used.

Sampling-Related Error
This is the error that arises from activities that are related
to the sampling process and are connected with
generalization or external validity.
Examples are the inaccurate sampling frame and non-
response.

27
 Error in Research
Data collection error
 This is the error that is connected with the
implementation of the research process.
 The source of error includes factor such as poor question
wording of self-completion questionnaires and structured
interviews; poor interviewing techniques; and flaws in the
administration of research instruments.

Data processing error

 This arises from faulty management of data, in particular
error in the coding of answers.
 This error is related more to validity of measurement than
sampling.

28
 Response Rate
 When a social survey is conducted whether by structured
interviews or by self-completion questionnaires invariably, some
people in the sample may not participate.
 On the other hand, some responses can not be taken seriously
thus may be classified as unusable.
 Response rate is calculated as follows;

number of usable questionnaires

X 100
total sample – unsuitable or uncontactable members of the
sample
 Response rates are very important in research because the
lower the rate, the more questions are likely to be raised about
the representativeness of the achieved sample.
 Low response rate introduces bias.
 Mangione (1995) cited in Bryman and Bell (2003) asserts that
response rates of below 50% are not acceptable.
29
 References

Bryman, A., and Bell, E., (2003) Business Research Methods, New
York: Oxford University Press.

Guy, R. F., Edgley, C. E., Arafat, I. and Allen, D.E. (1987) Social
Research Methods, Boston: Allyn and Bacon.

Kothari, C. R., (2007), Research Methodology: Methods and

Techniques, New Delhi: New Age International.

Leedy, P. D., (1993) Practical Research: Planning and Design, 5th

edition, New York: Macmillan.

Mouton, J. and Marais, H. C. (1996), Basic Concepts in Methodology of

the Social Sciences, Pretoria: Human Sciences Research Council.
30