PGDE RESEARCH METHODS CHAPTER THREE –

Part 3
Chapter Three

Methodology

Research Designs

Once the research problem has been defined, variables defined, objectives determined,
hypotheses formulated, the frame of reference described; the researcher confronts the problem of
constructing a research design which makes it possible to test the hypotheses. A research design
is the programme or plan which guides the researcher as he / she collects, analyses and interprets
the findings. It serves as the architectural blueprint or roadmap of the research project. It ensures
that the data collection and analysis activities used to conduct the study are tied adequately to the
research questions / hypotheses, and that the complete research agenda is executed.

A research design is the arrangement of conditions for collecting and analysis of data in a
manner that aims to combine relevance to the research purpose with economy in procedure. It is
in fact a conceptual structure within which research is conducted; it constitutes the blueprint for
the collection, measurement and analysis of data.

A research design should at least contain: -

(a) A clear statement of the research problem;
(b) Procedures and techniques to be used for collecting data;
(c) The population to be studied;
(d) Methods to be used in processing and analysing data.

The research design is needed because it facilitates the smooth sailing of the various research
operations, thereby making research as efficient as possible yielding maximum information with
minimum expenditure of effort, time and money. The research design stands for advance
planning of the methods to be adopted for collecting the relevant data and the techniques to be
used in their analysis, keeping in view the research objective and availability of necessary
resources.
Types of Research Designs

Research designs are categorized according to whether they belong to the quantitative or
qualitative research approaches.
A. Experimental Research Designs

In quantitative research designs, the findings are expressed in terms of numbers. These designs
are also called Experimental Research Designs. The primary purpose in conducting an
experimental study is to test the existence of a causal relationship among two or more variables.

In an experimental study, one variable, the IV, (also called treatment) is varied or manipulated
systematically so that its effects on another variable, the DV, can be measured. As the treatment
is being manipulated, the researcher controls for the extraneous variables, and then observes the
effect(s) on the dependent variable(s).

Experimental studies can take place in laboratories as well as outside social settings. An
experiment involves the comparison of the effects of a particular treatment with those of a
different treatment or no treatment. That is, in experimental research there is reference to an
experimental group which receives the treatment (IV) and a control group which is not exposed
to the treatment of interest.

The distinguishing characteristic of an experimental study is the random assignment of

individuals to groups or treatments. Random assignment is used to control all biases at the time
of assignment and to ensure that only one variable (treatment / IV) differs between the
conditions. An experimental study is the most appropriate approach to choose when studying
cause-and-effect relationships.

Types of Experimental Research Designs

There are two broad categories of Experimental Research Designs: -

(a) True Experimental Research Design (With random assignment of subjects or individuals
to groups and have control groups)
(i) Pretest – Post-test Control Group Design

Subjects are randomly selected from the population and randomly assigned to the groups. Both
groups are given a pre-test to determine if they are essentially the same on the DV. The
experimental group receives the treatment / IV, while the control group does not receive the
treatment – it receives something else. Then both groups are given a post-test. The treatment
effect for both groups is obtained by subtracting the results of the pre-test from the results of the
post-test.
(ii) Post-test Only Control Group Design

Subjects are randomly selected and randomly assigned to groups. The experimental group
receives the treatment (IV) while the control group receives something else. A post-test is
administered to both groups. The treatment effect is obtained by subtracting the post-test scores
of the control group from the post-test scores of the experimental group.

(b) Quasi-Experimental Research Designs

These designs have the same primary purpose as experimental studies, to test the existence of a
causal relationship among two or more variables. They become necessary when random
selection and assignment of subjects is impossible to achieve. However, intact groups can be
used. Therefore, a quasi-experimental design is a fallback strategy for situations in which
random selection and assignment are not possible.

The major strength of the quasi-experimental design is that it provides an approximation to the
experimental design and that, with care, it can support causal inferences. The greatest
vulnerability to quasi-experimental designs is the possibility that the comparison base created is
biased. It is often open to various types of threats to internal validity due to lack of
randomisation.

One common example of quasi-experimental research design is: -

 Non-Equivalent (Pre-test – Post-test) Control Group Design

This design is much like the Pre-test – Post-test Control Group Design except that it does not
involve random selection and assignment of subjects. There is some control over the influence of
extraneous variables though the use of the control groups.

In terms of procedure, there is no random selection of subjects from the population but intact
groups are used which may or may not be randomly assigned to the conditions. A pre-test is
administered to both groups to determine how nearly equivalent they are. The experimental
treatment (IV) is administered to the experimental group. Both groups are then post-tested. The
treatment effect for both the experimental and control groups is obtained by subtracting the
scores of the pre-test form the scores of the post-test. The two results are then analysed with a t-
test.

NON-EXPERIMENTAL RESEARCH DESIGNS

In non-experimental research there is no introduction and manipulation of treatments or IVs.

Rather the researcher obtains measures from the subjects to describe something that has
occurred, occurs or is occurring. Non-experimental research is not intended to show the cause-
and-effect relationship at that very moment of the study.

General examples of non-experimental research include: -

Ex-post facto (Causal Comparative Design)

Ex-post facto = from Latin: ‘After the fact’.

Thus this design attempts to determine the causes or consequences of differences which already
exist among groups or individuals. The researcher just looks at the phenomenon and tries to
explain the causes by looking back for possible reasons. It also attempts to establish the cause
and effect relationship without manipulation of a treatment. Subjects who have been exposed to a
stimulus and those who have not been so exposed are studied.

Example:

A researcher might hypothesise that attending a boarding primary school education is a major
contributing factor for the apparent differences in social adjustment among Senior One students.

The researcher would then select a group of S1 students who attended boarding primary schools,
and another group of those who did not. Both groups are then studied on social adjustment. The
observations about the two groups are the compared. If the group which attended boarding
primary schools exhibits a higher level of social adjustment, the researcher’s hypothesis would
be supported. The researcher needs to be cognisant of the influence of the uncontrolled variables
(EVs).

Descriptive Research Designs

The purpose of descriptive research is to describe the current state of affairs at the time of the
study in order to provide a picture of a phenomenon as it naturally occurs, as opposed to studying
the impacts of a phenomenon. It is a process of collecting data in order to test hypotheses or to
answer questions concerning the current status of the study subjects.

A descriptive researcher has no control over “what is”, and can only measure what already exists
(i.e. there is no manipulation of treatments). Descriptive research data are usually collected
through questionnaire surveys, interviews, discussions or observations. Thus randomisation is a
crucial characteristic feature of the descriptive design.

Descriptive approaches are appropriate when the researcher is attempting to answer “what is” or
“what was” questions, normative questions, as well as assessment of attitudes, opinions,
demographic information, conditions and procedures.

Advantages of descriptive research include lower costs, relative ease of implementation, and the
ability to yield results in a fairly short time.

The two broad categories of Descriptive Research Design are: -

Correlational Research Design

Correlational studies attempt to establish the degree of relationship between two variables in the
same individuals. Only one group of subjects is involved in the study and there is no introduction
and manipulation of the treatment (IV). The scores of the two variables are obtained from the
same individuals and then correlated.

The intent is to establish whether variables share something in common with each other. If they
do the variables are said to be positively correlated. They can also be negatively correlated.
Generalisation to a defined population is defensible if the sample was randomly selected and if
the mortality is very low.

Correlation is done using a suitable Correlation Co-efficient method from the available methods.

Note: the correlation method is best used when one individual is studied on the two or more
variables.
Survey Research Designs

A survey is an attempt to collect data from members of a population in order to determine the
current status of that population with respect to one or more variables of interest. This may
involve assessment of a variety of types of information such as attitudes, opinions, values,
beliefs, perceptions, practices, characteristics, demographic information, leadership dimensions,
mortality dimensions, performance dimensions, satisfaction, and so many others.

Two types of the Survey Research Designs are:

Cross-sectional survey
In this design data are collected from a random sample of subjects / participants drawn
from the accessible population. The random sample represents a cross-section of the
population. Thus a cross-section survey includes different categories of subjects who are
studied at one point in time. In cross-section surveys data may be collected by using
questionnaire, interview or observation methods.

Longitudinal survey research designs, which include: -

 Panel Survey Research Design
 Trend Survey Research Design
 Cohort Survey Research Design

Sampling Strategies / Techniques / Procedures / Methods

Population in the context of research

In the context of research, population refers to the entire group of individuals, objects, or events
that share common characteristics and are of interest to the researcher.

It represents the total set from which a researcher might draw conclusions. The population is the
focus of the study, but researchers typically do not study the entire population due to practical
constraints like time, cost, and accessibility.

Instead, they select a sample from the population to gather data, with the goal of making
inferences about the population as a whole.

Key concepts related to population in research:

1. Target Population: The larger group the researcher is interested in and aims to
generalize findings to (e.g., all secondary school students in a study area).
 2. Accessible Population: The subset of the target population that is realistically available
for the researcher to study (e.g., secondary school students from a specific selected
schools).
3. Sample: A smaller, manageable group selected from the accessible population for the
actual study. This sample should ideally be representative of the population to ensure
valid generalizations.
4. Sampling Frame: A list or database from which the sample is drawn, containing all the
members of the accessible population.

In sum, the population is the broad group the research aims to understand, and the sample
provides a practical way to study that population.

Researchers collect data to test hypotheses, answer research questions and provide empirical
support for explanations and predictions.

When data have been collected and analysed the subsequent explanations and predictions must
be capable of being generalised in order to be of scientific value.

Generalisations constitute a major stage in the research process. Typically, generalisations are
not based on data collected from all the observations, all the respondents, all the events, etc., that
are defined by the research problem.

Instead researchers use a relatively small number of cases, or a sample as the basis for making
inferences about all the other cases (a population). It is difficult, impractical, and very expensive
to collect data from all the potential units of analysis covered by the research problem.

That is why it is advisable to use a sample rather than the whole population.

Collecting data from every member of the population is a census.

The sample is drawn from the accessible population rather than from the target population. The
sample is carefully selected so as to be representative of the whole population with regard to the
relevant characteristics of interest.

Each member of the sample is called a subject / participant / respondent.

Generally, sampling enables the researcher to study a relatively small number of units in place
of the target population, and to obtain data that are representative of the whole target population.

Therefore, sampling is the process of choosing the units of the target population which are to be
included in the study.

Reasons for sampling

  In many cases complete coverage of the entire population is not practical.
 Sampling is better than complete coverage of the population since it addresses the survey
population in a short period of time and produces comparable and equally valid results.
 Studies based on samples are not time consuming and produce quick results.
 Sampling is less demanding in terms of workforce since it deals with a small portion of
the target population.
 It is less costly in most aspects.
 Studying a sample rather than the entire population is sometimes likely to lead to more
reliable results, mostly because fatigue is reduced, resulting in fewer errors in collecting
data form fewer elements.

Principles of sampling

For samples to be representative, they are expected to be chosen by means of sound

methodological principles, as shown here below: -
 Sample units should be chosen in a systematic and objective manner.
 Sample units must be easily identifiable and clearly defined.
 Sample units must be mutually exclusive, i.e. independent of one another, uniform and of
the same size, and should appear only once in the population.
 Sample units are not interchangeable, i.e. the same units should be used throughout the
study.
 The selection process should be based on sound criteria and should avoid errors, bias and
distortions.

Sampling errors

These are errors which arise because observations and measurements are taken from the sample
rather than the whole population.

They are defined as the differences between the statistic obtained from a sample survey and the
parameter obtained in a census.

Determining Sample Size

The sample size should be optimum – not too big nor too small.

An optimum sample size is one which fulfils the requirements of efficiency, representativeness,
reliability and flexibility.
In determining the sample size to use, the following considerations need to be bone in mind: -
(i) Homogeneity / Heterogeneity of the population. In a homogeneous population a
small sample can serve but in a heterogeneous population a larger sample is
required.
(ii) Nature of the study. In qualitative studies small samples can suffice while in
surveys large samples are needed.
(iii) Availability of qualified staff. Big samples require many research assistants.
(iv) Time available.
(v) Availability of other resources.

A big sample size is preferred to a small sample size because of representativeness. Small sample
sizes tend to increase sampling errors.

There are situations when large samples are required; as follows:

 When many variables are held constant in the study.
 When hypothesised relationships or differences between variables are very
small in real life.
 When the study requires the sample to be divided into sub-groups.
 When high mortality rate is anticipated.
 When the accessible population is highly heterogeneous.

Sample size can be determined in a number of ways. Two common methods are:-
(i) Using the Table developed by Krejcie, R. V., and Morgan, D. W. (1970).
Determining sample size for research activities. Educational and
Psychological Measurement, 30, 607 – 610.
Search the internet for this Table.

(ii) Using formulas. One of such formulas is Yamane’s formula: -

n = Sample size

N = Total population

e = Margin error (usually 0.05)

(iii) Some authors have suggested that 30% of the total target population is a suitable sample
size: cf. Mugenda, O.M. & Mugenda, A.G. (2005). Research Methods. Nairobi. East African
Education Publishers Ltd.

Sampling Techniques / Strategies / Procedures / Methods

These are the techniques / procedures used to select the study subjects to be included in the
sample. If the sample is selected scientifically it is highly possible to generalise the findings to a
larger population. Selecting a representative sample from the sampling frame (list of target
population) presupposes normal distribution of population characteristics. When the
characteristics of the population are not over-represented or under-represented in the sample that
sample will be effectively representative. The representative the sample is of the population, the
more generalisable are the study findings. Generalisability is one of the hallmarks of scientific
research.

Characteristics of a good sampling technique: It should: -

(a) Result in a truly representative sample.
(b) Result in a small sampling error.
(c) Be financially viable.
(d) Enable better control of systemic bias.
(e) Result in generalisability of the findings to wider populations with
reasonable confidence.

Sampling techniques / procedures or strategies are grouped in two categories:

 Probability Sampling
 Non – probability Sampling

PROBABILITY SAMPLING

In probability sampling the elements in the population (sampling units) have a known /
predetermined chance or probability of being selected as sample subjects and thus being included
in the study. The goal of probability sampling is to select a reasonable number of subjects,
objects or cases which represent the target population. Probability sampling provides accurate
information about groups which are too large to study in their entirety. It provides an efficient
system of capturing variations or heterogeneity in a small group, which exist in the target
population.
Probability sampling is essentially random sampling. Random sampling is the key to obtaining
a representative sample. Therefore, probability sampling strategies are used when
representativeness of the sample is of importance in the interest of wider generalisability.

Probability Sampling Techniques / Strategies

1. Simple random sampling

Here each sampling unit has an equal chance of being selected and included in the sample. If the
sample size is n and the accessible population is N, the probability / chance of each sample unit
being selected is . Here the presupposition is that the sample size has been determined,
and the target population is known; and the list of the sampling units (sampling frame) is
available.

Procedure

Lottery method (Gold fish bowl)

Each of the names of the sampling frame (sampling units / elements), or the serial number is
written on a slip / small piece of paper. These pieces of paper are folded properly and are put in a
container, and are mixed thoroughly. One piece of paper is picked at a time, without
replacement. The name or number of the selected element is recorded. The process continues
until the sample size is obtained.

Use of random numbers / digits

A Table of random numbers (random digits) already exists or it can be generated on a computer.
The names of the target population are compiled in a sampling frame, numbered from the first to
the last. For samples between 10 and 99, two columns of numbers will be used; between 100 and
999, three columns will be used; between 1000 and 9999, four columns will be used, and so on.

The researcher blindly points at a number in one column / row preferably with a pencil. The
number under the pencil is within the sample size it is recorded to be included in the sample. The
process continues in a uniform direction by rows or columns. If the number is higher than the
sample size it is dropped and the exercise continues until the sample size is obtained. Similarly,
the number which has already been picked is discarded.
N.B.

The starting point can be anywhere in any row or column and then move in any direction,
downwards, upwards, left or right; as long as the decision is taken in advance, strictly and
consistently followed.

Advantages of random sampling

i. It gives each member of the population an equal chance of
being selected and included in the sample.
ii. Has least bias.
iii. Offers the most generalisability.

Limitations of simple random sampling

i. It is cumbersome.
ii. It is costly.
iii. It does not guarantee an exact representation of a population
as it is possible to draw very ‘biased’ samples.
iv. An updated list of elements in the population (sampling
frame) may not always be available.

Systematic random sampling

The list of the elements in the population is randomly arranged. Alphabetical or numerical order
is not random.

To be systematic a sampling interval / ratio is determined by dividing the total target population
by the predetermined sample size. E.g. if the target population is 1500 and the sample size is
300, the sampling interval / ratio is 1500 ÷ 300 = 5. Randomly select a number between 0 and 5
to be starting point on the sampling frame. That random number will mark the first selected
participant. Continue by keeping adding 5 to the subsequent result: i.e. if the starting point is 3,
add 5 to 3, i.e. 3+5=8; 8+5=13; 13+5=18; and so on … until the sample size is reached. Locate
the names on the sampling frame of 1500 which correspond to the numbers drawn: 8, 13, 18 …
All those located names constitute the study sample of 300.
Advantages of systematic random sampling
i. It is simple, quick, easy and foolproof to perform.
ii. It is a convenient method especially for ill-trained research assistants.
iii. It is relatively easy to get samples from large populations.

Limitation

If the sampling frame possess a regular pattern occurring systematically at very sampling interval
point, the method may produce a very biased sample. If there is an identified systematic pattern
in the sampling frame, the list can be shuffled. Secondly the starting point can be randomly
selected again.

Stratified sampling

Stratified sampling is used for two main purposes:

a. When a researcher wishes to ensure that particular categories in the
population are represented in the sample in the same proportion as in the
population: e.g. gender, age groups, staff category, class level, marital status,
religion, etc. here the ultimate goal is to achieve the desired representation
from various sub-groups of the target population.
b. To ensure that there are sufficient numbers in the sample from all categories
that are to be studied. In stratified random sampling subjects are selected in
such a way that existing sub-groups in the target population are more or less
reproduced in the sample.

The sub-groups in both the target population are called strata (one is a stratum). Each sub-
group / stratum is expected to be reasonably homogeneous on the variable / characteristic of
interest. The sub-groups should be mutually exclusive (or non-overlapping), relevant,
appropriate and meaningful in the context of the study.

In using stratified sampling the researcher should first decide on which criteria the population
and thus the sample will be stratified. After deciding on the criterion of stratification, the next
step is to decide on the size of each stratum / sub-group in the sample. The number of subjects in
each stratum of the sample should conveniently be consistent with the proportion of the sample
to the whole target population.

After dividing the population into strata / sub-groups, the sampling frames of each of the sub-
groups are identified or developed if they are not available. Using either simple random sampling
or systematic random sampling, a sample is drawn from each sub-group. The separate samples of
the sub-groups are then merged into one to form the overall study sample. The identified subjects
/ participants are then approached for data collection.

Advantages of stratified sampling

i. Elements in each sub-group are homogeneous. This is based on the
theory that as the sub-groups get smaller and smaller, the elements in
those sub-groups become more and more homogeneous, and the
differences between them become greater and greater.
ii. If well utilized and applied the design yields more precise and accurate
results than other sampling designs. It is efficient in the sense that it
provides more information with regard to a given sample size. Thus,
stratified sampling results in more reliable and detailed information.
iii. It is relatively cheaper to execute.
iv. It ensures inclusion in the sample of sub-groups which would
otherwise be omitted by other sampling methods because of their small
numbers in the population.

Disadvantages of stratified sampling

i. It necessitates clear knowledge about the homogeneity or heterogeneity
of the characteristics in the population.
ii. Computing the proportions may be tedious.
iii. If an unrepresentative sample is obtained to get results for the variables
under study, there might arise sampling errors. The separate sampling
errors for all the strata might accumulate into a grand error which may
become significant enough to affect the study findings.

Cluster sampling

Cluster sampling is applied when it is not possible to obtain a sampling frame because the
population is either very big or scattered over a large geographical area, or when cluster criteria
are significant for the study.

It involves dividing the target population into a number of relatively small sub-groups which are
themselves clusters of still smaller units, and then some of these clusters are randomly selected
for inclusion in the overall sample. Ideally the clusters should have heterogeneity among the
elements within each unit. All members of small unit / cluster are then included in the sample
with each member becoming a sampling unit.
When several groups with intra-group heterogeneity and inter-group homogeneity are found,
then random sampling of the clusters can ideally be done and information is collected from each
of the elements in the randomly selected clusters.

Examples of clusters include schools, classes / streams, subjects, dormitories / hostels, etc.

Advantages
i. It is relatively less costly than other sampling strategies. Surveys are
concentrated in the selected clusters rather than scattering the
resources in widely distributed elements.
ii. It is used to overcome the problem of constructing a sapling frame
since in many cases reliable sampling frames are not available.
iii. It is convenient in the sense that elements in the cluster are found in
the same place.

Disadvantages
i. It renders itself to greater biases.
ii. It is the least generalisable because naturally occurring clusters the
organisational context do not contain heterogeneous elements, i.e. the
conditions of intra-cluster heterogeneity and inter-cluster
homogeneity are not often met.
iii. It does not offer much efficiency in terms of precision or confidence
in the results. That is, it is less precise than random sampling, and it
does not yield as information as in random sampling.

Multi-stage sampling

In this method a sequence of samples is drawn from samples already selected but only the last
sample of subjects is studied. With every additional drawing, the sample becomes more specific
and more relevant to the research problem, and the results are expected to become equally
relevant and more representative. The procedure of multi-stage sampling is as follows:

A sampling frame for the target population is identified. A large probability sample is selected.
The sampling units of this sample are usually called the primary selection units. Another sample
from the primary selection units is the selected. After the criteria of the respondents have been
identified another sample is drawn from within the previous sample. The procedure is repeated
until the targeted sample size is reached.
Example 1:

Suppose we are to study the attitudes of BSU university students towards campus residence. A
list of all university students is obtained from the Academic Registrar. From this list a large
number of respondents, say 2000 is selected. These respondents are screened with regard to
socially significant criteria, such as gender and year of study. A sampling frame is then
developed according to these criteria; and some 600 respondents are selected. These respondents
are further screened on the basis of the relevant criteria. A new sampling frame is developed on
the basis of these criteria. These respondents are further screened to identify additional relevant
criteria, and a sampling frame is developed on the basis of these criteria. From this sampling
frame a proportion of suitable respondents, say 30% are selected, who will constitute the final
research sample to be studied.

Example 2: In pilot surveys.

In a study to cover the whole country:

 Randomly select two districts
 In each of the selected districts randomly select two counties
 In each of the selected counties, randomly select two sub-counties
 In each of the selected sub-counties, select two parishes
 Develop a sampling frame of each selected parish from which
relevant final samples are drawn to be studied.

Advantage
 It allows the establishment of a sample that is directly related to the
research objective.

Limitation
 Achieving a degree of population validity / representativeness
between the selected clusters and the target population is often
difficult, thereby limiting making credible generalisations.

B. NON-PROBABILITY SAMPLING

In non-probability sampling the sampling units / elements do not have a known chance of being
selected and included in the sample as subjects. There is no way of specifying the probability of
each sampling unit’s inclusion in the sample. There is no guarantee that every sampling unit has
some chance of being included. Non-probability sampling techniques may be preferred when
time or other factors become critical and supersede generalisability.

Although researchers can make accurate estimates of the population’s parameters only with
probability samples, social scientists may use non-probability samples especially for reasons of
convenience and economy, which, under certain circumstances may outweigh the advantages of
using probability sampling. Non-probability sampling is also used when a sampling population is
not readily available; and when the researcher is not interested in selecting a representative
sample.

Most qualitative studies use non-probability sampling designs because the focus is on in-depth
information but not so much on making inferences or generalisations. In non-probability
sampling the elements / sampling units in the sampling frame (or population) have no
probabilities attached to their being selected as sample subjects. Thus the findings from the study
of a non-probability sample cannot be confidently generalised to the wider population.

Non-probability sampling techniques / designs / strategies / methods

Purposive or Judgement Sampling

In this sampling design the researcher selects sampling units subjectively in an attempt to obtain
a sample which appears to be representative of the population. Researchers purposely select
subjects, who, in their opinion, are thought to be relevant to the research problem. In this case the
judgement of the researcher is more important than obtaining a probability sample.

The chance that a particular sampling unit will be selected for the sample depends on the
subjective judgement of the researcher. Thus it is difficult to determine the inclusion of any
specific sampling unit in the sample. Purposive sampling allows the researcher to use subjects
who have the required information with respect to the objectives of the study. This is because
they are the only ones who possess the required information, or who conform to some criteria set
by the researcher, i.e. they possess the required characteristics, who are in the best position to
provide that required information. It is used when a limited number or category of subjects have
the require information.

E.G. when studying secondary school students’ discipline and teachers’ commitment to duty,
headteachers are selected purposively.

Advantage
It is the only viable method for obtaining the type of information that is required from very
specific pockets of people who alone possess the needed facts and can give the information being
sought.

Disadvantage

It curtails generalisation of findings because only a sample of experts who are conveniently
available are the ones used.

Quota Sampling

Quota sampling is a version of stratified sampling with the difference being that instead of
dividing the population into strata and randomly selecting a number of respondents, it works on
‘quotas’ set by the researcher. It is in effect a non-random stratified sampling procedure.

The main aim of quota sampling is to select a sample which is as similar to the sampling
population as possible. It is commonly practiced when it is impossible, difficult or costly to
identify members of a population. It also aims at including various groups or quotas of the
population in the study based on some identified criteria. A set of selection criteria is identified
because of their relevance to the research problem. For example, in a study of undergraduate
university students, three selection criteria might be used, namely: gender, year of study and
degree being pursued. By so doing a number of selection categories will be produced. The
researcher would then have to decide how many respondents are to be selected in each category.
There are two main possibilities: equal numbers in each category or numbers proportional the
category numbers in the population.

Example:

Supposing a researcher wishes to take a sample of 800 adults in Mbarara City. It is known that
adult males constitute 50% of the adult population and females also constitute 50% of the adult
population in Mbarara City. The researcher would take 50% of the adult male population (i.e.
400 adult males), and 50% of the adult female population (i.e. 400 adult females). Individual
adult males and females can be selected non-randomly until the targeted sample of 800 is
reached. Generally the quota fixed for each sub-group is based on the total number of each group
in the population.

Advantages of quota sampling

i. It can produce a sample with similar distribution of
characteristics to those that are considered to be important in
the population it is supposed to represent.
ii. It is conveniently economical, easy to administer and quick to
execute.
iii. It does not require a sampling frame, and it is relatively
effective.
iv. It can produce adequate results.
v. It also becomes necessary when a sub-set of the population is
under-represented in the organisation, such as females or males
among doctors, engineers, nurses, security organs, etc.
Therefore, quota sampling ensure that all the sub-groups in the
population are adequately represented in the sample.

Disadvantages
i. The sample selected with quota sampling may not be totally
representative.
ii. Since quota sampling is a non-probability sampling plan, the
generalisability of its findings is limited.

Convenience or accidental sampling

It involves the researcher taking the sampling units as they become available to him / her, from
whom information is sought and collected. The researcher can stand at the entrance of the school
to seek teachers’ and students’ views about the headteacher’s financial management practices,
social interactions with staff and students, etc. Some opinion palls studies use this method.
Media reporters use this method a lot.

Other terms used for this method are: volunteer, haphazard, grab and chunk sampling.
Advantages
i. It yields quick results.
ii. Convenience sampling is most often used during the
exploratory phase of a research project and is perhaps the best
way of getting some basic information quickly and efficiently.
Limitations / Drawbacks
i. Convenient samples are not representative of any population.
ii. The findings (collected views) can be biased.
iii. Findings of convenient sampling have least generalisability.

Network or Snowball sampling

This method is employed when the lack of sampling frames makes it impossible for the
researcher to achieve a probability sample, when the target population is unknown, or when it is
difficult to approach the respondents in any other way. Researchers begin with few respondents
who are available to them. These initial subjects with the desired characteristics are identified
using purposive sampling or other techniques. The few identified subjects name others who have
the required characteristics until an exhaustive sampling frame is obtained. Every time new
subjects are found they are interviewed and information is got from them.

Examples of such subjects are drug addicts, defiled children, sex workers, homosexuals,
witchdoctors, robbers, buttered men, etc.

Advantage / Justification

The method is useful when the population that possesses the desired characteristics is not known
and difficult to find.

N.B.

It is a sensitive sampling method which must be used carefully and cautiously.

DATA COLLECTION METHODS

Data collection methods and corresponding instruments:

 Data collection method Instrument used
Questionnaire method Questionnaire
Observation Observation checklist
Interview Interview guide
Focus Group Discussion FGD guide
Documents review Documents review checklist
Physical measurement Actual / relevant tools

QUESTIONNAIRES

A questionnaire is a printed / typed self-report form designed to elicit information from

respondents in written form. The respondents write their responses on the questionnaire against
the written questions / statements / items. Questionnaires are commonly used in surveys.

Questionnaires may be constructed by researchers, but there are also several existing
questionnaires, called standard questionnaires, which were developed and tested by earlier
expert researchers. Search the internet extensively to establish whether your research problem
has an existing / standard questionnaire.

The questionnaire method may be applied in three ways:

i. Self-administered questionnaires
ii. Mail questionnaires
iii. Electronic questionnaires (delivered electronically)

Advantages of Questionnaires in General

1. Mail questionnaires are less costly than other methods even when
the population is large and widely spread over a large
geographical space.
2. Questionnaires produce relatively quick results.
3. Respondents can complete them at their convenience when there
is no stress or pressure of work.
4. They offer greater assurance of anonymity.
5. They are a stable, consistent and uniform measure, without
variations.
6. They offer less chances for bias and errors caused by the presence
or attitudes of the researcher / interviewer.
7. They offer a considered / true and objective view on the issue,
since respondents can consult records of reflect.
Limitations
1. Only limited to literate respondents who can read and write.
2. Time consuming.
3. Self-administered questionnaires are costly in terms of resources
and energy.
4. No opportunity for spontaneous responses which are at times good
and valid, such as facial expressions or certain sounds.
5. Responses are taken as final regardless of their validity, even
when the respondent misunderstood the question or statement in
the questionnaire.
6. They do not allow probing, prompting or clarification of questions
or statements.
7. There is no opportunity for useful supplementary observation data.
8. The identity of the respondent and the conditions under which the
questionnaire was completed are not known. One cannot be sure
that it was the selected subject who completed the questionnaire,
and that undue assistance was sought.
9. It is not possible to ascertain that the order of items was followed.
10. It is likely to have a high rate of mortality and partial responses.

Questionnaire construction / design

The first step in designing a questionnaire is to identify and specify the needed information to
achieve the stated objectives, answer the research question(s) and / or test the hypotheses.
Therefore, the questionnaire items should necessarily be related to the study objectives, as
exemplified below:

Objectives Questionnaire items

Objective 1 1.......................................................

2.

………………………. 3.

4.

5.

6.
 Objective 2 7…………………………………...

8.

……………………….. 9.

10.

11.

12.

Objective 3 13………………………………….

14.

……………………….. 15.

16.

17.

18.

In the example above each objective has six items. Each item attempts to solicit information
from the respondent which will enable the researcher to realise the corresponding objective, as
stated in the column on the left.

*****This step calls for a lot of constructive and proactive thinking and extensive reading.
Questionnaire construction is a demanding task which requires methodological competence and
adoption of questioning techniques.

The basic requirement of a questionnaire format is that the items should be arranged in a logical
order, allowing for a smooth transition and flow. That is, a smooth passage from one idea to the
next, avoiding distortions, repetitions and problems.

It is important that the items are interesting and relevant to the problem being investigated.
The presentation and structure should make the respondents feel at ease and worthy, rather than
the subjects of a strict interrogation.

An adequate and appropriate format ensures a frictionless / smooth completion of the

questionnaire; it allows the respondent to feel part of the research process; helps to avoid fatigue
and boredom; maintains a positive attitude to the study; and makes it possible and more likely for
the respondent to complete and return the questionnaire.

Structure of the questionnaire

The first part of the questionnaire is the Introduction Statement. This includes the introduction of
the researcher (names, institution and programme of study being pursued), study title and its
purpose, reason(s) for selecting the respondent, assurance of anonymity and confidentiality,
appeal for sparing time to complete the questionnaire as candidly as possible.

In this part the researcher should spell out the instructions on how the respondent can express
their choices / preferences. Instructions should offer as much information as possible and need to
be written in a simple language. Inadequate instructions are a major cause of non-responses.

The second part is one of relevant demographics, i.e. the personal bio-data of the respondent.
These include: gender, age, occupation (if not implicit), length of service, educational
qualifications, marital status, type of institution, etc. Include only those that are relevant to the
study which will improve understanding of the findings. ***Avoid sensitive information if it may
not add value to the study.

The third part, which is the main body of the questionnaire, includes the items which respondents
have to deal with. To be effective, this part has to be worked out very carefully with regard to
content, structure, wording, flow, format, etc., and to adhere to the basic principles of
questionnaire construction. It is the part that will enable the researcher to collect the relevant data
to complete the study.

Size of the questionnaire

The questionnaire size depends on factors like research objectives / questions, types of
respondents, methods of analysis and availability of resources, among others. However, the
golden rule with regard to questionnaire size is that one should include as many items as necessary
and as few as possible.

Essentially, the questionnaire size depends on the number of variables and the number of
indicators considered in the study objectives, as well as the number of items required to fully
address the identified indicators.
Example:

Supposing one variable of the study is socio-economic status. Here indicators can be:
occupation, income and education. Each one of indicators will need a number of items to address
it as shown below:

Occupation: - Item 1: Occupation of respondent

Item 2: Occupation of respondent’s spouse

Item 3: Occupation of respondent’s parent

Income: - Item 4: Income of respondent

Item 5: Income of respondent’s spouse

Item 6: Income of respondent’s parent

Education: - Item 7: Education of respondent

Item 8: Education of respondent’s spouse

Item 9: Education of respondent’s parent

N.B.

If information about the respondent’s second parent of siblings, other items can be added on
each indicator.

Questionnaire items

Questionnaire items are of different types.

Primary items: These elicit information directly related to the research problem. Each item
provides information about a specific indicator of a particular variable.

Secondary items: These are items which do not relate directly to the research problem. They are
of secondary importance in that they provide information on secondary issues. They do not add
new information about the research problem.
Direct and indirect items: These ask the respondent to give information about self, such as bio-
data, opinions, beliefs, practices, etc. On the other hand, indirect items are mainly used when
respondents are unable or unwilling to offer direct information on the research problem. This is
particularly so when the items deal with sensitive, embarrassing or threatening issues; or when
the research problem is too difficult for the respondent to give a direct response.

Suggestive items: These presuppose that the respondent holds a particular view similar to that of
the researcher on an issue in question; and contain an implied attempt to tempt the respondent to
confirm that view. E.g. a direct item can be: Do you believe that examinations should be
abolished in all post graduate programmes? In a suggestive mode, this item could read thus:
Don’t you also think that examinations in post graduate programmes should be abolished?
Suggestive items lead the respondent and bias the direction of the findings.

Main considerations in formulating questionnaire items

(i) To be successful the questionnaire should be comparatively short,
containing only the essentials.
(ii) Statements / words should be simple, direct, clear and familiar to the
respondents. Every statement matters / is important. Avoid difficult
words, unnecessary jargon and cumbersome phrases. Avoid too general,
ambiguous, vague and embarrassing statements.
(iii) Items should proceed in a logical sequence from easy to relatively
difficult items. Personal and intimate items can be left to the end.
(iv) Avoid double negatives. E.g. “Wouldn’t you be opposed to invigilate
students whom you don’t teach?”
(v) Avoid double barreled items. These are items which address two or more
ideas at a time. E.g. Teachers teach their lessons regularly and also
participate in co-curricular activities. One item should address only one
idea at a time. Each item should have a single purpose and seek only one
piece of information.
(vi) Avoid loaded / charged items. These are charged with words which
threaten self-esteem of respondents. E.g. How many times does your wife
batter you in a week?
(vii) Avoid irrelevancy. Only those items that are essential to the research
problem should be included in the questionnaire.
(viii) Items should convey a positive attitude towards the respondent and the
study in general, based on friendliness and collegiality. That is, it should
be user friendly.
(ix) Items should be formulated in the language of research understood by the
respondents. Use correct grammar and spellings, etc. It may necessitate
translating the questionnaire in a language of the respondents. E.g. When
 collecting data from rural parents, or on a refugee camp where refugees
do not speak Uganda’s research language, English.
(x) The questionnaire should be sufficiently attractive in terms of physical
appearance and quality of paper so as to attract the attention of the
respondents, and enlist their cooperation.

RESPONSE FORMAT
This is the format in which the respondent will indicate the responses to the items in the
questionnaire. There are basically two types of response formats: -
1. Free Response / Open-ended / Unstructured Format
2. Fixed Response / Closed-ended / Structured Format

Free response / open-ended format

Respondents fill in blank spaces or write unstructured responses to the items in their own
language, and their responses are recorded in full.

Advantages of Open-ended items

1. They allow freedom to express feelings and thoughts, especially when complex issues
are being studied.
2. They do not force the respondent to adapt to preconceived responses.
3. They offer more details than pre-coded responses, especially clarifications and
justifications.
4. They allow the respondent to show creativity, self-expression and initiative.
5. They offer information in areas that might not have been foreseen by the researcher.
6. They generate rich information.
7. They are less time consuming in construction.
8. They allow conclusions about the respondents’ way of thinking and logic.

Limitations
1. They are not suitable for sensitive questions.
2. They generate large amounts of information which require extensive time, effort and
expertise to code and analyse.
3. They are time consuming and allow no comparisons.
4. They may generate useless /irrelevant information.
5. They are not suitable for respondents who have difficulties with articulation in written
work.
 Fixed response / Closed-ended / Structured Formats

Here the respondent has to choose from the pre-determined responses. I.E. responses are fixed
and the respondent is expected to choose the option with which he / she agrees most. Fixed
response format items come in different forms. The most commonly used ones are:
 Rankings
 Multiple choice
 Scaled items

Rankings

A ranking forces respondents to place responses in a rank order according to some criterion.
Value judgements are then made and the rankings of each response can be summed up across
subjects thus obtaining an overall or group rank order of alternatives which can be analysed
using suitable statistical techniques.

Example

Rank the following activities with regard to how you like to use your leisure time:

Use numbers 1 – 5 with 1 as the most preferred and 5 being the least preferred.

--------- Watching television

--------- Taking a walk

--------- Reading a novel

--------- Playing in-door games

--------- Praying in prayer groups

Data processing and analysis

Respondent TV Walk Read Pla Pray

y
1 3 5 1 2 4
2 2 1 3 5 4
3 5 3 2 4 1
4 1 5 4 2 3
 5 4 3 2 5 1
n=5 15 17 12 18 13

∑ 3 3.4 2.4 3.6 2.6

Mean=
n

Final 3 4 1 5 2
Rank

Interpretation

In a sample of five respondents the order in which the subjects wish to spend their leisure time is
as follows: - 1. Reading, 2. Praying, 3. Watching, 4. Walking, 5. Playing indoor games. As
postgraduate students choosing ‘reading’ is an indication that they give priority to their academic
work. As adult learners they have very little time for playing. When they are not reading they get
committed with other job-related or domestic activities.

Multiple Choice Responses

This is where respondents are asked to choose one of the possible alternatives from those given.
Respondents are expected to either encircle / tick or write, the number of the selected alternative.
Respondents should be given specific instructions as to how they should show their choices.
They should use one uniform method. A two-response format (Yes/No, True/False) is only
useful for items of fact, but it is not recommended for measuring attitudes, opinions, perceptions,
feelings, etc., because people’s subjective reactions are not so clear-cut sufficiently enough to
fall neatly into one or the other category.

Construction of the response categories / descriptors requires much resourceful thinking. The
descriptors are primarily constructed so that they are accurate, uni-directional, exhaustive,
mutually exclusive, and clear (i.e. unambiguous).

The ultimate goal of such a construction should be: easy application, reduced stress on the
respondent, reduced time required for completion, high precision and making the instrument
user-friendly. Too many descriptors may overtax the respondents’ ability to differentiate,
especially those with below-average ability. Including a neutral option (undecided / no opinion)
depends on the nature of the study and encourages laziness.
Scaled Items Response Format

When items are set about behaviours, thoughts, feelings, attitudes, opinions, etc. that can vary in
intensity or frequency a scaled items response format should be used. A scale is a series of
gradations, levels or values that describe various degrees. Scaled items are some of the most
intensively used types of closed-ended statements in questionnaires because they allow fairly
accurate assessment of strength of beliefs or opinions. This is because most of people’s beliefs
and opinions are thought in terms of graduations: - very strong, strong, weak, very weak.

The usual format of scaled items is a statement or question followed by a scale of potential
preconceived responses. The respondent then ticks / crosses / encircles the place on the scale
which best reflects his / her belief / opinion, about the statement.

The most widely used example of a Scaled Items Response Format is the Likert Scale.

The Likert Scale

There are different Likert Scale types but the 5 response choice format is the most frequent and
preferred type.

The most common 5 – point Likert Scale reflects the strength of agreement or disagreement: -
Strongly Agree (SA); Agree (A); Undecided (U); Disagree (D); and Strongly Disagree (SD).
Items may be positively or negatively stated. E.g. A positive statement is: Research is very
important in post graduate university training. A negative statement is: Research is not
important in post graduate university training. The scoring of both types is as follows:

Positive items: SA A U D SD

5 4 3 2 1

Or 4 3 2 1 0

Negative items: SA A U D SD
 1 2 3 4 5

Or 0 1 2 3 4

The rule is:

 Positive or favourable items are scored 5 to 1 (or 4 to 0) with 5 representing
“strongly agree”.
 Negative or unfavourable items are scored 1 to 5 (or 0 to 4) with 1 representing
“strongly agree”).

All score values for all the responses in one hypothesis are then summed up to yield the
respondent’s score for the whole scale. An appropriate statistical / analysis technique is applied
to analyse the numerical data.

In the case of positive statements the higher scores show more positive attitudes towards the
subject / construct of interest.

The Likert Scale provides great flexibility since the descriptors on the scale can vary to suit the
nature of the item statement, and the research as a whole.

[Link]. Always, Most times, Sometimes Rarely, Never

Excellent Good Fair Poor Very poor

Advantages of fixed / closed-ended / structured questionnaires

1. They are easy to administer, to code and to answer.
2. They allow comparisons.
3. They allow quantitative scoring.
4. They help to produce fully completed questionnaires.
5. They help to avoid irrelevant responses.
6. They are quick and relatively inexpensive to analyse.
7. They are most amenable to statistical analysis.

Limitations
1. All possible alternatives / descriptors may not be covered.
2. They restrict the freedom, creativity and expressiveness of the respondent.
3. They allow a high likelihood of guess work.
4. Construction of the response categories / alternatives / descriptors, is difficult.
5. They may force a statement of opinion on the respondent about which he / she does
not in fact have any opinion.
6. They are inappropriate for complex issues.
Pre-testing the questionnaire

There are times when the researcher is not very sure about the effectiveness or suitability of all
or part of the instrument developed for the study; such as inadequate response categories or the
wording of items. In such cases pre-tests are employed.

Pre-tests are predominantly used to check eventual ‘mechanical problems; determine the
effectiveness of instructions; clarity of items / statements; and time required to complete the
questionnaire. For example, the researcher may be concerned that the inclusion of response
categories like ‘undecided, no opinion or other, please specify’ might not be effective; and would
wish to avoid such a problem.

In pre-testing the instrument it is best to locate a sample with similar characteristics similar to
those of the study population. A pre-test sample size of around 30 can do.

The administration of the pre-test should be the same as that to be used in the final study. The
pre-test respondents should be given space to write comments about individual items, and the
whole instrument in general.. this is to check if it takes long to complete the questionnaire, if
instructions and items are clear, and indeed if the whole instrument is effective and suitable.

Pilot Studies

A pilot study is a small-scale replication and a “dress rehearsal” of the main study. While pre-
tests help to solve isolated “mechanical problems” of an instrument, pilot studies are concerned
with administrative and organisational problems related to the whole study and the respondents.

Pilot studies serve many purposes which include the following:

1. To estimate the costs and duration of the study.
2. Test the effectiveness of the study organisation.
3. To test the suitability of the research methods and the instrument.
4. To show the adequacy of the sampling frame.
5. To estimate the levels of response and drop-out / mortality.
6. To gain information about homogeneity and heterogeneity of the study population.
7. To familiarize the researcher with the research environment.
8. To enable the researcher and assistants to practise research in real situations before
the main study begins.
9. To test the response of the respondents and the data collection methods, and through
that, the adequacy of its structure.

The structure and purpose of pilot studies vary from case to case, depending on the type of
research and the structure of the methodology used.
For example in the context of case studies pilot studies aim to establish whether respondents are
accessible, whether the study site is convenient enough, whether the data collection techniques
generate enough, insufficient or too much information, whether the plan is well adjusted, and
whether any changes or adjustments are needed.

In a nutshell, the purpose of pilot studies is to discover possible strengths, weaknesses,

inadequacies, ambiguities, challenges and problems, among others, in all aspects of the research,
so that they can be corrected before actual data collection takes place.

Instrument Reliability and Validity

Before any research instrument is used to collect data, the researcher must establish that it
possesses the two instrumental qualities of Reliability and Validity.
(1) Instrument Reliability
Instrument reliability refers to the consistency of measurement, i.e.: -
 The consistency of the instrument in measuring whatever it measures.
 The extent to which the results are similar from different forms of the instrument or
occasions of data collection.
 The degree to which an instrument will measure the same attribute more than once.
 The ability of an instrument to produce consistent results whenever it is used.

The reliability of an instrument may be conceptualised in terms of the amount of error of the
instrument.
- If the instrument has little error, it is reliable.
- If it has big error, it is unreliable.

Reliability of an instrument is expressed as a Correlation Co-efficient; usually the Pearson

Product Moment Correlation Co-efficient.

Reliability Co-efficient ranges in value from 0 to ± 1

-1_______________________________0______________________________ +1

Ideal reliability co-efficient should be 0.7 and above ˃ 0.7

Estimate of Reliability
Methods of estimating instrument reliability include, but not limited to the following:
(1) Measures of stability: -
A co-efficient of stability is obtained by correlating scores from the same test on two different
occasions of a group of same individuals. If the responses of the individuals are consistent, then
the correlation co-efficient and thus the reliability will be high. This is called a test-retest
procedure. It assumes that the variable / attribute being tested remains constant over time. The
time interval between the two tests should be long enough to avoid the effect of memory or
practice.

(2) Measures of equivalence: -

Two forms of instruments meant to measure the same variable are administered to the same
group, one after the other. The two obtained scores are then correlated using the Pearson Product
Moment Correlation Co-efficient. The reliability obtained reliability is a co-efficient of
equivalence.

(3) Measures of Internal Consistency ….. Also called Internal Consistency Reliability.
This is the extent to which the instrument has homogeneous content, i.e. the degree to which
items in the instrument measure the same thing.
There are a number of methods used to estimate the Internal Consistency Reliability of an
instrument. Three of them are: -
(i) Item–total Correlation Method
The researcher may examine Inter – Item reliability by looking at item – total correlation for
each item in the instrument. The researcher’s pertinent question is: (How) Does each item in the
instrument correlate with the sum of all other items?

Example:
Take item 1 in a questionnaire with many items. Item – total correlation would be computed as
follows: -

Respondent Score for @ respondent Total sum for @ respondent

On Item 1 on all other items

1 t1 T1

2 t2 T2

3 t3 T3

. . .

Key: t = Score on Item 1

 T = Score on remaining items (excluding Item 1)

The two measures are them correlated using the Pearson Product Moment Correlation Co-
efficient.

(ii) Split-half Reliability Method

Here the researcher administers the instrument to the target sample population. The items in the
instrument ate divided into two halves – usually the odd and even numbered items. For each
subject the scores for the odd and even numbered items are computed. The summations are then
correlated.
Note: this type of correlation co-efficient gives the instrument reliability for only a half of the
instrument.
In order to get the full length instrument reliability, a correlation must be made by the Spearman
– Brown Prophecy Formula, as follows: -

(iii) Cronbach Alpha Co-efficient

Cronbach Alpha Co-efficient is particularly useful when the items in the instrument are scorable
on more than two-point basis (Likert Scale). E.g. Strongly Agree, Agree. Undecided, Disagree,
Strongly Disagree.

The formula for Cronbach Alpha Co-efficient is: -

K = No. of items in the instrument

SD²i = Variance of individual / single items

SD²t = Variance of total instrument

Ideal 𝞪 = ≥ 0.70 i.e. 0.70 +

Cronbach Alpha is generally the most appropriate type of reliability for survey research and
other questionnaires with a range of responses to choose from.

(2) Validity of an instrument

Validity refers to the extent to which a measurement procedure actually measures what it is
intended to measure. Measurement in social sciences, with very few exceptions, is indirect.
Under such circumstances, researchers are never completely certain that they are measuring the
variable for which they designed their measurement procedure.

One prominent type of evidence of validity is Content related evidence of validity. This refers
to the degree to which an instrument measures the intended content area. I.e. it indicates the
nature of the content that is included within the instrument.

Content related evidence of validity asks questions like: -

 How appropriate and comprehensive is the content?
 Does the content logically get at the intended variables? Are the items related
to what is needed?

Content related evidence of validity is largely a function of how the instrument is developed.

Thus, to determine the content related evidence of validity the list of objectives which guided the
construction of the instrument, and a separate list of items designed specifically to test the
hypothesis are given to two knowledgeable persons in the domain of the content. These
knowledgeable persons are then asked to each independently have a thorough inspection or
critique of the items, and specifically: -
(i) Link each objective with its respective objective.
(ii) Assess the relevance of the items to the content addressed by the objectives.
(iii) To judge if they (knowledgeable persons) believe that the items in the
instrument adequately represent the content or behaviours in the domain of
interest.

The two knowledgeable persons are independently to rate each item on a 4-point scale:
1. Not relevant
2. Somewhat relevant Not applicable

3. Quite relevant
4. Very relevant Applicable
The researcher can then compute the level of agreement on the items between the two
knowledgeable persons as the inter-rater agreement.

Modern thinking holds that there should be a proper quantification of the content related
evidence of validity.

This quantification is called the Content Validity Index (C V I).

Computation of the Content Validity Index is as shown in the following example.

Example: Using an instrument of 10 items

Knowledgeable 1 OR 2 3 OR 4 Items

persons

No. 1 3 7 10

No. 2 1 9 10

Total = 2 4 16 20

Mean (÷ 2) 2 8 10

The ‘Not relevant’ and ‘Somewhat relevant’ choices by the knowledgeable persons are ignored.
Only the ‘Quite relevant’ and ‘Very relevant’ choices are considered in the computation of the
Content Validity Index, as follows:

CVI = 8 ÷ 10 = 0.8

The ideal Content Validity Index is ≥ 0.70

Therefore, in the example above the instrument is valid. It can therefore, be used to collect data.

Interviews
An interview is a conversation between an interviewer and an interviewee with the purpose of
obtaining information from the interviewee. It may be done by face-to-face interaction or by
telephone. Interviewing is a form of questioning characterised by the fact that it employs verbal
questioning as its principal technique of data collection.

When is it necessary to use interviews as a method of data collection?

When: -
 A small number of subjects are involved.
 Sample subjects are accessible.
 Most of the items are ‘open’ and require an extended response with prompts and
probes.
 Everyone is ‘key’ whom the researcher cannot afford to lose.
 The content is sensitive in nature so that trust is a requirement.
 Anonymity is not necessary but confidentiality is important.
 Depth of meaning is essential.
 Research aims and objectives require insight and understanding.

Types of interviews
1. Formal / Structured / Standardised Interviews
Structured interviews are based on a strict procedure and use highly formally structured
schedules, called interview schedules or interview guides. Questions are read to the interviewee
by the interviewer. The number of questions, their sequence and wording are identical for all
respondents with no deviations. This rigid structure determines the operation of the research
instrument and allows no freedom to make adjustments to any of its elements such as content,
wording or order of questions. Strict adherence to the questions and instructions is paramount.
Researchers using the interview method make two assumptions: -
(i) That the questions in the interview schedule are sufficiently comprehensive to
elicit all possible information relevant to the research problem from the
respondents.
(ii) That all questions have been worded in a way clearly and equally understood
by the respondents.

Researchers use the structured interviews to make sure that any variations between responses can
be attributed to the actual differences between respondents and not to variations in the interview.

In this type of interviewing the interviewer is expected to perform like a ‘robot’, acting in a
neutral manner, keeping the same tone of voice, offering the same impression to the respondents,
and using the same style, appearance, prompts, probes, etc., and showing no initiative,
spontaneity or personal interest in the research topic.. The essence of this is to reduce interviewer
bias to a minimum thereby achieving the highest degree of uniformity in procedure.

2. Informal / Unstructured / Unstandardized Interview

This interview has no strict procedures to follow with no restrictions in the wording and order of
questions. The interviewer acts freely, formulating questions as and when required and
employing neutral probing. The structure of the interview is flexible, and the restrictions
minimal, being presented in most cases in form of guides rather than rules. The questions are
open-ended. The researcher has a great deal of freedom to probe various areas and to raise
specific queries during the course of the interview. Not all respondents will find similar or equal
meaning in like-worded questions.

With little or no direction from the interviewer, respondents are encouraged to relate their
experiences, describe what they consider to be important events, provide their meanings of
situations, and reveal their opinions and attitudes as they see fit.

Interviewers must develop, adapt and generate questions and follow-up probes appropriate to
each situation and central theme of the study. Unstructured interviews are at times used to
augment field observations.
3. Semi-structure / Focused / Semi-standardised Interview
This lies somewhere between the structured and unstructured interview, and it contains elements
of both. It involves the implementation of a number of predetermined questions and topics.

Although the interview follows a systematic and consistent order of questions, the interviewer
has the freedom to digress and probe beyond the answers to the standardised questions.
Similarly, respondents are given considerable freedom in expressing their definition of a
situation that is presented to them. Wording of questions is flexible and the level of language
may be adjusted. It is focused on the respondents’ experiences regarding the situations under
study.

Focused interview is meant to focus attention on the given experience of the respondent and its
effects. Under this the interviewer has the freedom to decide the number and sequence of the
questions; and also the freedom to explore reasons and motives.

Nature of interview

For the interview to take place the respondent should know the answer and is willing to respond.
This greatly depends on a number of factors: -
(i) Introductory statements before the actual interview – should make self-introduction,
explain the study and its reasons, give some assurances and plead for acceptance to
be interviewed.
(ii) The way questions are asked – clarity.
- Tone, speed, wording / language, sequence, level of questions.
(iii) Motivation of respondent to ensure willingness to respond.
(iv) Conducive environment for both parties: e.g. disturbing noise / music, presence of
playful or stubborn children.
(v) Appearance / presentation of the interviewer cf. dress, mannerisms, age, gender,
hair style, size, etc.

The researcher should look for the pre-selected / sampled respondent - no substituting. All
responses should be reported verbatim as given by the respondents in the language of research.
Tape recording can be done, with the respondents’ consent, to avoid the problem of slow writing.

Distinctive characteristics of quality interviews

 They use open-ended questions only.
 They are predominantly single / personal interviews.
 Flexible question structure, allowing change of question order, and addition of new ones
where necessary.
  They allow the interviewer more freedom in presenting the questions and adjusting the
interview so that it meets the goals of the study.

Principles of Interviewing

The first step in the interviewing process is getting the respondent to cooperate and to
provide

the desired information. Three factors help in motivating respondents to cooperate: -

(1) Respondents must feel that their interaction with the interviewer will be pleasant and
satisfying.
(2) Respondents need to see the study as being worthwhile. They should feel not only that
the study may benefit them personally, but also that it deals with a significant issue and
that their cooperation is important. Therefore, interviewers should interest the
respondents in the study by pointing out its significance and contribution which the
respondents can make by cooperating.
(3) Barriers to the interview in the minds of respondents need to be overcome. Interviewers
must correct misconceptions of some respondents who may be suspicious of their
intentions. The onus is on the interviewer to explain in a friendly way, the purpose of the
study, the method of selecting the respondents, and the confidential nature of the
interview.
(4) Interviewers should introduce themselves to the respondents and create rapport with
them.

Advantages of interviews
1. Interviews are flexible as they can be adjusted to meet many diverse situations, whereby
questions can even be restructured.
2. They provide in-depth, complete and honest data in contrast to questionnaires, by using
probes.
3. Very sensitive and personal information can be obtained from respondents by honest and
personal interaction between the interviewer and interviewees.
4. Using personal skill, the interviewer can overcome any resistance from respondents.
5. Interviews allow a higher response rate since respondents answer the questions face-to-
face.
6. Interviews are easy to administer as they do not require respondents to be literate or to
handle complex documents or questionnaires.
7. They allow opportunity to observe non-verbal behaviour.
8. The interviewer is able to control the conditions under which questions are asked.
9. The interviewer is able to correct / clear misunderstandings by respondents.
10. The interviewer can control the order of questions.
11. There is an opportunity to record spontaneous answers / responses.
 12. The identity of the respondent is ascertained, and background information can be verified.
13. There is a guarantee that all the questions will be answered.
14. The interviewer can usually control which person(s) will answer the questions.
15. There is control of the date, time and place of the interview, especially when information
must be collected on a specific day and at a specific time and place.
16. More complex questions can be asked since the interviewer can assist in answering them.
17. The language of the interview can be adopted to the ability or educational level of the
interviewee, and as such any misinterpretations of questions can be avoided.
18. Greater permissible length is more possible than in other methods.
19. In standardised / structured interviews, respondents give responses which are easy to
compare.
20. The interviewer can collect supplementary information about the respondent’s personal
characteristics which is often of great value in interpreting results.

Limitations
1. Interviews are costly and time consuming, especially when respondents are widely
dispersed geographically.
2. Interviewing requires high level communication and interpersonal skills which novice
interviewers might not possess.
3. May be affected by interviewer bias.
4. If different interviewers are involved, each one has unique characteristics and differences
in interviewing techniques which may affect respondents’ answers.
5. Lack of anonymity since the interviewer knows the details of the interviewee. This may
make the respondent / interviewee threatened. Some respondents may develop or feel
some form of “examination fever”.
6. There is possibility of response effect whereby the respondent may be over-stimulated
and too eager to please the interviewer, even to the extent of giving imaginary
information just to make the interview interesting. On the other hand the interviewer may
be inclined to ask only questions which tend to support his /her perceived notions.
7. Interviews are only suitable for small samples.
8. The respondent’s perception of, and reactions to, the interviewer can influence responses.
E.g. young people interviewing old people, gender biases, ethnic sentiments, etc.
9. It is less effective than other methods when sensitive issues are discussed. People prefer
to write about sensitive issues rather than discuss them.
10. Standardised interviews overlook different forms of information which would be
obtained by observation or probing, such as non-verbal information.
11. Interviews may take place when the subjects’ environment has been disturbed socially
and physically.
12. Certain types of respondents such as important officials or executives and wealthy
individuals may not be easily approachable, to the extent that the data may prove
inadequate.
13. The organisation required for selecting, training and supervising the field staff is more
complex with formidable problems.
 14. Interviewing at times may introduce systematic errors.
Focus Group Discussions (FGD)
This is a method of data collection in which the researcher moderates a discussion of some
people focusing on a topic of interest. It provides information about perceptions on a specific
topic or subject of interest.

Focus Group Discussions are done on the premise that opinions can never be made in isolation of
one another. One’s opinions can influence the ideas of others. In the end many ideas and
opinions can be generated. The moderator of the FGD keeps the group focused on the topic of
discussion. He / she encourages participants to express their perceptions or opinions by asking
them open-ended questions and telling others to respond to the ideas expressed by other group
members. The goal is to get as many people as possible to express their different perceptions and
not necessarily to have them reach a consensus.

Focus Group Discussions are conducted for a number of reasons:-

(i) To identify issues which are important to the people in respect of the research
topic and research question.
(ii) To cross-check the accuracy and reliability, consistency and validity of other
methods used to collect the same data.
(iii) To get more details from the respondents in case of any unexpected or missing
ideas. That is, to complement / supplement data obtained using other methods.

Main considerations in conducting FGDs:-

1. Group composition
The group should comprise of a homogeneous group of people who are more familiar with one
another. The group size should be reasonable – not too big nor too small. The recommended
ideal size is 06 – 10 participants – but in any case not less than five (05). The more the people the
more the opinions. Reasonably fewer people allow a chance of getting more details from one
another. Too few participants end up dwelling on too few opinions. Members should not be shy
or fear to talk about themselves and others.
 Participants should come from similar background to be comfortable to talk
about themselves. - Peers in terms of age, education, socio-economic status,
etc.
2. Moderator’s characteristics. Cf. Fluency in language of discussion. The role of the
moderator is to encourage members to talk openly, keep them on track, control
dominant characters, involving everyone in the discussion, posing questions, probing,
among others.
3. Venue of Discussion. The FGD should take place in a neutral place free of
distractions / disruptions like noise, rain, mosquitoes, flies, foul smell, food /drink
flavours, or even bias like a church or mosque, etc.
4. Group seating arrangement. Let everyone sit seeing one another, in a semi-circle.
 X X

X X X

X X

X X X

5. Questions.

Welcome members; make self-introductions; explain why they were invited; assure them that
there are no wrong answers, every view is important; emphasise the expression of honest views
an feelings; and assure them of anonymity of their views.

Establish rapport with participants. The first question may be a general and open-ended one
which may be answered by each participant. Other questions can follow systematically. The
moderator may pick a relevant point from the participants’ responses and continue the
discussion.

6. Motivating participants. Pay close attention and show interest in anyone who
contributes. Encourage the quiet ones also to speak – point them out by their
numbers. Reward the contributions with encouraging remarks. Accept all
contributions without making any distinctions. Do not grill participants on their
contributions. Do not dominate the discussion or express personal opinions. Remain
 calm and reserved, and let others talk. Seek different opinions deliberately where
everyone seems to agree.

7. Recording the Discussion. From the outset explain the mode of recording the
discussion and obtain informed consent – by video / tape recording or video
recording, etc. make sure that records remain anonymous. Let the assistant record
other forms of non-verbal communication, such as facial expressions, gestures and
other forms of body language.

8. Analysing the Discussion. Generate lists of all different contributed ideas – reasons,
effects, preferences, and indeed whatever was discussed. Try to estimate how
important an issue seems to be by noting how much time was spent on it or how
many people expressed opinions on it. Get someone else to listen to the audio-taped
record, or read the transcript of the discussion and check whether they agree in your
conclusions.

Advantages of Focus Group Discussions

1. The Focus Group Discussion method is primarily a way of gaining information in
a short period of time about the breadth or variation of opinions, and of
establishing a mechanism of opinion formation.
2. FGDs allow significant points of view to be presented in a real, emotional and
summated form as spontaneous expressions.
3. Group environments encourage participants to express their views and to evaluate
situations, especially when encouraged / motivated by the moderators.

Limitations
1. Group conditions might force people to hide their real opinions, especially if their
views can have effects on their personal life or professional career.
2. Manual recording of the discussions can be problematic.
3. Domination of the discussion by some persons might affect the direction and
outcome of the discussion.
4. Some members might not participate in the discussion.
5. A trend of group to please the moderator might occur for a variety of reasons.
6. Success of the method heavily depends on the qualities of the moderator and the
composition of the group.
7. There may be difficulties with keeping the discussion on track.
8. The group might have reasons to attempt to offer to the moderator a consolidated
front and mislead him / her.
Observation Method
This is a method of data collection which employs vision as its major means of getting data
without asking participants any questions. However, not all observation is scientific. Observation
becomes a scientific tool and the method of data collection for the researcher, when it serves a
formulated research purpose, is systematically planned and recorded, and is subjected to checks
and controls on validity and reliability.

People’s behaviour, activities and other of interest aspects are observed in their natural settings /
environments. This method is particularly suitable in studies which deal with subjects who are
not capable of giving verbal reports of their feelings for one reason or the other. Observation
may be used as the only technique of data collection, or jointly with other methods.

Data collected by observation describes the observed phenomena as they occur in their natural
settings, which eliminates the element of artificiality found in other data collection methods.

Whatever the purpose of the study or observational procedure used, researchers must deal with at
least four major considerations to ensure that the data obtained is systematic and meaningful.
(i) What should be observed?
(ii) When to observe and how to record the observations?
(iii) How the accuracy of observations can be ensured?
(iv) How much inference is required?

Types of Observation Method

1. Structured Observation
This employs formal strictly organised procedure, with a set of well-defined observation
categories. It is organised and planned before the study begins. It is subjected to high levels of
control and differentiation. A controlled observational system is typified by clear and explicit
decisions as to what, how, and when to observe.

Controlled observations can be carried out in the laboratory as well as in the field. In both
settings the researcher wishes to infer the causal effect by maximizing control over threats to
internal and external validity of the study, while employing one of the experimental research
designs and systematically recording the observations. This type of observation is also called
scientific observation because it is systematically planned and executed; it is related to certain
goals and subjected to tests and controls.

Furthermore, structured observation is one in which the observation is characterised by a careful

definition of the units to be observed, the style of recording the observed information,
standardised conditions of observation and the selection of pertinent data of observation.
Structured observation is appropriate in descriptive studies.
 2. Unstructured Observation
This is loosely organised and the process of observation is largely left up to the observer to
define. In an exploratory study the observational procedure is most likely to be relatively
unstructured.

3. Simulation Observation
Here, the researcher creates a situation to be observed and tells the study subjects what activities
they are to be engaged in. it allows the researcher to observe the required behaviour though not
in the natural setting. Its limitation is that the setting is not natural and the exhibited behaviour
may not be what would occur ordinarily in the natural setting (it can be artificial). Cf. the
Hawthorne effect.

4. Participant Observation
Here observers become members of the group under study either overtly or covertly. The
researcher is a participant observer.

5. Non-participant Observation
The researcher does not become one of the participants under study. The observer / researcher is
detached from the study group with or without their knowledge. Detachment of the researcher /
observer is expected to reduce bias and allow him / her to develop and provide an accurate
description of events. This method is useful for describing events which people do not describe
clearly either because they do them habitually without much thought, or they do not often these
actions into words.

An ideal narrative record includes details of events as they unfold in their natural sequence.
When narrations or events are being recorded continuously, times can be recorded against the
events to later know the duration of the event or the interval between events. The observer may
also record the consequences if they are observable.

Collection of Observational Data

In participant observation data collection starts after entry to the study setting has been gained.
Notes are taken in ways that best suit the circumstances. When the framework is qualitative, data
collection and analysis take place simultaneously. Observation focuses on research units over a
set period of time.

Data collection takes different forms based on the focus of the observation:
(i) Continuous observation

In this case observation is continuous and does recording of occurrences for the entire duration
of the event. E.g. Observer may record the activity say between 7.00 am and 8.00 am daily for a
period of four weeks.
 (ii) Time-point observation
Data collection may focus at a specific point in time (time-point). The observer / researcher is
interested in what happens at that point in time. E.g. Recording what happens in a school at
6.00 pm every day for three months or a whole school term.

(iii) Time-interval observation

Here data collection is focused on what happens between two set times.
(iv) Event observation
This relates to behaviours which occur because other behaviours / events also occur. E.g. The
observer may be interested in seeing how students behave when the teacher enters the
examination room carrying examination papers.

Advantages / strengths / merits of observation

1. It provides information when other methods are not effective.
2. Behaviours and other observable events can be described in details
maintaining the natural sequence and consequences of the events.
3. It approaches reality in its natural structure, and studies events as they evolve.
4. Data obtained through observation is more amenable to statistical analysis by
coding each unit of observation under a category and analysing the frequency
counts for each category.
E.G.

Laughin Smiling Clapping Standing Nodding

g

//// ///// /// // ///////

Categories can then be compared, and statistical relations between them be computed and
inferred.

5. It is less time consuming. Many events and big numbers of subjects can be
observed simultaneously.
6. It can offer data where subjects are unable or unwilling to cooperate or to give
information, i.e. it is independent of subjects’ willingness to respond.
7. It generates first-hand information without having to rely on reports of others.
8. It is relatively inexpensive.
9. Subjective bias is eliminated to some extent if observation is done accurately.
 10. Information obtained under this method relates to what is currently happening.
It is not complicated by either the past behaviour or future intentions or
attitudes.

Disadvantages / limitations of observation

1. Subjects may be uncomfortable with a stranger in their midst recording their
actions.
2. Since the non-participant observer does not ask the subjects about their
actions or their understanding of events, the observed events become only
clues to their meaning.
3. Important events or statements may not be recalled because of observing and
recording many things at the same time.
4. Non-verbal reactions may be easily missed. E.g. Facial expressions of
acceptance, rejection or scorn. Such expressions may influence subsequent
events / actions. If they are missed the interpretation of the subsequent
observations may be affected or may be inconsistent.
5. It cannot be employed when extensive events or sizable samples are studied.
6. It cannot provide information about past, future or unpredictable events.
7. It is amenable to observer bias, selective perception and memory.
8. It does not offer controls measures for bias, attitudes and opinions of the
observer. When researchers are sole observers, they unconsciously tend to see
what they expect to see. Their own values feelings and attitudes based upon
past experiences may distort their observations.
9. It cannot offer information about frequency of behaviour.
10. It cannot study opinions or attitudes directly.
11. It is inadequate when studying certain phenomena such as sexual behaviour or
domestic violence.
12. It is a laborious and time-consuming method.

Document Analysis
This sometimes referred as content analysis. It consists of analysing the contents of documentary
materials such as books, journals, research reports, magazines, newspapers, conference papers,
and others.

In terms their closeness to the event they describe, one can distinguish between:
 Contemporary documents, namely those compiled at the time the events took
place;
 Retrospective documents, which have been produced after the event took
place;
  Primary documents, which are sources compiled by eye witnesses of the
described events;
 Secondary documents, being sources derived from primary documents.

Types of documents
1) Public documents, including census statistics, statistical Year Books,
departmental / organisational records, mass media and literature, etc.
2) Personal documents, such as life histories, diaries, memoranda
autobiographies, etc.
3) Administrative documents, such as proposals, policies, circulars,
memoranda, progress reports, minutes of meeting, and other internal
documents.
4) Research Reports, etc.

Procedure of documentary research

Documentary analysis is accomplished in a number of successive steps, as summarised here

below:

Stage 1: Identification of relevant documents

The choice of documents depends on availability, accessibility, relevance the personal interest of
the researcher. The use of documents may constitute a part of a larger review of literature.

Stage 2: Data collection

The process of data collection also depends on availability, accessibility and the method of
analysis, among others. For example if description in qualitative research is the purpose of the
study, reading and taking of notes may be sufficient. The most important thing is to ensure that
the targeted information is extracted from the consulted documents. This ought to be done
systematically, under the identified themes in conformity with the study objectives.

Stage 3: Data analysis

This is a very diverse step. The type of evaluation depends on the nature of the documents, the
methodology chosen and the purpose of the study. This diversity becomes even more complex
when very many documents are considered. Generally, data will be related to the assumptions
made before or during the study, and assessed with regard to the degree to which these
assumptions are valid. Whether the findings can be generalised in statistical or analytical terms
depends on the evidence obtained, and on the type of methodology used.

Stage 4: Interpretation of the findings

Interpretation of the findings will be made in the light of the research topic, and will depend on
the nature and purpose of the study, the type of methodology used. While In some cases
inductive generalisations can be made, in other cases analytical generalisations will be produced.

Advantages of document analysis

1. Retrospectivity:

Documentary methods enable researchers to study past events and issues.

2. Quick and easy accessibility: especially in the case of many available
documents. It has been made much easier by the introduction of
electronic media, increasing availability of diverse digital sources and
data banks and sophisticated computer programs and packages.
3. Spontaneity: In many cases writers produce documents when they have
not been requested to do so by researchers. This reduces researcher bias
significantly.
4. Low cost:
5. Sole source: Often documents are the only source of information, as in
the case of studying past events.
6. High quality of information.
7. Non-reactivity: The method itself and the act of measurement do not
affect the results.

Limitations
1. Documents are not necessarily representative of their kind and thus they
do not allow generalisations, as in the case of personal documents like
autobiographies.
2. Some of the documents are not always readily and easily accessible, as in
the case of personal documents like letters, and documents containing
highly sensitive information.
3. Some documents are not complete or up to date.
4. The reliability of some documents is questionable.
5. Documents present methodological problems such as coding problems.
6. Documents are biased since they represent the views of authors.
 RESEARCH PROPOSAL

What is a Research Proposal?

A Research Proposal is a written plan which identifies the major elements of the study. It spells
out in detail what the researcher intends to do, why and how it will be done. It is a formal way to
communicate ideas about a proposed study so as to obtain approval to conduct the study, or to
seek funding from funding agencies. It enables other people (e.g. colleagues, lecturers, or
reviewers) to think through the proposed study so that they can give advice / input on how to
improve the study before it is conducted.

COMPONENTS OF A RESEARCH PROPOSAL

1. The Title / Front Page
2. Preliminary pages: Declaration, Approval, Table of Contents
3. Chapter One: Introduction
 Introduction / Overview
 Background to the Study, consisting of: -Brief description of global, regional and
local state of affairs with regard to the topic of the study, justification for the study
 Theoretical, conceptual and contextual backgrounds.
 Statement of the problem
 Objectives of the study
 Purpose of the study (General aim)
 Scope of the study (Content, population, geographical and time)
 Area of the study (Brief description)
 Significance of the study (For whom and how, will the study be useful?)

 Explanation of terms
 Abbreviations / acronyms Optional
 Conclusion

4. Chapter Two: Review of Related Literature

  Introduction
 Theoretical Review
 Review of Literature (By objectives)
 Conceptual Framework (and its explanation)
 Hypotheses / Research questions for the study. May be placed after Objectives.
 Conclusion

5. Chapter Three: Methodology

 Introduction
 Research design
 Sample selection
 Sample size
 Research instrument(s)
 Instrument reliability
 Instrument validity
 Data collection method(s)
 Data collection procedure
 Data analysis
 Presentation of findings (According to hypotheses / research questions)
 Ethical considerations
 Conclusion

N.B. (1): Give justification for all choices in this Chapter.

6. References: -Strictly only those cited in the text

7. Appendices:

 Letter of Authority

 Budget estimates

 Time frame

 Instrument(s)

 Map(s) of Uganda showing the location of the District / Municipality / City, in

which the study will be carried out. Another one showing the actual locality.

N.B. (2): Strictly adhere to the Academic Writing guidelines of APA 7th Edition.
 8. References (Only those cited in the text)

More Appendices (As necessary)

ACADEMIC WRITING FOR A RESEARCH PROPOSAL

APA 7th Edition: American Psychological Association

The Research Proposal contains only three Chapters, namely:

Chapter One: Introduction

Chapter Two: Review of Related Literature

Chapter Three: Methodology

 A research proposal is developed and written according to specified institutional
guidelines which ought to be adhered to strictly.
 The content of the research proposal is written in the future tense except for the
Literature Review Section which cites and describes the results of previous studies.
 Personal pronouns (like I, me, we, my, our …) are avoided. The researcher-to-be writes
impersonally.
 References are written in the alphabetical order of the first author’s surname. The
surnames are followed by the initials of the other names.
 Ensure consistent flow of ideas throughout the proposal – connectivity of the ideas from
Chapter One through to Chapter Three is a very crucial feature of the research proposal.
 Be precise in description by using very clear and explicit language. Be economical with
words by avoiding verbosity, irrelevance, vagueness, ambiguity, colloquial or slang
words.
 Use correct language in terms of grammar, spellings, punctuations, capitalisations, etc.
 Work hand in hand amicably with the supervisor.
 Preparing a good research proposal requires a lot of creative thinking with thoughtful
anticipation of successfully executing it later. At the proposal stage the researcher is
expected to form a full clear picture of the whole study in his / her head right from the
beginning.

REFERENCES
Ary D.L.C & Razarieh. (1990). An introduction to research in Education, Harcourt, Brace and
Javanrich. New York

Bell, J (1999). Doing Your Research Project, Oxford University Press Buckingham,
Philadelphia.

Best. J.W. and Kalin J.V. (1993). Research in Education. Prentice Hall of India PVT Ltd. New
Delhi.

Blakie, N. (2010). Designing Social Research: The Logic of Anticipation. 2nd Ed. Polity Press.
Cambridge, UK.

Bryman.A (2000) Social Research Methods Oxford University Press. New York

Creswell, J. (2003). Research design: Qualitative, quantitative &mixed methods approaches.

2nd edition. London: Sage Publishers

Denzin N.K and Lincoln Y.S (ed.). (2000). Handbook of Qualitative Research (2nd Ed.)
Thousand Oaks, California.

Kakooza T. (2002) Research: An Introduction to Research methodology. National Adult

Education Association, Kampala.

Kothari, C.R. (2004). Research Methodology, New Age International, New Delhi Kumar (1999)
Research Methodology Sage Publications,London

Mugenda, O. & Mugenda, A. (1999). Research Methods: Quantitative and Qualitative

Approaches. ACTS Press: Nairobi, Kenya.

Sarantakos, S. (1998). Social Research 2nd Ed Macmillan Press Ltd, London

Tashakkori, A. & Teddie, C. (1998). Mixed Methodology: Combining qualitative and

quantitative approaches. Applied Social Research Methods Series, 46. Thousand Oaks, C4: Sage
Publications, Inc.