LECTURE FOUR - RESEARCH PROCESS.ppt which

Research Methodology

1. LECTURE FOUR –
THE RESEARCH
PROCESS

2. 1.Formulating the Research Problem
2.Extensive Literature Review
3.Developing the Objectives
4.Preparing the Research Design including Sample Design
5.Collecting the Data
6.Analysis of Data
7.Generalization and Interpretation
8.Preparation of the report or Presentation of Results-Formal write ups of
conclusions reached.

3. THE RESEARCH PROCESS
Problem
1,2,3
Generalization (7
& 8)
Data Analysis (6)
Data Collection
(5)
Measurement
Design (4)
Research
Design (4)

4. 1. Selecting and Definition of the
Research Problem
This stage involves:
•Selecting the topic of research (Main RQ).
•Acquiring knowledge on current theories related to
the problem identified. This is also referred to as
preliminary review of the related literature (to
identify variables)
•Formulation of the research problem
•Formulating research questions, objectives and
hypotheses

5. 2. Determining the Research Design

A design is the ‘blueprint’ that enables the
researcher to come up with solutions to the
research questions and also guides him/her in the
various stages of the research (Nachmias and
Nachmias, 1996).

A research design enables a researcher to
determine the methods of sample selection,
instruments to use and methods of analyzing data

6. 3. Measurement Design
To carry out a research investigation, researchers
must gather data with which to answer research
questions or test hypothesis. The tools used to
collect this data are referred to as research
instruments. In this stage, a researcher determines
the research instruments to use to collect data.
Researchers must also provide evidence that the
research instruments will yield data that is both
valid and reliable.

7. VALIDITY AND RELIABILITY

8. To carry out a research investigation, researchers must
gather data with which to answer research questions
or test hypothesis. The tools used to collect this data
are referred to as research instruments. Researchers
must therefore provide evidence that the research
instruments will yield data that is both valid and
reliable.
Introduction
09/26/24Wambugu L.N.

9. Validity refers to the appropriateness, meaningfulness
and usefulness of the inferences a researcher makes.
Validity pertains to the results of an instrument not the
instrument.
There are three categories of validity:
Content, Criterion-related and Construct
Validity
09/26/24Wambugu L.N.

10. This refers to the degree to which an instrument
measures the subject matter and behaviors the
researcher wishes to measure. It is also the
representativeness or sampling adequacy of the
content of a measuring instrument. There are two
types of content validity;

face validity and

sampling validity.
Content validity
09/26/24Wambugu L.N.

11. Face validity is concerned with the extent
to which an instrument measures what it appears to
measure according to the researcher’s subjective
assessment while sampling validity is the degree to
which a measure adequately samples the subject
matter under consideration.
Content validity yields a logical judgment as to
whether the instrument covers what it is supposed to
cover.
Contd…
09/26/24Wambugu L.N.

12. To ensure content validity, the researcher
should ask a panel of experts (at least
3)in the field to critically examine the
items for their representativeness or Peer
Review
Contd…
09/26/24Wambugu L.N.

13. This attempts to measure the extent to which a
measure relates / correlates with an external
criterion. There are two types of criterion-related
validity;

concurrent validity and

predictive validity.
Criterion-related Validity
(Empirical Validity)
09/26/24Wambugu L.N.

14. 
Concurrent validity deals with measures that can
be administered at the same time as the measure to
be validated. The concurrent validity for an
instrument is determined by administering both the
instruments to a group of respondents, then finding
the correlation between the two.
Contd…
09/26/24Wambugu L.N.

15. Predictive validity refers to the extent to which a
measuring instrument predicts future outcomes that
are logically related to the construct for instance
do pupils who score high in KCPE also score high in
KCSE? Or Mock and KCSE?

Unlike content validity, criterion-related validity
yields a numeric value, correlation coefficient,
sometimes also called a validity coefficient.
Contd…
09/26/24Wambugu L.N.

16. An instrument has construct validity if its relationship to
other information corresponds well with some
theory.
A theory is a logical explanation of phenomena after
it has been observed for a long time.
Theories can be used to determine construct validity
of an instrument.
Construct validity
09/26/24Wambugu L.N.

17. For instance, you can measure attitude by:

Answering questions

Attending classes

Participating in class
Triangulation increases construct validity.
Contd…
09/26/24Wambugu L.N.

18. Reliability is concerned with the question ‘does the
instrument yield the same or similar results
consistently?’ Reliability is the degree of consistency
that the instrument or tool demonstrates on repeat
trials. There are two main methods of estimating
reliability;

repeated measurements and

internal consistency.
Both involve the procedure of correlating two sets of
scores.
Reliability
09/26/24Wambugu L.N.

19. The higher the correlation coefficient, the greater the
reliability. Correlation coefficient varies on a scale
of 0.00 (indicating total unreliability) and 1.00
(indicating perfect reliability). 0.8 -0.9 indicates
high reliability, 0.6 -0.8 indicates acceptable
reliability value while below 0.5 is unacceptable.
Contd…
09/26/24Wambugu L.N.

20. This is concerned with the ability of the instrument to
measure the same thing at different times. Three
methods are used:
a) Test-Retest method
b) Alternative forms method
c) Parallel forms/ equivalent forms method
Repeated measurement
09/26/24Wambugu L.N.

21. a) Test-Retest method is estimated by measuring
individuals on the same instrument on different
occasions and correlating the scores obtained by
the same person’s on the two administration.
However it is affected by time between the test
and retest, the mood of the respondent during
either of the test and also some respondents may
memorize answers especially if the second
administration is given too soon. Pearson Product
Moment Correlation Coefficient may be used to
determine the coefficient
Contd…
09/26/24Wambugu L.N.

22. b) Alternative forms method – this is similar to test-
re-test except that the questions on the second
instrument are renumbered to create an alternative
form of the first instrument
c) Parallel forms/ equivalent forms method – in this
method, two alternative instruments designed as
equivalent as possible is administered to the same
group of respondents during the same time period.
Contd…
09/26/24Wambugu L.N.

23. This form of reliability indicates the degree of
homogeneity among items in an instrument. It is
estimated by determining the degree to which each
item in a scale correlates with each other item.
Internal consistency is based on a single
administration of a measure. There are three types
of this reliability:
Internal Consistency
09/26/24Wambugu L.N.

24. a) Split half reliability- a measure is split into two
parts and each of them is treated as a separate
measure. Each part is scored accordingly and the
scores correlated. The spearman-Brown Prophesy
formula is used to estimate this reliability
b) The Kuder-Richardson Method – this is used for
knowledge questions
c) Cronbach-Alpha – this is used with instruments in
which there is no right or wrong answer to each item,
such as an attitude scale
Contd…
09/26/24Wambugu L.N.

25. 4. Data Collection
Once the researcher determines the tools to use and
constructs them, the fourth stage is to collect data
using the tools. There are various methods of
collecting data including conducting a one-on-one
interview, mailing the questionnaires to the
respondents, conducting focus group discussions etc.

26. 5. Analysis and Interpretation of Data
Data collected from the field must be analyzed,
presented and interpreted to make it meaningful
to the audiences of the research work. Methods of
data analysis depends on the type of data
collected be it nominal, ordinal, interval or ratio
data. There are two statistics that a researcher
may use; descriptive and inferential statistics.

27. 6. Writing a Research Report and
Generalization
It is during this stage that a researcher draws
conclusions about the target population based on
the data analyzed and interpreted.
The validity of generalization may be threatened by
two kinds of statistical errors:

28. Contd…

Bias error which occurs when the sample is not
representative

Random error which occurs when the sampling
design is flawed.