Tant a Un ive r sity

Fac u lty of De n tis try

RESEARCH
METHODOLOGY
A Systematic Guide to Scientific Inquiry

10 Dec, 2025
 AGENDA OVERVIEW
01 Definition
02 Steps
03 Hypothesis

04 Variables

05 Control Groups

06 Bias
07 Sample & Sampling Techniques
 A structured and scientific approach to solving
a research problem.

It Guides through:

Planning & Design : Creating a Detailed

Plan for the study.
Data Collection & Analysis : Gathering
Information and using Statistical Methods to
reveal patterns.
Interpretation: Drawing Conclusions to
produce Valid and Reliable findings.
Research Methodology
Essentially, it is the Framework that guides a
researcher's choices and ensures a study’s
is the Roadmap for Scientific Integrity.

Discovery !
Systematic Steps of the Research Process
Slide 5 Slide 6

Slide 7 Slide 8

Slide 9

1. Research 2. Literature 3. Research 4. Research 5. Population &

Problem Review Q uestion/ Design S ample
Hypothesis
Slide 13
Slide 12 Slide 11 Slide 10

9. Report Writing 8. I nterpretation 7. Data Analys is 6. D ata

Collection
1 . Identifying the Research Problem

Identify & Develop your Topic

Clearly Define the problem or issue

that needs investigation and explain
its importance
2 . Literature Review

Find Background Information

Review previous studies, in books,

encyclopedias, and databases to
understand what is already known
and identify gaps in the existing
knowledge related to your topic.
3 . Formulating Research Question/Hypothesis

Developing specific queries

Develop a clear research Question

or testable Hypothesis that guides
the study.
4 . Choosing Research Design

Create detailed plan for the study

Including
1. Type of Research
(experimental, observational, etc.),
2. Variables involved,
3. Control Mechanisms.
5 . Selecting population & sample

Defining the target group

Choose a Representative Sample using
suitable Sampling Techniques.
6 . Data Collection

Use Reliable Tools

Gather the necessary information or data
using Specific Tools like surveys, interviews,
or experiments.
7 . Data Analysis

Process the collected data

Often using Statistical Methods, to reveal
patterns, relationships and insights.
8 . Interpretation

Drawing meaning from the analysis

Interpret the results, Explain their meaning,
and Determine whether they support or
contradict the hypothesis.
9 . Report Writing

Report the Findings

Present the study’s results in a structured
report, including methods, analysis,
discussion and conclusions.
Hypothesis
It is a specific, testable prediction
about the relationship between
variables included in the study.
 Types of Hypotheses
Function Directionality

Null (H0) Alternative Directional Non-Directional

No significant (H1/Ha) Specifies the States a relationship
relationship Significant expected direction exists without
nor relationship or of the relationship predicting its specific
difference exists
difference between variables direction (e.g., "Cats
between two between
variables. (e.g., "Girls perform and dogs differ in the
variables. better than boys"). amount of affection
they express").
 Types of Hypotheses
Complexity Relationship

Complex Simple Associative Causal

Predicts a relationship Predicts a relationship the association cause-and-effect
involving two or more between a single between variables, interaction
independent variables independent variable but it does not between variables.
and/or two or more and a single determine which indicates that
dependent variables dependent variable variable is causing manipulating one
(e.g., "Higher poverty (e.g., "Lower fertilizer the changes in the variable directly
other. causes changes in
and higher illiteracy use leads to lower
rates in a society lead agricultural another variable
to higher crime productivity").
rates").
2 2 1 1
 Control Group
A baseline group in an experiment that does not receive the experimental treatment
or intervention. It provides a standard for comparison to the experimental group,
helping to isolate the effect of the independent variable & establish causality.
Types of Control Groups

Positive Control GP
Receives a treatment that is Already Known to produce the Expected Effect.

Negative (No-treatment)
Control GP
Receives No Intervention At All (No Effect Is Expected).

Placebo Control GP

Receives a Fake Treatment (e.g., a sugar pill) that has No Therapeutic Effect
(improvement based on expectation).
 Definition
Variable • Any characteristic und er stud y o r investig atio n,
related to d ifferent elements( sub ject , ob ject or
event).

Statistically T y pes Role

• Dependent vs independent
• Quantitative Vs qualitative
• Confounding variables.
• Quantitative (Continuous
• Predictor variables.
vs discrete)
• Input variables.
• Qualitative (ordinal Vs
• Outcome variables.
nominal)
• Latent variables.
• Composite variables.
• Preceding variables.
Variables acc. to Statistics

QUANTITATIVE Variables
• Numerical/ Metric (can be measured
numerically).
• Example: Height (inches/cm), Blood
pressure (mmHg), Age (years), Weight
(Kg), Number of beds in hospital.

QUALITATIVE Variables
• Categorial/ Yield observations acc. to
some characteristic or quality.
• Example: Gender, Marital status,
Educational level, Occupation.
Variables acc. to Statistics
Continuous Variables
• Can take any value within a given
range; values are infinite and
uncountable.
• Unlimited numbers in any interval.
• No gaps in the real values.
• Example: Height, Weight, Time
Discrete Variables
• Have separate, distinct, and countable
values; gaps exist between possible
values.
• Specific points along its scale.
• Counting is often used with it.
• Examples: Number of cars, shoe size
(e.g., 7, 7.5, 8, but not 7.2), outcomes
of dice rolls (1, 2, 3, 4, 5, 6).
Variables acc. to Role
Independent Variables
• Cause/Manipulated factor/Predictor.
• Controlled or intentionally varied by
the researcher.
• Plotted on the horizontal x-axis
• Ex: amount of water used in watering
plants.
Dependent Variables
• Effect/Outcome/Response.
• Observed and measured (of 1ry
interest); changes in response to the
independent variable.
• Not controlled by the researcher.
• Plotted on the vertical y-axis.
• Ex: plant’s growth.
Variables acc. to Role
Confounding Variables
• Unmeasured third variable.
• If not accounted for
-< Influences both independent variable
& dependent variable
–< false/misleading +ve/-ve correlation.
• Ex: Coffee drinking and heart disease:
Independent variable: Coffee consumption.
Dependent variable: Heart disease rates.
Confounding variable: Smoking, if patients
are smoking.
• Solution: Properly addressing & isolating
them strengthens the Credibility & Validity
of research outcomes ( clarifying the direct
relationship between the dependent and
independent variables ).
Variables acc. to Role
Predictor Variables
• Used to predict/estimate the value of
another variable
(outcome/dependent).
• Does not necessarily imply causality
with the dependent variable.
• Generally observed as it naturally
occurs; not actively controlled by the
researcher.
• All independent variables are
predictors, but not all predictors are
independent variables.
Variables acc. to Role
Input Variables
• any data/parameter, provided to
produce an outcome or output.
• A synonym for the independent variable or
predictor variable.

Outcome Variables
• A common synonym for the dependent
variable.
Variables acc. to Role
Latent Variables
• Cannot be measured directly; its value
is inferred indirectly.
• Uses observable variables to estimate
its presence.
• Represents abstract, theoretical
concepts (e.g., intelligence,
socioeconomic status) that lack a
single perfect measurement.
Variables acc. to Role
Composite Variables
• Created by combining two or more
separate variables into a single score
(e.g., summation/averaging).
• Measures complex concepts (e.g.,
"socioeconomic status").
• Simplifies analysis into one
manageable measure.
• Example: Socioeconomic Status,
combining income, education, and
occupation.
Variables acc. to Role
Preceding Variables
• Occurs earlier in time or in a causal
sequence than another variable.
• Mostly, it’s a synonym for independent
variable.
BIAS
• Systematic Error in the design, conduct,
or analysis of a study that distorts the
findings.
• Unlike random error (which is due to
chance), bias pushes results in a specific
direction, often leading to incorrect
conclusions about the relationship
between variables.
 Categories of Bias
Bias Type Phase Key Characteristic Example Mitigation Strategy
A. Design/Selection Bias
Surveying only volunteers, who may Random sampling;
1. Selection Selection Sample ≠ Population be more motivated than the average Defined inclusion
person. criteria
Excluding non-English speakers from a
Systematically removing specific Defining inclusive
2. Exclusion Design health study, missing cultural health
groups from the study. eligibility criteria
factors.
Researcher may
3. Occurs when randomization is (consciously/unconsciously) choose to
Design Randomization
Allocation inadequate or broken. allocate certain patients to the desired
group.
4. Differences in care between study
One group gets extra attention or
Performanc Design groups (not related to the Blinding
better care.
e intervention).
 Categories of Bias
Bias Type Phase Key Characteristic Example Mitigation Strategy
B. Data Collection & Measurement Bias
Data Memory errors; Participants provide Use real-time data
Patients struggling to remember
1. Recall Collectio inaccurate data due to imperfect (diaries/objective
exactly what they ate two weeks ago.
n memory. records)
Data Researcher subconsciously influences
[Link] A doctor smiling more when a patient Blinding: (Double &
Collectio participants data to fit their
nter Bias) reports improvement on a new drug. triple-blind studies)
n hypothesis.
Anonymizing
3. Social Data
Participants answer in a way they Under-reporting alcohol consumption surveys; indirect
Desirability Collectio
think makes them look good. or over-reporting charitable giving. questioning
Bias n
techniques.
Validating survey
4. Data A thermometer that is consistently 1°
The tool used to measure is flawed, questions &
Instrument Collectio off; A survey question that asks, "Is X
uncalibrated, or leading. Regularly calibrating
Bias n terrible?"
equipment.
Data Knowing who is in the experimental
5. When outcome assessment differs Blinding &
Collectio group leads to more careful
Detection between groups. Randomization
n assessment.
 Categories of Bias
Bias Type Phase Key Characteristic Example Mitigation Strategy
C. Analysis & Publication Bias
Analyzing only successful
startups to find
Focusing only on "successes" or those Analyzing both successes
1. "common traits" while
Analysis who remain, ignoring those who failed and failures; tracking
Survivorship ignoring failed ones with
or dropped out. drop-outs.
the same traits.

Dismissing data points Pre-registering the

2. Interpreting data in a way that confirms
that don't fit the theory analysis plan; Having a
Confirmation Analysis pre-existing beliefs while ignoring
as "errors" while keeping third party analyze the
Bias contradictions.
those that do. data blindly.
An outside variable (confounder) distorts Smoking while studying
3. Randomization;
Analysis the association between the studied relationship between
Confounding Multivariate analysis
variables. coffee and heart attack
4. Publication Publishing only the 2
The tendency for journals to publish
(The File successful studies out of Pre-registering studies;
Reporting positive results, while negative results
Drawer 10 makes a failing drug Publishing null results
(go unpublished.
Problem) look effective
 Keys to Mitigate Bias

Randomization Blinding Pre-registration Triangulation

• Randomly assign • Single-blinding • Publicly state your • Use Multiple

participants to (patient is blinded) hypothesis and Methods (e.g.,
control and • Double blinding analysis plan mixing interviews
treatment groups (neither the before you collect with surveys) to
to average out participant nor the data to prevent verify findings.
unknown researcher knows "p-hacking"
confounders who is in the control (manipulating data
group). until you find a
• Triple blinding significant result).
(patient,
experimenter & data
analysis party).
 Population vs. sample

• Population is • Sample is the

the entire group specific group
that you want to of individuals
draw conclusions that you will
about. collect data
from.
 Probability Sampling
• Random selection Sampling
• Highly representative Techniques
• Quantitative research Carefully selecting subset of
individuals that is representative
of the group as a whole -< to draw
valid conclusions from results.
Non-Probability Sampling
• Non-random selection

• Cheaper & easily collect

data.
• Higher risk of Bias
 Probability Sampling Non-Probability
Sampling
1. Simple random 1. Convenience sampling
sampling 2. Voluntary response
2. Systematic sampling sampling
3. Stratified sampling 3. Purposive sampling
4. Cluster sampling 4. Snowball sampling
5. Quota sampling
Probability/Random Sampling
1. Simple random sampling
• Every population member has an
Equal Chance of selection.
• Sampling frame should include the
whole population.
• Choose participants randomly
(e.g., random number generator).
• Most suitable for homogenous
population, as representativeness
of sample is uncertain.
Probability/Random Sampling
2. Systematic sampling
• Similar to simple random
sampling
but Easier
• list everyone, assign numbers,
then select individuals at fixed
intervals (e.g., every 10th
person).
• First selected individual is
randomly selected.
Probability/Random Sampling
3. Stratified sampling
• Divide the population into
subgroups (strata) by relevant
characteristics, then sample
from each to ensure all groups
are well represented.
• More suitable for heterogenous
population.
Probability/Random Sampling
4. Cluster sampling
• Divide the population into
clusters that each would mirror
the whole sample,

Then
Randomly select entire clusters
instead of individuals.
• Mostly applied in very large
population.
Non - Probability/ Non - Random Sampling

1. Convenience sampling
• Includes the Easiest-to-reach
individuals.
• It’s Quick and Cheap,
BUT
• Not Representative and highly
prone to sampling and
selection bias.
Non - Probability/ Non - Random Sampling

2. Voluntary response sampling

• Similar to convenience
sampling,
BUT
• people Volunteer themselves
(e.g., responding to an online
survey), making it easy but
biased.
Non - Probability/ Non - Random Sampling

3. Purposive sampling
• The researcher uses their
expertise to choose a sample
Most Relevant to the study.
• Common in qualitative
research; it focuses on gaining
detailed insights rather than
generalizing statistically.
• Always state your inclusion and
exclusion criteria.
Non - Probability/ Non - Random Sampling

4. Snowball sampling
• Snowball sampling recruits
hard-to-reach participants
through existing ones.
• Its Representativeness is
uncertain.
Non - Probability/ Non - Random Sampling

5. Quota Sampling
• Sample characteristics are
predetermined.
• Selects a set number of units
from subgroups (quotas) that
you formed based on specific
characteristics.
• Allows the researcher to
control the sample composition.
THANK YOU