Experimental Study Design
Design and Key Considerations
Overview of Experimental Studies
§ Definition of experimental studies
§ Role in hypothesis testing
§ Intervention and outcome relationships
§ Experimental studies involve researchers actively
applying an intervention.
Types of Epidemiological Studies
• Observational vs. experimental
Observational studies:Studies in which the researcher
does not intervene but observes and analyzes natural
variations in exposures and outcomes.
Key Characteristics:
§ No manipulation of variables
§ Subjects are not randomly assigned
§ Used when experiments are unethical or impractical
Types:
– Cross-sectional studies (e.g., a health survey)
– Case-control studies (e.g., lung cancer in smokers
vs. non-smokers)
– Cohort studies (e.g., following exposed vs.
unexposed over time)
Example:
A researcher studies whether people who drink sugary drinks daily
are more likely to develop diabetes, without changing their habits
What is an Experimental Design?
Studies where the researcher actively assigns an intervention (e.g., a
treatment or exposure) and observes the outcomes.
Research design that involves
-manipulation of variables
-Control over confounding factors
-Randomization and control groups
This design tests cause-effect relationships.
ü Researchers manipulate variables and observe outcomes.
ü Example: Giving a vaccine to test its protection against disease.
Basic Features of Experimental Studies
Intervention (treatment vs control)
-Intervention refers to a deliberate action or treatment introduced by
the researcher to study its effect on a specific outcome.
-It is the core component that distinguishes experimental studies from
observational ones
Key Features of an Intervention:
Controlled: Applied to one group (intervention
group) and not to another (control group)-Placebo
Standardized: Delivered uniformly to ensure
consistency
Ethically approved: Must be justified and safe for
participants
Randomization
-is the process of assigning participants to different study
groups (e.g., intervention or control) purely by chance, rather
than by choice or any systematic method
Purpose of Randomization:
§ To ensure comparability between groups
§ To eliminate selection bias
§ To distribute known and unknown confounding variables equally
across groups
§ To increase the validity of statistical conclusions
TYPES OF RANDOMIZATION
Blinding
• Blinding is a technique used in experimental
research to prevent participants, researchers, or
both from knowing which intervention a
participant receives.
• It helps reduce bias in data collection, outcome
assessment, and reporting.
Why Is Blinding Important?
Without blinding:
§ Participants may alter their behavior if they
know their group.
§ Researchers may unintentionally influence
measurements.
§ Outcome assessors may interpret data
differently based on expectations.
Types of blinding
Examples of blinding
üVolunteers get either the vaccine or a placebo.
üNurses administering it and participants do
not know who got what.
üOnly after data analysis is complete is the
code broken
NOTE
INTENTION OF EXPERIMENTS ARE USNURED THROUGH VALIDITY
o Validity refers to the accuracy and truthfulness of the conclusions
drawn from a research study.
o It determines whether the study truly measures what it intends to
measure and whether the findings can be trusted and generalized.
Internal Validity
Definition: The extent to which the results of a study are attributable to the
intervention and not to other factors (confounders or biases).
External Validity (Generalizability)
Definition: The extent to which the study results can be applied to populations,
settings, or conditions beyond those studied.
Types of Experimental Studies
§ Randomized Controlled Trials (RCTs): Subjects randomly
assigned to intervention or control.
§ Field Trials: Applied to individuals in a community
§ Community Trials: Entire communities receive interventions.
Randomized Controlled Trials (RCTs)
It is an experimental study design in which participants are randomly
assigned to either an intervention group (receives the treatment) or a
control group (receives standard care or placebo), to assess the
effectiveness or safety of an intervention
Purpose of an RCT:
a. To determine cause-and-effect relationships between an
intervention and outcome
b. To eliminate selection bias through randomization
c. To provide high-quality evidence for decision-making in Phamacy
• .
Key features
• Gold standard in experimental research.
• Ensures comparability between groups.
• Controls for known and unknown confounders.
• Used in drug testing, interventions, behavioral
studies
Phases of RCTs
Phase I: Assess safety and dosage in small groups.
Phase II: Test effectiveness and side effects.
Phase III: Compare with current treatments in larger
groups.
Phase IV: Monitor long-term effects after approval.
Field Trials
• Test interventions on healthy individuals.
• Aim: prevent disease onset (e.g., vaccines).
• Require long follow-up and large sample sizes.
• Examples: oral polio vaccine trials.
Community Trials
• Entire communities are study units.
• Focus on public health measures (e.g., sanitation,
health education).
• Randomization may be by cluster or community.
• Longer duration and complex logistics.
Quasi-Experimental Studies
• Lack randomization, but use control groups.
• Used when RCT is unethical or impractical.
• Includes before-and-after, time series designs.
• Prone to selection bias.
Steps in Designing an Experimental
Study
1. Formulate hypothesis and objectives.
2. Define population and eligibility criteria.
3. Design intervention and control conditions.
4. Select outcomes and follow-up period.
5. Plan data collection and analysis.
Ethical Considerations
a) Informed consent must be obtained.
b) Ethical approval by IRB or Ethics Committee.
c) Participants’ rights and confidentiality respected.
d) Risks must be justified by potential benefits.
Sample Size Determination
a. Critical to ensure study power.
[Link] on effect size, alpha, beta.
c. Too small: underpowered; Too large: not
feasible.
be cautious
In research, we typically test:
Null Hypothesis (H₀): There is no effect or difference.
Alternative Hypothesis (H₁ or Hₐ): There is an effect
or difference.
When we conduct a statistical test, we risk making
two types of errors:
types of errors
• A Type I error occurs when the null hypothesis is
true, but we incorrectly reject it.
• A new malaria drug is not better than the
current one,but the study concludes it is better
→ Type I error
Type II Error
Type II error occurs when the null hypothesis is
false, but we fail to reject it.
Inclusion and Exclusion Criteria
• Defines who can and cannot participate.
• Ensures target population relevance.
• Helps manage confounders and improve safety.
• Should be specific and justified.
Quality Assurance and Control
• Train all research staff thoroughly.
• Pilot testing of tools before study.
• Supervision and spot checks during data collection.
• Regular audits and data verification.
Data Management
a. Design secure and accessible databases.
b. Implement data entry protocols and cleaning.
c. Backup data regularly and ensure confidentiality.
d. Use of statistical software for analysis.
Outcome Measurement
a. Primary: Main effect of interest.
b. Secondary: Additional related outcomes.
c. Ensure outcomes are valid, measurable, relevant.
d. Minimize missing or subjective data.
Limitations of Experimental Designs
a. May not reflect real-world settings (low external
validity).
b. Expensive and time-consuming.
c. Ethical challenges in withholding treatment.
d. Dropouts and non-adherence can affect results.
Advantages of Experimental Designs
• Strong evidence for causality.
• Control over variables enhances validity.
• Minimizes bias and confounding.
• Foundation for policy and practice decisions.
Summary and Conclusion
• Experimental designs are rigorous and valuable.
• Proper planning and ethics are key.
• Statistical analysis must match the design.
• Results inform practice, policy, and future research.