Research Design Explained: Descriptive, Correlational, Experimental, and More
You’ve finally reached the methodology section of your thesis proposal. You know you need to name your research design, but suddenly every term sounds the same. Descriptive? Correlational? Experimental? Quasi-what?
You’re not alone. This is where many Kenyan students get stuck, often copying and pasting definitions from the internet without truly understanding what they mean. The result? A weak methodology section that confuses your supervisor and weakens your entire proposal.
Here’s the truth: Your research design is the backbone of your entire study. Choose the wrong one, and your findings may not actually answer your research questions. Choose the right one, and everything else—data collection, analysis, conclusions—falls into place.
This guide will explain every major research design in simple terms, with real examples relevant to Kenyan students. By the end, you’ll know exactly which design fits your study and how to justify it in your proposal.
And if you’d rather have an expert help you craft your entire methodology chapter, Proposal Writers Kenya is here to help. Our experienced academic writers specialize in guiding Kenyan students through every stage of their thesis proposal.
What Is a Research Design?
Think of a research design as a blueprint for a house. Before any construction begins, the architect draws a detailed plan showing where every room, door, and window will go. The blueprint guides every decision made during construction.
Your research design works exactly the same way. It’s your overall plan for answering your research questions. It tells your reader:
What kind of data you will collect
How you will collect it
Who you will collect it from
How you will analyze it
What conclusions you can (and cannot) draw
Many students confuse research design with research methods. Here’s the distinction: Your design is the overall strategy; your methods are the specific tools (like questionnaires or interviews). The design comes first—it determines which methods are appropriate.
The Three Main Categories of Research Design
All research designs fall into three broad families. Each serves a different purpose and answers a different kind of question.
| Design Type | Purpose | Key Question | Control Level |
|---|---|---|---|
| Descriptive | To describe | What is happening? | Low |
| Correlational | To predict | What goes together? | Low to Medium |
| Experimental | To explain cause | What causes what? | High |
Let’s explore each one in detail.
Descriptive Research Design
What Is Descriptive Research?
Descriptive research does exactly what its name suggests: it describes. You observe and measure variables as they naturally occur, without manipulating anything. Your goal is to paint an accurate picture of a situation, population, or phenomenon.
Imagine you want to know how many hours per week University of Nairobi students spend studying. You would survey a sample of students, calculate the average, and report your finding. That’s descriptive research. You’re not trying to explain why some study more than others—just describing what exists.
Types of Descriptive Designs
Cross-sectional design: Data collected at one point in time. This is the most common design for undergraduate theses in Kenya.
Longitudinal design: Data collected from the same subjects over multiple time points (e.g., tracking student performance over a full academic year).
Normative design: Comparing your results against established norms or standards.
When to Use Descriptive Design
Choose descriptive design when:
You want to describe “what is” happening
Your research questions start with “What,” “How many,” “How often,” or “To what extent”
You have no intention of manipulating variables or establishing cause-and-effect
Examples for Kenyan Students
“What is the average study time among third-year students at Kenyatta University?”
“How many students use digital libraries at Moi University?”
“What are the common challenges faced by small-scale farmers in Kiambu County?”
“To what extent do final-year students at JKUAT experience research anxiety?”
Strengths and Limitations
| Strengths | Limitations |
|---|---|
| Simple to understand and execute | Cannot explain why something happens |
| Flexible and adaptable to many topics | Prone to bias if sampling is poor |
| Provides real-world, practical data | Cannot establish causation |
| Excellent for exploratory research | May lack depth |
Correlational Research Design
What Is Correlational Research?
Correlational research measures the relationship between two or more variables. You’re asking: When one variable changes, does the other variable change in a predictable way?
Important note: You do not manipulate anything. You simply measure both variables as they naturally exist, then calculate the strength and direction of their relationship.
Understanding Correlation Coefficients
The correlation coefficient (usually written as “r”) is a number between -1.0 and +1.0 that tells you two things:
Strength: How close the number is to 1.0 (or -1.0). Closer to 1.0 means a stronger relationship.
Direction: Positive (+) or negative (-).
| Correlation Value | Meaning | Example |
|---|---|---|
| +0.80 | Strong positive relationship | More study time → higher exam scores |
| +0.20 | Weak positive relationship | More study time → slightly higher scores |
| -0.70 | Strong negative relationship | More screen time → lower sleep quality |
| 0.00 | No relationship | Shoe size and exam scores (unrelated) |
The Cardinal Rule: Correlation Does Not Equal Causation
This is the single most important rule in all of research methods. Just because two variables are related does NOT mean one causes the other.
Consider this classic example: Ice cream sales and drowning incidents are positively correlated. As ice cream sales increase, drowning incidents also increase. Does eating ice cream cause drowning? Of course not. The third variable is summer heat—people buy more ice cream AND swim more when it’s hot.
As a student, you must never claim causation from correlational data. Your supervisor will spot this mistake immediately.
When to Use Correlational Design
Choose correlational design when:
You want to predict one variable from another
You cannot ethically or practically manipulate variables
Your research questions include words like “relationship,” “association,” “prediction,” or “link”
Examples for Kenyan Students
“What is the relationship between study time and exam performance among undergraduate students?”
“Does daily smartphone usage predict sleep quality among university students in Nairobi?”
“Relationship between lecturer feedback frequency and student motivation at Strathmore University”
“Is there an association between family income and access to online learning resources?”
Strengths and Limitations
| Strengths | Limitations |
|---|---|
| Can study variables you cannot manipulate | Cannot prove causation |
| Good for making predictions | Directionality problem (which variable influences which?) |
| Often more practical and ethical than experiments | Possible third-variable problem |
| Can analyze many variables at once | Requires larger sample sizes |
Experimental Research Design
What Is Experimental Research?
Experimental research is the gold standard for establishing causation. In a true experiment, you actively manipulate one variable (the independent variable) to see its effect on another variable (the dependent variable).
If you want to prove that X causes Y, you need an experiment.
Core Elements of True Experiments
For a design to be called a “true experiment,” it must have three essential elements:
Manipulation: You control the independent variable (e.g., you decide which students get a new teaching method and which get the traditional method).
Control group: A group that does NOT receive the treatment, providing a baseline for comparison.
Random assignment: Participants are randomly placed into experimental or control groups, ensuring the groups are equivalent at the start.
Types of True Experimental Designs
Post-test only control group design: Measure both groups AFTER the treatment only.
Pre-test post-test control group design: Measure both groups BEFORE and AFTER the treatment.
Solomon four-group design: Combines the above to control for testing effects.
When to Use Experimental Design
Choose experimental design when:
You want to prove that X causes Y
You can actually manipulate the independent variable
You have the resources for control groups and random assignment
Your research questions include words like “effect,” “impact,” “causes,” or “influence”
Examples for Kenyan Students
“Does using mnemonic techniques improve memory retention among Form 4 students in Kiambu?”
“Effect of gamification on student engagement in online learning at the University of Nairobi”
“Impact of financial literacy training on savings behavior among university students”
“Does a 10-minute mindfulness exercise reduce exam anxiety among final-year students?”
Strengths and Limitations
| Strengths | Limitations |
|---|---|
| Can establish causation with confidence | Often artificial (lab settings) |
| High internal validity | Many variables cannot be ethically manipulated |
| Replicable and rigorous | Expensive and time-consuming |
| Considered the “gold standard” | May not generalize to real-world settings |
Quasi-Experimental Research Design
What Is Quasi-Experimental Research?
Quasi-experimental design looks like an experiment but lacks one crucial element: random assignment. You still manipulate an independent variable and have a control group, but participants are assigned to groups in non-random ways (e.g., using existing classrooms, schools, or communities).
Types of Quasi-Experimental Designs
Non-equivalent control group design: You use intact groups (e.g., Classroom A and Classroom B) rather than randomly assigning individuals.
Time series design: You measure the same group multiple times before and after a treatment.
Single-subject design: You track one individual or a small group over time.
When to Use Quasi-Experimental Design
Choose quasi-experimental design when:
You cannot randomly assign participants (e.g., studying students in different schools)
You are evaluating a real-world intervention
True experiments are impossible due to practical or ethical constraints
Examples for Kenyan Students
“Effect of a new teaching method on students in School A compared to School B” (you cannot randomly assign students across schools)
“Impact of a youth mentorship program on participants in one community” (no control group exists)
“Effect of a financial literacy workshop on employee productivity in one department”
Strengths and Limitations
| Strengths | Limitations |
|---|---|
| More practical and ethical than true experiments | Weaker causal claims than true experiments |
| Works in real-world settings | Risk of selection bias |
| Often the only option for field research | Groups may differ in important ways |
| Easier to implement with existing groups | Cannot completely rule out alternative explanations |
Other Research Designs
Causal-Comparative (Ex Post Facto) Design
You study cause-and-effect without any manipulation because the cause has already happened. The phrase “ex post facto” means “after the fact.”
Example: “Do students with working mothers perform differently than students with stay-at-home mothers?” You cannot assign children to different mother types—you simply compare existing groups.
Case Study Design
An in-depth investigation of a single person, group, event, or situation. You gather rich, detailed, contextual data from multiple sources.
Example: “A case study of student retention strategies at one Kenyan university” or “The journey of one first-generation university student at Moi University.”
Survey Design
Collecting data from a sample using questionnaires or interviews. Surveys can be descriptive (simply reporting what exists), correlational (examining relationships), or comparative (comparing groups). This is the most common design in Kenyan undergraduate theses.
Action Research Design
Practitioner research aimed at solving immediate, practical problems. Very common in education and business theses.
Example: “Improving student participation in my classroom using peer feedback strategies” or “Reducing customer wait times at my workplace using a new queuing system.”
How to Choose the Right Research Design for Your Study
Follow these four steps to choose your design:
Step 1: Write down your research question clearly.
Step 2: Ask yourself three questions:
Do I want to describe, predict, or explain cause?
Can I manipulate my independent variable?
Can I randomly assign participants?
Step 3: Match your answers to the design using this guide:
| Your Goal | Can You Manipulate? | Can You Randomly Assign? | Recommended Design |
|---|---|---|---|
| Describe what exists | N/A | N/A | Descriptive |
| Predict relationships | No | No | Correlational |
| Explain cause | Yes | Yes | True Experimental |
| Explain cause | Yes | No | Quasi-Experimental |
| Explain cause (past events) | No | No | Causal-Comparative |
| Deep understanding of one case | N/A | N/A | Case Study |
Step 4: Consider your practical constraints (time, budget, ethics, access). Be honest about what you can realistically accomplish.
Common Mistakes Kenyan Students Make With Research Design
Mistake 1: Calling your study “experimental” when you didn’t use random assignment. That’s quasi-experimental.
Mistake 2: Claiming causation from correlational data. Your supervisor will catch this immediately.
Mistake 3: Choosing a design that doesn’t match your research question. Design follows question—never the reverse.
Mistake 4: Describing your design but never justifying why you chose it.
Mistake 5: Calling every small study a “case study” when you’re actually doing a descriptive survey.
How to Write the Research Design Section of Your Proposal
Here is a simple template you can adapt:
This study adopted a [name of design] research design. This design was appropriate because [reason based on your research question]. Unlike [alternative design], this design allows the researcher to [specific advantage]. The study did not use [alternative design] because [reason].
Example for a descriptive study:
This study adopted a descriptive cross-sectional survey design. This design was appropriate because the research aimed to describe the current state of student engagement with e-learning platforms at the University of Nairobi. Unlike a correlational design, the study was not seeking to predict relationships between variables but simply to describe existing conditions.
Frequently Asked Questions
Which research design is easiest for an undergraduate thesis?
Descriptive cross-sectional surveys are the most common and manageable for undergraduates.
Can I use more than one design?
Yes, mixed-methods studies often combine qualitative and quantitative elements. However, for an undergraduate thesis, keep it simple.
Do I need an experimental design for my MBA thesis?
Rarely. Most MBA theses use descriptive or correlational designs unless you’re testing a specific business intervention.
What design should I use if I cannot access a control group?
Use descriptive or correlational design. Quasi-experimental requires a comparison group but not necessarily random assignment.
Is survey research always descriptive?
No. Surveys can be descriptive, correlational, or even experimental if you embed an experiment within the survey.
Conclusion
Your research design is the backbone of your entire thesis proposal. Choose it carefully, justify it clearly, and be honest about its limitations. Remember the simple rule: Your research question drives your design—never the other way around.
If you’re a Kenyan student feeling overwhelmed by your methodology chapter or any part of your thesis proposal, you don’t have to struggle alone. At Proposal Writers Kenya , we specialize in helping students craft clear, well-structured proposals that impress supervisors. From research design to data analysis, our experienced writers are here to support you every step of the way.
Visit Proposal Writers Kenya today for a free quote and let’s get your thesis proposal approved.