Research Design Explained: Descriptive, Correlational, Experimental, and More
Picture this: You’re sitting at your desk, staring at your thesis proposal template. You’ve written your topic, your problem statement feels solid, and your objectives are finally clear. Then you hit the section labeled “Research Design.” Your confidence evaporates.
What is a research design anyway? Is descriptive the same as correlational? Can I do an experiment for my education thesis? What does my supervisor actually expect to see here?
If these questions sound familiar, you’re not alone. Research design is one of the most confusing topics for Kenyan students writing their thesis proposals. But here’s the good news: once you understand the basics, choosing the right design becomes straightforward.
This guide will explain every major research design type in plain language—descriptive, correlational, experimental, quasi-experimental, causal-comparative, and the main qualitative designs. You’ll learn what each design does, when to use it, and how to write about it in your proposal. By the end, you’ll know exactly which design fits your research and how to justify it to your supervisor.
And if you still feel stuck after reading? At Proposal Writers Kenya, we help students like you craft complete, supervisor-ready proposals—including perfectly written methodology chapters. But first, let’s get you educated.
What Is Research Design? (A Simple Definition)
Let’s start with a clear definition. Research design is the overall blueprint or roadmap for your study. It answers three fundamental questions:
What kind of data do you need? (Numbers? Words? Both?)
How will you collect that data? (Surveys? Interviews? Experiments?)
How will you analyze it? (Statistics? Thematic analysis?)
Think of research design as the architectural plan for a house. Before a single brick is laid, the architect draws up a plan showing where every room, door, and window will go. Your research design does the same thing for your study—it shows your supervisor exactly how you will answer your research questions.
Important distinction: Research design is NOT the same as research methodology. Methodology is the broader strategy; design is the specific framework within that strategy. Your methodology might be quantitative, but your design could be descriptive or correlational.
The Major Types of Research Design: An Overview
Before we dive deep into each design, here’s a quick reference table. This will help you see the big picture before we break everything down.
| Design Type | Purpose | Data Type | Control Level |
|---|---|---|---|
| Descriptive | To describe what exists | Quantitative or Qualitative | Low |
| Correlational | To examine relationships | Quantitative | Low |
| Causal-Comparative | To explore causes after the fact | Quantitative | Medium |
| Quasi-Experimental | To test interventions without full control | Quantitative | Medium-High |
| Experimental | To test cause-effect with full control | Quantitative | Very High |
| Case Study | To deeply understand a single case | Qualitative | Low |
| Phenomenological | To understand lived experiences | Qualitative | Low |
Now let’s explore each design in detail.
Design Type 1: Descriptive Research Design
Definition and Key Characteristics
Descriptive research design is exactly what it sounds like—it describes a population, situation, or phenomenon as it naturally exists. The researcher does not manipulate anything. They simply observe, measure, and document.
Key characteristics:
No manipulation of variables
No control group
Answers questions like “what,” “who,” “where,” “how many,” and “how often”
When to Use Descriptive Design
You should consider descriptive design when:
You want to describe the characteristics of a specific population
You want to measure the prevalence of attitudes, behaviors, or conditions
You are doing exploratory research with no need to test relationships or causes
Subtypes of Descriptive Design
Cross-sectional descriptive: Data collected at one point in time (most common in Kenyan undergraduate theses)
Longitudinal descriptive: Data collected from the same population over multiple time points
Normative descriptive: Comparing your findings to established norms or standards
Real Example from Kenya
“A descriptive survey of financial literacy levels among university students in Nairobi County”
In this study, the researcher would simply measure how much students know about saving, budgeting, and investing. They would not try to change anyone’s behavior or test any relationships—just describe what exists.
Advantages and Disadvantages
| Advantages | Disadvantages |
|---|---|
| Simple to understand and implement | Cannot explain why something happens |
| Good for large populations | Cannot establish cause-and-effect |
| Relatively quick and inexpensive | Prone to self-report bias |
| Provides baseline data for future research | Limited depth of insight |
How to Write This in Your Proposal
*”This study will adopt a descriptive cross-sectional survey design. According to Creswell (2018), descriptive design is appropriate for studies that aim to describe the characteristics of a population as they exist naturally without manipulation. This design is suitable for this study because the objective is to describe the financial literacy levels and saving behaviors of university students in Nairobi. Data will be collected at one point in time using a structured questionnaire.”*
Design Type 2: Correlational Research Design
Definition and Key Characteristics
Correlational research design measures the statistical relationship between two or more variables. The key question is: When one variable changes, what happens to the other?
Key characteristics:
No manipulation of variables (like descriptive)
Measures direction (positive or negative) and strength of relationships
Cannot prove causation (this is the most important point to remember)
The Golden Rule: Correlation ≠ Causation
This is the single most important concept in correlational research. Just because two things are related does NOT mean one causes the other.
The ice cream and drowning example: As ice cream sales increase, drowning deaths also increase. Are ice cream sales causing drownings? No. Both are caused by a third variable—hot weather. More people eat ice cream in summer AND more people swim (and drown) in summer.
When you use a correlational design, you can NEVER conclude that one variable causes another. You can only say they are “related” or “associated.”
When to Use Correlational Design
You want to identify relationships between variables
You cannot manipulate variables for ethical or practical reasons
You want to make predictions about one variable based on another
Your research questions ask about “relationship,” “association,” or “prediction”
Subtypes of Correlational Design
Bivariate correlation: Examining the relationship between exactly two variables
Multiple regression: Using several predictor variables to predict one outcome variable
Path analysis: Examining complex relationships among many variables
Real Example from Kenya
“The relationship between study habits, sleep patterns, and academic performance among secondary school students in Kiambu County”
The researcher would measure each student’s study habits (hours studied, consistency), sleep patterns (hours slept, bedtime consistency), and academic performance (exam scores). Then they would calculate whether students with good study habits also tend to have good grades, and whether sleep patterns predict performance.
How to Write This in Your Proposal
“A correlational research design will be employed in this study. This design is used to determine the statistical relationship between two or more variables without manipulating them (Fraenkel et al., 2019). The design is appropriate because this study aims to examine the relationship between study habits, sleep patterns, and academic performance. The researcher will measure all variables using standardized instruments and compute Pearson correlation coefficients to determine the strength and direction of these relationships.”
Design Type 3: Experimental Research Design
Definition and Key Characteristics
Experimental research design is the gold standard for proving cause and effect. In a true experiment, the researcher manipulates an independent variable and observes its effect on a dependent variable.
Key characteristics:
Independent variable is manipulated by the researcher
Random assignment of participants to groups (this is what makes it a TRUE experiment)
A control group for comparison
Can establish causation (unlike correlational design)
When to Use Experimental Design
You want to prove that one thing causes another
You can ethically manipulate the independent variable
You have the resources to randomly assign participants to groups
Your setting allows for controlled conditions
True Experimental Subtypes
1. Posttest-Only Control Group Design:
Random assignment → Intervention → Measure
Simple, but no baseline to compare against
2. Pretest-Posttest Control Group Design:
Random assignment → Pretest → Intervention → Posttest
More powerful because you can measure change
3. Solomon Four-Group Design:
Combines both approaches to control for the effect of taking a pretest
Complex but very rigorous
Real Example from Kenya
“The effect of gamified learning on student motivation in mathematics: A randomized controlled trial in Nairobi primary schools”
The researcher would randomly assign students to two groups. The experimental group uses gamified learning apps; the control group uses traditional worksheets. After several weeks, the researcher measures both groups’ motivation levels. If the experimental group is significantly more motivated, the researcher can conclude that gamified learning caused the increase.
Advantages and Disadvantages
| Advantages | Disadvantages |
|---|---|
| Can establish cause-and-effect | Often impractical in real-world settings |
| High internal validity | Can be expensive and time-consuming |
| Results are easy to interpret | Many variables cannot be ethically manipulated |
| Widely respected in academic circles | Random assignment is often impossible in schools |
Design Type 4: Quasi-Experimental Research Design
Definition and Key Characteristics
Quasi-experimental research design is the compromise between true experiments and correlational studies. The researcher still manipulates an independent variable, but WITHOUT random assignment.
Key characteristics:
Independent variable is manipulated
NO random assignment (uses existing groups like classrooms)
Still tries to establish causation (but weaker claim than true experiments)
When to Use Quasi-Experimental Design
You cannot randomly assign participants (e.g., entire classrooms)
Ethical constraints prevent random assignment (e.g., you can’t randomly assign smoking)
You are working in real-world settings where control is limited
Quasi-Experimental Subtypes
1. Non-Equivalent Control Group Design:
Existing groups → Pretest → Intervention → Posttest
The most common quasi-experimental design in Kenyan theses
2. Time Series Design:
Multiple measurements before and after intervention
Helps show that changes happened after the intervention
3. Single Subject Design:
One participant measured repeatedly over time
Common in clinical psychology and special education
Real Example from Kenya
“The impact of a financial literacy program on savings behavior: A quasi-experimental study of university students in Mombasa (using existing classes as groups)”
The researcher cannot randomly assign individual students because they are already organised into classes. So one class becomes the experimental group (receives financial literacy training) and another similar class becomes the control group (no training). The researcher measures savings behavior before and after. This is strong evidence, but not as ironclad as a true experiment.
How to Write This in Your Proposal
*”This study will use a quasi-experimental non-equivalent control group design. Due to the school setting, random assignment of individual students is not feasible; therefore, existing intact classes will serve as the experimental and control groups. Both groups will complete a pretest, followed by a 6-week gamified learning intervention for the experimental group. Both groups will then complete a posttest to measure changes in mathematics motivation.”*
Design Type 5: Causal-Comparative (Ex Post Facto) Design
Definition and Key Characteristics
Causal-comparative design (also called ex post facto, which means “after the fact”) explores possible causes of existing differences between groups.
Key characteristics:
No manipulation (the “cause” has already happened)
The researcher identifies existing groups
Compares groups on a dependent variable
Difference from Correlational
Correlational design looks at relationships between continuous variables. Causal-comparative design compares distinct groups.
Difference from Experimental
Experimental design has manipulation and (ideally) random assignment. Causal-comparative has neither—the groups already exist.
When to Use Causal-Comparative Design
You cannot ethically assign participants to groups (e.g., smoking vs. non-smoking)
The “cause” happened in the past
You are exploring possible reasons for existing differences
Real Example from Kenya
“Differences in career readiness between students who participated in internship programs and those who did not at Kenyan technical universities”
The researcher identifies two existing groups: students who did internships and students who did not. Then they compare career readiness scores. If the internship group scores higher, the researcher can suggest that internships might cause higher career readiness—but cannot prove it because other factors may differ between the groups.
How to Choose the Right Research Design for Your Thesis
Here is a simple 5-question decision framework:
| Question | Your Answer | Suggested Design |
|---|---|---|
| Are you describing what exists? | Yes → | Descriptive |
| Are you measuring relationships without manipulation? | Yes → | Correlational |
| Are you testing whether an intervention causes change? | Yes → | Experimental or Quasi-Experimental |
| Are you exploring causes of existing differences? | Yes → | Causal-Comparative |
| Are you deeply understanding a single case or experience? | Yes → | Qualitative (Case Study, Phenomenology) |
Remember: Your research design must align with your research questions. A correlational design cannot answer a causal question. An experimental design is overkill for a descriptive question. Match the tool to the job.
Common Mistakes Kenyan Students Make
| Mistake | Why It’s a Problem |
|---|---|
| Claiming experimental design without random assignment | That’s quasi-experimental, not true experimental |
| Using correlational design but claiming causation | This will cost you marks and confuse your findings |
| Choosing a design that doesn’t match research questions | Your proposal will be logically inconsistent |
| Describing the design but not justifying why you chose it | Your supervisor will wonder if you understand your choice |
| Confusing research design with data collection methods | These are separate sections in your proposal |
Conclusion
Research design doesn’t have to be intimidating. At its core, it’s simply the blueprint that answers three questions: What kind of data do you need, how will you collect it, and how will you analyze it?
Remember the key distinctions:
Descriptive tells you what exists
Correlational tells you what is related
Experimental tells you what causes what
Quasi-experimental gives you strong evidence without random assignment
Causal-comparative explores possible causes after the fact
Most undergraduate and master’s theses in Kenya use descriptive or correlational designs, and that’s perfectly fine. You don’t need a complex experiment to produce valuable research. What matters is that your design matches your questions and that you can justify your choice.
Now you have the knowledge to choose the right design and write about it confidently in your proposal.
Feeling stuck anyway? Writing the methodology chapter—especially justifying your research design—is one of the hardest parts of any thesis proposal. At Proposal Writers Kenya, our expert academic writers can help you craft a clear, rigorous methodology chapter that impresses your supervisor. We understand Kenyan university requirements and can work with any design type. Get your free quote today and let us help you move forward with confidence.