Validity in Research: How to Ensure Your Findings Are Trustworthy
Imagine spending six months collecting data, analyzing results, and writing your thesis. You submit your work to your supervisor. Their feedback arrives: “How do you know your findings are actually true? Have you addressed validity?”
Your heart sinks. You are not sure what validity means, let alone how to defend it.
This scenario plays out every semester in Kenyan universities. Students pour their energy into data collection but neglect the fundamental question: Does your research actually measure what you claim to measure?
Validity is the answer to that question. It is the difference between trustworthy research that earns approval and questionable findings that leave your supervisor unconvinced.
In this guide, I will explain exactly what validity means, walk you through the four types of validity, and show you practical strategies to establish each one in your own research. By the end, you will know how to write a methodology chapter that defends your findings with confidence.
Struggling with your research methodology or validity section? At Proposal Writers Kenya, we help students design rigorous research that stands up to supervisor scrutiny. Get expert help with your thesis proposal today.
What Is Validity in Research?
Validity is the degree to which your research measures what it claims to measure. In simple terms, validity asks: Are your conclusions actually true?
Consider a bathroom scale. If you step on it and it reads 70 kilograms, you want to know that 70 kilograms is your actual weight. If the scale consistently reads 65 kilograms when you actually weigh 70, the scale is reliable (it gives the same number every time) but not valid (that number is wrong). A valid scale gives you the true measurement.
The same principle applies to your research. A valid study produces conclusions that genuinely reflect reality.
Validity vs. Reliability: A Critical Distinction
Many students confuse validity and reliability. Here is the difference:
Reliability: Your instrument produces consistent results
Validity: Your instrument produces correct results
A thermometer that reads 2 degrees too high every single time is reliable (consistent) but not valid (incorrect). A broken clock that always shows the same wrong time is reliable but not valid. For your research to be trustworthy, you need both reliability and validity.
The Four Types of Validity
Researchers generally recognize four types of validity. Each answers a different question about your research.
| Type | Question It Answers |
|---|---|
| Conclusion Validity | Is there actually a relationship between my variables? |
| Internal Validity | Did one variable actually cause the other? |
| Construct Validity | Does my instrument measure the right concept? |
| External Validity | Can I generalize my findings to other people or settings? |
Let us explore each one in detail.
Type 1: Conclusion Validity
Definition: Conclusion validity is the extent to which your conclusions about relationships between variables are reasonable.
When it matters: This applies to quantitative studies, especially those using correlation or regression analysis.
Example: You conclude that study time is related to exam scores. Conclusion validity asks whether that relationship actually exists in your data or whether you are seeing a statistical illusion.
Threats to Conclusion Validity
The biggest threat to conclusion validity is low statistical power. This happens when your sample size is too small to detect a real relationship. If you study only 10 students, you might miss a real relationship between study time and exam scores simply because you do not have enough data.
Other threats include:
Unreliable measurement instruments
Violating assumptions of statistical tests
Too much variation within your sample
How to Enhance Conclusion Validity
Ensure adequate sample size: Use sample size formulas or tables (Yamane, Krejcie & Morgan)
Use reliable instruments: Pilot test before full data collection
Check statistical assumptions: Do not just run tests blindly
Reduce measurement error: Train research assistants, standardize procedures
Type 2: Internal Validity
Definition: Internal validity is the extent to which you can conclude that changes in one variable caused changes in another variable.
When it matters: This is critical for experimental and quasi-experimental studies that aim to make causal claims.
Example: You implement a new teaching method and then observe improved test scores. Internal validity asks whether the teaching method caused the improvement or whether something else is responsible.
Classic Threats to Internal Validity
| Threat | Explanation | Example |
|---|---|---|
| History | Outside events affect outcomes | A university policy change during your study |
| Maturation | Participants naturally change over time | Students get better simply because they are older |
| Testing | Taking a pretest affects the posttest | Students remember answers from the pretest |
| Instrumentation | The measurement tool changes | Different observers used at different times |
| Selection | Groups are not equivalent | Comparing two different schools without randomization |
| Mortality | Participants drop out | The weakest students leave the study |
How to Enhance Internal Validity
Use control groups that do not receive the treatment
Randomly assign participants to groups
Use pretest-posttest designs
Standardize all procedures
Blind participants and researchers when possible
Important note for Kenyan students: Most undergraduate and master’s research cannot claim strong internal validity because random assignment is often impossible. Be honest about this limitation rather than overstating your conclusions.
Type 3: Construct Validity
Definition: Construct validity is the extent to which your instrument measures the theoretical concept it claims to measure.
When it matters: This applies to any study using questionnaires, tests, or scales.
Example: You develop a “job satisfaction” questionnaire. Construct validity asks whether your questions actually measure job satisfaction or whether they accidentally measure something else like mood, workplace relationships, or general happiness.
How to Establish Construct Validity
There are several ways to demonstrate that your instrument measures the intended construct.
Face Validity (Simplest): Ask experts to review your questions. Do the questions appear to measure what you want? If three lecturers in your field agree your questions make sense, you have established face validity.
Content Validity: Ensure your instrument covers all aspects of the construct. If you are measuring “study habits,” your questionnaire should address time management, note-taking, revision strategies, and concentration—not just one of these dimensions.
Criterion-Related Validity: Show that your instrument correlates with an existing gold standard. For example, if you develop a new depression screening tool, you would show that scores on your tool correlate highly with scores on an established depression scale.
Convergent and Discriminant Validity:
Convergent: Your measure correlates with other measures of the same construct
Discriminant: Your measure does NOT correlate with measures of different constructs
Practical Steps for Kenyan Students
Have your supervisor review your questionnaire (face validity)
Base your questionnaire on a thorough literature review (content validity)
Pilot test and ask participants if questions make sense
Report all validity evidence in your methodology chapter
Type 4: External Validity
Definition: External validity is the extent to which your findings can be generalized to other populations, settings, and times.
When it matters: Any study that aims to make claims beyond the specific sample studied.
Example: You study 200 students at the University of Nairobi. External validity asks whether your findings apply to students at Moi University, to students in Tanzania, or to students five years from now.
Threats to External Validity
Non-representative sample: Your sample differs from the broader population
Selection bias: You used a convenience sample (as most students do)
Specific setting: Your findings may only apply to your university
Time-bound findings: Your findings may only apply to this specific year or context
How to Enhance External Validity
Use random sampling whenever possible (though this is rare in student research)
Describe your sample in great detail so readers can assess generalization themselves
Be honest about limitations in your methodology chapter
Acknowledge that your findings may not generalize broadly
The Kenyan Context
A study conducted only in Nairobi may not apply to rural Kenya. A study of private university students may not apply to public university students. This does not make your research invalid. It simply means you must honestly describe your sample and acknowledge limitations.
Validity in Qualitative Research
Validity looks different in qualitative studies. Qualitative researchers often prefer the term “trustworthiness” instead of validity, using criteria developed by Lincoln and Guba.
| Quantitative Term | Qualitative Alternative | What It Means |
|---|---|---|
| Internal validity | Credibility | Are your findings believable? |
| External validity | Transferability | Can findings apply elsewhere? |
| Reliability | Dependability | Are findings consistent? |
| Objectivity | Confirmability | Are findings free from researcher bias? |
Strategies for Qualitative Trustworthiness
Member checking: Ask participants to review and confirm your interpretations
Triangulation: Use multiple data sources (interviews, documents, observations)
Peer debriefing: Discuss your findings with colleagues
Prolonged engagement: Spend sufficient time in the field
Thick description: Provide rich, detailed descriptions of your context
Reflexivity: Keep a journal acknowledging your own biases and assumptions
If you are conducting interviews or focus groups in Kenya, these strategies will strengthen your qualitative research significantly.
Common Validity Mistakes Kenyan Students Make
| Mistake | Why It Is a Problem | How to Fix |
|---|---|---|
| Claiming causality without experimental design | Supervisors will reject your conclusions | Use cautious language like “associated with” instead of “causes” |
| Ignoring validity in qualitative studies | Findings lack credibility | Address trustworthiness explicitly |
| Using instruments without validation | May not measure what you think | Pilot test and report validity evidence |
| Convenience sampling without acknowledging limitations | Overstates generalizability | Honestly discuss limitations in your chapter |
| Confusing reliability with validity | A scale can be reliable but not valid | Address both separately in your methodology |
How to Write About Validity in Your Thesis
Here is a sample paragraph for a quantitative methodology chapter:
“To ensure validity, this study addressed the four types of validity. Conclusion validity was enhanced through an adequate sample size of 200 respondents, determined using the Yamane formula. Internal validity was addressed through the use of a control group and random assignment of participants. Construct validity was established through face validity (review by three lecturers from the Department of Education) and content validity (the questionnaire covered all five dimensions of the construct based on a thorough literature review). External validity was limited by the use of a convenience sample, and this limitation is acknowledged in Chapter Five.”
For a qualitative study:
“Trustworthiness was established using Lincoln and Guba’s criteria. Credibility was achieved through member checking, where five participants reviewed and confirmed interview transcripts and preliminary findings. Transferability was enhanced through thick description of the research context, including detailed descriptions of the school, community, and participant characteristics. Dependability was addressed through an audit trail of all research decisions, from sampling to analysis. Confirmability was ensured through reflexivity journaling and explicitly acknowledging the researcher’s position as a former teacher in similar schools.”
Conclusion
Validity is not an optional extra for advanced researchers. It is the foundation of trustworthy research. Without validity, your findings cannot be believed. Your conclusions cannot be defended. Your thesis cannot be approved.
The good news is that establishing validity is entirely achievable with careful planning. Define your constructs clearly. Use established instruments when possible. Pilot test everything. Get expert reviews. Acknowledge your limitations honestly. And always remember that a valid study is one where your conclusions genuinely reflect reality.
Your supervisor will notice the attention to validity. Your examiners will appreciate the thoroughness. And you will graduate knowing that your research actually means something.
Need help designing a valid research instrument or writing your methodology chapter? At Proposal Writers Kenya, our experts specialize in research methodology across all academic levels. From questionnaire design to validity testing to complete thesis proposal writing, we help Kenyan students produce research that stands up to scrutiny. Visit Proposal Writers Kenya to learn how we can support your thesis journey.