Direct comparison
Internal vs external validity — what is the difference?
Internal vs external validity explained: the difference is a sound cause-effect link within a study versus how well results generalise beyond it.
Side-by-side comparison
| Dimension | Internal validity | External validity |
|---|---|---|
| What it is | Confidence that the independent variable caused the observed effect. | Confidence that findings generalise beyond the study sample and setting. |
| Core question | Did the treatment really cause the outcome here? | Will this result hold for other people, places and times? |
| Main threat | Confounding variables and uncontrolled alternative explanations. | Unrepresentative samples or artificial, atypical conditions. |
| How it is strengthened | Randomisation, control groups, blinding and tight control. | Representative sampling, varied settings and replication. |
| Typical strong design | Tightly controlled laboratory experiment. | Field study or large, diverse real-world sample. |
| What it protects | The causal claim within the study. | The applicability of the claim outside the study. |
| Direction of inference | Inward — about the study itself. | Outward — about the wider population. |
| Relationship to the other | Often increases as control increases. | Can decrease as artificial control increases. |
| Trade-off | Maximising control can reduce real-world realism. | Maximising realism can let confounds creep in. |
Common questions
FAQ
Why is there a trade-off between internal and external validity?+
Tightening experimental control removes confounds and raises internal validity, but the more artificial and controlled the conditions become, the less they resemble real-world settings, which lowers external validity. Researchers rarely maximise both at once and instead balance them to suit the study’s aim.
Which type of validity matters more?+
It depends on the goal. Internal validity is usually prioritised when the aim is to establish causation, because a confounded result is uninterpretable. External validity matters more when the aim is to apply findings broadly. Replication across settings is the strongest way to build both over time.
How do you improve external validity?+
Use representative sampling, test in varied and realistic settings, and replicate the study across different populations and times. Field experiments and large, diverse samples help, as does clearly specifying the conditions under which the effect is expected to hold so others can test generalisability.
