Internal and external validity, these terms pop up all the time in research methods courses, but what do they actually mean?
If you’ve ever wondered how to tell whether a study is truly solid or just looks good on paper, you’re in the right place.
In this article, we’re cutting through the jargon and getting to the heart of what internal and external validity really mean, clearly explained, with examples that actually make sense.
So whether you’re working on a research project or just trying to wrap your head around these concepts for an exam, let’s break it down together.
What was validity again?
Exactly—it’s one of the three quality criteria used to assess the rigor of scientific studies.
Validity answers the question:
How well does a method measure what it is supposed to measure?
Or put differently: Can I answer my research question using this approach?
In another article, I’ve already given a general overview of the three quality criteria: objectivity, reliability, and validity.
But to clarify what validity means, here’s a simple example:
Imagine you step on the scale every morning to check your body weight. However, the scale is off by 2 kg every time. It provides consistent readings, so it’s reliable. But because it doesn’t show your true weight, it lacks validity.
The question of validity mainly comes up in quantitative research. Within this context, the characteristics of validity have been refined over time. In general, we distinguish between internal and external validity.
#1 Internal validity (quantitative research)
Internal validity refers to the degree to which researchers can draw causal conclusions from the results of a particular empirical study.
The most cited work on this is the classic “Experimental and Quasi-Experimental Designs for Research” by Campbell and Stanley (1963). Since experiments are considered the gold standard for testing causal relationships, internal validity is especially important in this context.
In an experiment, we test the causal link between an independent variable and a dependent variable. Internal validity is achieved when valid causal relationships can be derived from the results.
But how do we determine that?
Essentially, by ruling things out. Are there any other factors that might have influenced the dependent variable—factors that could explain the change even without manipulating the independent variable? In other words, are there any confounding variables that blur the causal relationship?
Let’s take an example:
In an experiment, the temperature inside a car is adjusted. People sitting in the car are asked how comfortable they feel at each temperature. Temperature is the independent variable. The reported comfort level is the dependent variable.
If you want to measure the effect of temperature on comfort, you better make sure the seat massage function is turned off! Otherwise, people might feel more comfortable—not because of the heat, but because of the massage. That doesn’t mean temperature has no effect, but it does make it harder to separate the influence of heat from the massage.
So, internal validity is all about minimizing the influence of potential confounding variables (Hussy et al., 2010).
#2 External validity (quantitative research)
Once internal validity is ensured, we can think about external validity. This refers to how well the results can be transferred to other contexts—such as different people, places, or points in time.
More casually, external validity is about whether the findings can be generalized to other situations (Hussy et al., 2010).
The term isn’t only used in experimental research. It’s also relevant in survey research, where generalizability matters too.
So how do we establish external validity?
Again, we use a kind of elimination process to identify and reduce threats to external validity.
One such threat is sampling bias—whether the sample is skewed in some way. This might happen, for example, if participants didn’t volunteer freely, or if certain conditions make the sample unrepresentative of the broader population.
#3 Internal and external validity in qualitative research
In quantitative research, internal validity has top priority. Only once that’s established do we look at external validity.
In qualitative research, it’s the opposite—or more accurately, internal validity doesn’t apply. Qualitative research isn’t about testing causal relationships, so internal validity isn’t relevant.
External validity, however, can play a role. Generalizing to a broader population might be important, especially if the sample is randomly selected.
For example, if you’re conducting a content analysis and select a sample of 100 videos from a population of 1,000 YouTube videos, you might ask whether the findings generalize to the other 900. But this kind of generalization isn’t usually the goal of qualitative studies.
Where external validity does matter is in generalizing to other contexts and situations. That is a goal of qualitative research. Just think about Grounded Theory: the aim is to generalize the findings of an empirical study by developing useful theoretical concepts and identifying their relationships (Hussy et al., 2010).