Internal Validity & External Validity

What is internal validity?

The extent to which a study rules out alternative explanations by controlling unwanted effects of extraneous variables.

When is internal validity threatened?

When factors outside the study could explain or influence the results.

What are the seven common threats to internal validity?

History, maturation, testing, instrumentation, statistical regression, selection bias, and attrition.

: What is the history threat?

When outside events occur between observations that influence participants' behavior.

Give an example of a history threat.

Participants attend a support group or watch related YouTube videos during the study.

What is the maturation threat?

When natural changes in participants (growing older, healing, or becoming tired) occur over time.

Give an example of a maturation threat.

Aphasia recovery might improve naturally after a stroke, not just due to treatment.

What is the testing threat?

When participants' performance changes because they become familiar with the test or expect certain outcomes.

What is the instrumentation threat?

When the measurement tools or observer criteria change during the study.

How is testing different from instrumentation?

Testing affects participants' responses, while instrumentation affects the measurement process itself.

What is statistical regression?

When participants with extreme scores naturally move closer to the average on retesting.

What is selection bias?

When groups differ before the study begins, making comparisons unreliable.

What is attrition (mortality)?

When participants drop out during a study, possibly changing the group's characteristics.

Why is attrition especially problematic?

If one group loses more participants than another, it can create selection bias.

What is external validity?

The extent to which study results can be applied to other people, settings, or conditions.

When is external validity threatened?

When unique features of a study limit how findings apply beyond it.

What are the two main categories of external validity threats?

Subject characteristics and contextual characteristics.

What are examples of subject characteristic threats?

Differences like animal vs. human, male vs. female, mild vs. severe disorder, or WEIRD vs. non-WEIRD populations.

What does WEIRD stand for in research?

Western, Educated, Industrialized, Rich, Democratic.

Why is the WEIRD problem important?

Many studies use WEIRD populations, which may not represent other cultures.

What are contextual characteristics?

Factors in the study's setting that make it different from real-world conditions (e.g., lab vs. clinic).

What are reactive effects of experimental arrangements?

When participants act differently because they know they're being studied.

What are novelty effects?

When participants respond differently because the treatment feels new or exciting.

What is assessment reactivity?

When being observed or tested changes how participants behave.

Can external validity threats ever be completely eliminated?

No, researchers reduce them through replication and diverse participant samples.

Why is replication important for external validity?

It confirms whether findings hold true in different settings and populations.