Testing is a threat to internal validity. Which scenario illustrates this threat?

Using the same test multiple times in a study can introduce a threat to internal validity known as testing effects. This scenario occurs when participants take the same test more than once, which can lead to improved performance due to familiarity with the test format or questions rather than any actual change in the underlying construct being measured. This learning effect can distort the results, making it challenging to accurately assess the true impact of the intervention being studied.

For example, if a group of participants is tested on their knowledge of a subject, and they re-take the same test, their scores may improve simply because they remember the answers from the first test rather than due to increased knowledge. This can create a misleading impression that the treatment or intervention was effective when, in fact, the improvement may not reflect a genuine change.

In contrast, changing the instruments used for measurement, while potentially problematic, specifically relates to measurement validity rather than a direct testing effect. Randomly selecting participants is a method to enhance external validity and reduce selection bias, and collecting data at different times may be a necessary method for longitudinal studies, though it can introduce variability based on when data is collected rather than the threat presented by repeated testing.