Which statement describes a threat to external validity?

Which statement describes a threat to external validity?

• A. Nonrepresentative samples are not a threat to external validity.
• B. Artificial settings always improve external validity.
• C. Large sample size alone determines external validity.
• D. Nonrepresentative samples and artificial settings together threaten external validity.

Answer: (D)

External validity is about whether study findings would generalize beyond the exact people, settings, and times of the research. A nonrepresentative sample threatens this generalizability because if the participants don’t reflect the broader population of interest, the results may not hold for others. Artificial settings threaten external validity because behaviors can change when people are in an unusual or controlled environment, so findings might not transfer to real-world contexts. When both occur—a nonrepresentative sample studied in an artificial setting—the threat to generalizability is even stronger, since the results are unlikely to apply to the broader population or to normal, real-world environments. For example, a lab study using mostly college students in a quiet, controlled setting might not apply to older adults in everyday life. The other statements miss these nuances: nonrepresentative samples do threaten external validity, artificial settings do not always improve it, and simply having a large sample size does not determine generalizability.

External validity is about whether study findings would generalize beyond the exact people, settings, and times of the research. A nonrepresentative sample threatens this generalizability because if the participants don’t reflect the broader population of interest, the results may not hold for others. Artificial settings threaten external validity because behaviors can change when people are in an unusual or controlled environment, so findings might not transfer to real-world contexts. When both occur—a nonrepresentative sample studied in an artificial setting—the threat to generalizability is even stronger, since the results are unlikely to apply to the broader population or to normal, real-world environments. For example, a lab study using mostly college students in a quiet, controlled setting might not apply to older adults in everyday life. The other statements miss these nuances: nonrepresentative samples do threaten external validity, artificial settings do not always improve it, and simply having a large sample size does not determine generalizability.