What is construct underrepresentation and construct-irrelevant variance in the context of test validity?

• A. Underrepresentation means missing key aspects of the construct; irrelevant variance means including noise unrelated to the construct.
• B. Underrepresentation means too many items; irrelevant variance means including only relevant items.
• C. Underrepresentation means noise in data; irrelevant variance means too few items.
• D. Underrepresentation means sampling bias in the test population.

Answer: (A)

Construct validity hinges on how well a test measures the intended construct without being distorted by irrelevant factors. Two key ideas describe common threats: underrepresentation and construct-irrelevant variance.

Underrepresentation means the test doesn’t cover all important aspects of the construct. If essential facets are left out, the test provides an incomplete or biased picture of what it’s meant to measure, so scores don’t fully reflect the domain being assessed. For instance, measuring problem-solving with only one kind of problem leaves out other meaningful ways people solve problems, so the construct isn’t fully represented.

Construct-irrelevant variance is the part of test scores that comes from factors unrelated to the construct. This is noise that can include language difficulty, test-taking speed, cultural biases, or confusing item wording. When these extraneous factors influence performance, the scores don’t cleanly reflect the construct, harming validity by muddying the link between what the test is supposed to measure and what the score represents.

Both issues weaken construct validity: one by failing to cover the full construct, the other by letting irrelevant factors seep into scores. The other options mischaracterize these terms (for example, suggesting too many or too few items or focusing on sampling bias), which doesn’t capture what underrepresentation or irrelevancy variance means.