Predictive validity is the degree to which scores on a test accurately forecast future performance or outcomes, such as an aptitude test predicting later college grades. In AP Psychology, it's one of the psychometric principles a test must meet to be considered useful (Topic 2.8, Unit 2).
Predictive validity asks one simple question about a test: do its scores actually forecast what they claim to forecast? If a college admissions test is supposed to predict college success, predictive validity is the evidence that high scorers really do succeed in college later. Researchers check it by correlating test scores with a future outcome (the criterion). A strong correlation between intelligence test scores and later academic performance is evidence of predictive validity. A weak one means the test might be measuring something, but not something that matters for that prediction.
In the AP Psych Revised CED, predictive validity lives inside the bigger idea that all psychological assessments, including intelligence tests, must follow sound psychometric principles to be useful (2.8.B). It pairs naturally with aptitude tests, which by definition try to predict future performance (2.8.D). And it's where the bias conversation gets real. A test can have decent predictive validity overall but still be shaped by poverty, discrimination, and educational inequities that affect both the scores and the outcomes (2.8.C).
Predictive validity sits in Topic 2.8 (Intelligence and Achievement) in Unit 2: Cognition, supporting learning objectives 2.8.B (explain how intelligence is measured) and 2.8.D (explain how academic achievement is measured compared to intelligence). The CED is blunt that a test isn't useful just because it exists. It has to be standardized, reliable, and valid. Predictive validity is the specific flavor of validity that justifies real-world decisions like college admissions and identifying students for educational services. It also connects directly to 2.8.C, because when scores are used to predict outcomes for individuals or groups, sociocultural bias can distort both the prediction and how people interpret it. If you can explain why a high score-to-outcome correlation supports a test's usefulness, and why bias can undermine that, you've got the heart of this topic.
Keep studying AP® Psychology Unit 2
Construct validity (Unit 2)
Construct validity asks whether a test measures the concept it claims to measure, like intelligence itself. Predictive validity asks whether the scores forecast a future outcome. A test needs both to be trusted, and the exam loves making you tell them apart.
Test-retest reliability (Unit 2)
Reliability is consistency, validity is accuracy. A test that gives you the same score in September and October (a 0.92 correlation, say) is reliable, but that says nothing about whether the score predicts anything. Reliability is necessary for validity but never sufficient.
Aptitude vs. achievement tests (Unit 2)
Aptitude tests exist to predict future performance, so predictive validity is literally their job description. Achievement tests measure what you already know, so they lean more on content and construct validity. This pairing comes straight from 2.8.D.
Bias in intelligence assessment (Unit 2)
Under 2.8.C, poverty, discrimination, and educational inequities can lower scores and distort the score-to-outcome relationship. So a test's predictive validity can look fine on paper while still reflecting systemic disadvantage rather than pure ability.
Predictive validity is classic multiple-choice territory. The stem gives you a scenario where someone correlates test scores with a future outcome, like a new intelligence test that correlates highly with later academic performance, and you have to name the concept. The trap answers are other psychometric principles: test-retest reliability, construct validity, standardization. The signal word is future or predict. Another common stem flips it and shows a limitation, like an admissions test that predicts first-year grades but not graduation rates, and asks what's weak about the test (its predictive validity for long-term outcomes). On the free-response side, the Article Analysis Question (AAQ) and Evidence-Based Question (EBQ) often ask you to evaluate whether a measure in a study is sound, and naming predictive validity correctly, with the score-to-outcome correlation as your evidence, is exactly the kind of psychometric reasoning that earns points.
Construct validity is about what the test measures (does this test actually capture intelligence?). Predictive validity is about what the scores foretell (do high scores lead to high grades later?). Quick check for MCQs: if the scenario compares test scores to a future outcome like college GPA or job performance, it's predictive validity. If it asks whether the test taps the underlying concept at all, it's construct validity.
Predictive validity means a test's scores accurately forecast a future outcome, like an aptitude test predicting college grades.
It's measured with a correlation between test scores and a later criterion, so a high positive correlation with future performance is the evidence.
Reliability and validity are different things. A test can give consistent scores every time and still predict nothing.
Aptitude tests live or die by predictive validity because their whole purpose is forecasting future performance, while achievement tests measure what you already know.
A test can predict some outcomes but not others, like first-year grades but not graduation rates, which limits its predictive validity.
Sociocultural bias, poverty, and educational inequities can distort both scores and outcomes, so strong-looking predictions don't automatically mean a fair test.
Predictive validity is the extent to which a test's scores forecast future performance or outcomes. The textbook example is an admissions or aptitude test whose scores correlate with later college grades. It's one of the psychometric principles in Topic 2.8 that make an assessment useful.
Reliability is consistency, validity is accuracy. If a test given in September and October produces a 0.92 correlation between the two score sets, that's test-retest reliability, not predictive validity. Predictive validity requires correlating scores with a separate future outcome, not with the test itself.
No. A test can produce nearly identical scores every time and still fail to predict anything meaningful. Reliability is necessary for validity, but it never guarantees it. This is one of the most common MCQ traps in Topic 2.8.
A psychologist validates a new intelligence test by showing its scores correlate highly with students' later academic performance. The exam also tests the limitation side, like a college admissions test that predicts first-year grades but not graduation rates, which shows weak predictive validity for long-term outcomes.
No. Construct validity asks whether a test measures the concept it claims to measure (does this really capture intelligence?), while predictive validity asks whether scores forecast a future outcome (do high scorers earn high grades later?). Look for the word 'future' in the question stem to spot predictive validity.
Connect this key term to the AP exam workflow: review the course, practice questions, and check related study tools.
Review units, study guides, and course resources.
Check this vocabulary in multiple-choice context.
Apply key concepts in written AP responses.
Estimate the exam score you are working toward.
Review the highest-yield facts before practice.
Put the full course together before test day.