Face validity is whether a test looks like it measures what it claims to measure, from the perspective of the people who encounter it, including test-takers, teachers, administrators, parents, and other stakeholders. It is not a technical measurement property but a perception-based judgment about test credibility.
Face validity occupies an awkward position in testing theory. Strictly speaking, it is not "real" validity at all; Bachman & Palmer (1996) deliberately excluded it from their test usefulness framework, arguing that validity should be evidence-based, not impression-based. Messick (1989) does not treat it as a separate validity category.
Yet it has real consequences. Hughes (2003) and Brown & Abeywickrama (2010) both argue that face validity matters pragmatically, even if it lacks theoretical standing:
| Type | Question | Who judges | Method |
|---|---|---|---|
| Face validity | Does this look right? | Non-experts (students, parents, administrators) | Subjective impression |
| Content Validity | Does this sample the content domain adequately? | Subject matter experts | Systematic specification matching |
| Construct Validity | Does this measure the target ability? | Researchers, test developers | Statistical and theoretical analysis |
A test can have strong face validity but weak construct validity. A grammar translation test looks like an English test to many stakeholders (it has English sentences, grammar rules, right/wrong answers), but it may not validly measure communicative ability. Conversely, a test can have strong construct validity but weak face validity: an innovative task type that genuinely measures the target ability may look unfamiliar and therefore suspicious to test-takers.
High-stakes decisions. When test results determine university admission, immigration status, or employment, stakeholders demand that the test look credible. IELTS invests heavily in appearing to measure "real" English ability: the Speaking test involves a face-to-face conversation (not just reading aloud), and Writing tasks require extended composition (not just gap-fills).
New or unfamiliar test formats. When introducing a novel assessment approach (e.g., portfolio assessment, computer-adaptive testing), face validity concerns are heightened. Stakeholders need to understand what the test is doing and why.
Institutional contexts. If parents or administrators see a test and cannot understand how it relates to English ability, they may lose confidence in the program regardless of the test's actual validity.
Face validity alone is never sufficient. A test that looks good but does not actually measure the target construct is worse than useless because it creates false confidence. The most important validity questions require technical analysis (construct validity, item analysis, correlation studies), not just stakeholder impressions.
The ideal is a test that has both strong technical validity and strong face validity: it measures what it should and it looks like it does.