Content validity is whether a test's items adequately cover the full range of the concept it claims to measure, rather than sampling only part of it.
Imagine a conscientiousness test made up entirely of questions about keeping a tidy desk. Tidiness is part of conscientiousness, but so are punctuality, follow-through on commitments, and careful planning. A test that only samples the tidiness slice has a content validity problem — it's measuring a narrower thing than it claims to, even if every individual item is well written. Good content validity means the item set acts like a representative sample of the whole construct, not just the parts that were easiest to write questions about.
Content validity is usually built in during test construction rather than checked afterward with a statistic. Test developers typically start with a detailed definition of the construct, break it into its known sub-components, and then deliberately write items across all of them, often checking coverage with subject-matter experts before anything gets tested on real people.
It's easy to confuse with face validity, but the two are different questions. Face validity asks whether a test looks right to the person taking it. Content validity asks whether the test is right — whether it actually spans the construct — regardless of how it appears on the surface. A test can look impressive and still have poor content coverage, or look plain and cover the construct thoroughly.
Content validity is one of several pieces that feed into the larger question of construct validity: whether a test measures what it claims to measure overall.