Validity shows the extent to which a test measures what it claims to, and so the meaning that users can attach to test scores. There are many different types of validity, though in organisational settings the main ones are content, construct and criterion validity. Reference may also be made to other types of validity such as face validity, which concerns the extent to which a test looks job-relevant to respondents.

Content validity relates to the actual questions in the test or the task that test takers need to perform. The more closely the content matches the type of information or problems that a test taker will face in the workplace, the higher its content validity. For tests such as personality or motivation, content validity relates more to the relevance of the behaviours assessed by the test rather than the actual questions asked.

Construct validity shows how the constructs measured by the test relate to other measures. This is often done by comparing one test against another. Where tests measure multiple scales, as is the case with assessments of personality and motivation, it is also common to look at how the measure's scales relate to each other.

Criterion validity looks at the extent to which scores on the test are statistically related to external criteria, such as job performance. Criterion validity may be described as 'concurrent' when test scores and criterion measures are taken at the same time, or 'predictive' when test scores are taken at one point in time and criterion measures are taken some time later.

Construct and criterion validity are often indicated by correlation coefficients which range from 0, indicating no association between the test and criterion measures, and 1, indicating a perfect association between the test and criterion measures. It is difficult to specify precisely what an acceptable level of validity is, as this will depend on many factors including what other measures the test is compared against or what criteria are used to evaluate its effectiveness. However, for criterion validity, tests showing associations with outcome measures of less than 0.2 are unlikely to provide useful information and ideally criterion validity coefficients should be 0.35 or higher. The samples used for criterion validity studies should also be at least 100.

Overall, whilst a publisher should provide validity evidence for their test, validity comes form using the right test for the right purpose. Therefore, users need to use available validity evidence to evaluate the relevance of the test for their specific purpose.