The NSA recognizes three types of evidence which can be collected to ensure test validity:
1. Criterion-Related Evidence
Evidence demonstrating the systematic relationship of test scores on a predictor test to a predicted criterion variable.
2. Construct-Related Evidence
Empirical evidence that (1) supports the posited existence of a hypothetical construct and (2) indicates an assessment device does, in fact, measure that construct.
3. Content-Related Evidence
Evidence indicating that an assessment instrument suitably reflects the content domain it is supposed to represent
Choosing the type of evidence to collect is based on the type of test.
Because criterion-related evidence of validity is concerned with aptitude tests, the NSA does not use this type of evidence in assessing validity of items on the National Spanish Examinations. Similarly, the NSA does not collect construct-related evidence, because there are no hypothetical constructs assessed on the NSA.
Content-related evidence is the most important in assessing validity for the National Spanish Assessment, because the achievement section of the National Spanish Assessments is designed from specific learner outcomes based on content standards for specifications for vocabulary and grammar.
Furthermore, content-related evidence of validity is ensured through both (1) development care and (2) external reviews.
Additionally, items in the proficiency section (Interpretive Reading & Interpretive Listening) are validated through a high number of samples which are placed into the applicable level; similarly to the National Spanish Exams.
More information can be found here.