The extent to which a measure correlates with one or more criterion variables.
A type of test validity that addresses the degree to which the test adequately samples the relevant material.
A type of test or research validity that addresses the question of what is really being assessed by a particular test, or manipulated or investigated in an experiment, for example, the validity of an abstract expression, term, or concept that is presumed to explain or give meaning to the relationship between X and Y in experimental research.
The extent to which observations or measures are consistent or stable.
The effect of uncontrolled variables that often can be specifically identified; such effects are, theoretically speaking, not self-canceling (in contrast to the self-canceling nature of random errors).