Standardized tests: Test validity
The internal validity of a test is the extent to which
it measures what it is supposed to measure. The external validity of a test
refers to how well it can be generalized to others in the population for which
it was developed. There are 3 additional criteria or types of evidence for
test validity: construct, content, and criterion validity.
validity - refers to the association
of the test with an underlying theory. Does the test produce a
result that is in accord with an established theory? For example,
according to psychoanalytic theory, anxiety is a common response
to stressful events. A test that shows an increase in anxiety in
people who have been traumatized would possess construct validity.
|Content validity -
sometimes known as "face validity." On the face of it,
does the test or inventory look like it measures what it is supposed
to measure? Does the content make sense? An item asking about how
gears work would appear to measure mechanical ability. Asking someone
to indicate on a checklist whether they are feeling happy or sad
sounds like a good way to find about that person's mood.
validity - do the test results match up with other known measures
of a characteristics (the criterion)? There are two (2) types of criterion
validity: concurrent and predictive.
validity - a test result is compared with another
indicator (criterion) at the same time. For example,
to check the concurrent validity of the State Anxiety Scale,
it could be administered at the same time as the older, established
Taylor Manifest Anxiety Scale. If both scales give the same
result, then the newer one possesses criterion validity.
It is valid because it gives the same result as a known criterion
validity - the test result predicts a later outcome.
Police selection example: Applicants who scored low on the
test (but passed and were hired) had more discipinary citations
in their records a year later than person who scored high
on the test. The test has predictive validity in that it
predicts (fortells) subsequent behavior.
Next section: Projective tests