Section Four
Selection procedures also must be valid. Validity means
that the procedure predicts what it is supposed to predict. With
selection procedures, this means that the test or interview
predicts job performance in the position for which applicants
are being selected. In order for selection procedures to be
valid, they must first be reliable. There are several forms of
validity relevant to selection. The first is predictive validity. With
predictive validation, potential selection procedures, such as
batteries of tests, are given to job applicants. However, during
the validation procedure, selection is based on procedures
other than the tests. After sufficient time for those hired to
have learned their jobs, their job performance is then
correlated with test scores. Tests having a significant and
substantial statistical relationship with job performance are said
to have predictive validity.
Another form of validity is concurrent validity, which has
the practical advantage of not requiring a time lag before tests
can be used. With concurrent validity, tests are given to current
employees and their test scores are compared with their job
performance. Both predictive and concurrent validation are
called criterion-related forms of validity because they are
validated on the criterion of job performance. A third form of
validity is content validity. There is no statistical basis for
assessing the content validity of a selection procedure, such as
a test. Instead, content validity is established by having a panel