Strategic Human Resource Management

Section Four

Reliability and Validity

A number of requirements must be met before firms can make
good selection decisions. First of all, selection procedures need
to have the psychometric properties of reliability and validity.
Reliability means that the procedure, be it a test or an
interview, needs to produce approximately the same result
when the procedure is repeated. Thus, if a test is used, an
applicant’s score from one administration of the test should
correlate well with his or her score on the next administration
of the test. This form of reliability is called test–retest reliability.
Similarly, when a manager interviews job applicants, his or her
assessments of the applicants should correlate well if the
interviews are repeated at a later date. In addition, better
results are achieved when multiple interviewers are used.
Interviews are reliable when there is agreement among the
interviewers about an applicant, which means that their ratings
are correlated. This form of reliability is called interrater
reliability. Greater reliability is possible when interviewers can
be trained to gather information from applicants on dimensions
that can be assessed with some degree of accuracy, such as
communication and interpersonal skills. Interviewers also gain
useful information by using interviews to fill in gaps in
information obtained from the résumé, application, and
references.