A practical application of reliability theory to family practice research

The practical need for short instruments to assess constructs such as general health results in potentially imprecise measurement. According to reliability theory, a consequence of this imprecision is the inclusion of random error in the observed measurement of any construct. To the extent that measures are unreliable, clinicians may find that large changes in obtained measurements must be seen before one can have confidence that a true change in behavior has occurred. Discussion of the standard error of measurement and its application to functional assessment research is included.