Reliability Estimation When a Test Is Split Into Two Parts of Unknown Effective Length

In some measurement settings internal consistency reliability of a measure must be based on a partition of the instrument into only 2 parts that cannot be further subdivided. Each of these 2 parts yields only a single score. If the functional lengths of the parts appear to be unequal or the parts are scored on different scales, the setting calls for a congeneric coefficient. It is shown that a single-valued estimate of the total score reliability is possible only if an assumption is made about the comparative size of the error variances of the parts. Without such an assumption, a range of reliability estimates is consistent with the part-test variances and covariance. But if the reliability of 1 part can be estimated independent of scores on the 2nd part, then a single-valued congeneric estimate of total score reliability is possible.