Internal consistency: reports of its death are premature.

ABSTRACT

McCrae, Kurtz, Yamagata, and Terracciano (Personality and Social Psychology Review 1528-50, 2011) presented data supporting the position that test-retest reliability is more important for methods of personality assessment than is internal-consistency reliability. We raise two points not adequately considered in their findings. First, not all personality assessment is concerned with assessing traits, for which temporal stability is clearly important. Much research (e.g., therapy outcome studies) is concerned with assessing states, in which an emphasis on temporal stability can lead to the underestimation of change. Second, we suggest that using the literature as a basis for their position causes the role of internal consistency to be underestimated. This is because manuscripts are frequently rejected because their measures possess low internal consistency, but are rarely rejected because of low temporal stability. Thus, the effects of internal consistency are range-restricted to an unknown amount, because investigators do not have access to studies that have been rejected due to low internal consistency.