On reliability in psychological measures
“The repeated stress on the need for reliability in psychological measures is a sign of the only half-realized but persistent dominance of trait psychology and the belief that what we must seek and find are fixed characteristics of an object rather than an understanding of process . . . . The most obvious feature of persons is that they change, and grievous though this may be to the champions of reliable psychological measurement, it must be recognized that we must seek to understand change and to measure our degree of understanding by the degree to which we can predict it…” (pp.161–162)
Reference:
Bannister, D. & Bott, M. (1973) Evaluating the person. In P. Klein (Ed.), New Approaches to Psychological Medicine. Chichester: John Wiley & Sons.