ABSTRACT
The problem of diagnostic reliability lay with clinical practitioners who disagreed with one another about the nature of clients' problems. To the researchers who began to study this issue, unreliability was the product, not of any one deficit, but of a confluence of errors made by clinicians. The typical clinical setting in which many of the early reliability studies were conducted was viewed as an unsupervised nursery for clinical errors. Another source of diagnostic error, called criterion variance, involves how clinicians make diagnostic decisions. Even if information variance can be controlled or eliminated completely, clinicians might still disagree about what pieces of information are relevant, how they should be interpreted, and what diagnostic category the disorder, if any, should be placed in. Two teams of investigators, who would ride this problem into positions of prominence and influence, began working on ways to control the behavior of freewheeling clinicians.
