A study of reading specialists conducted in 1977 revealed that their diagnostic reliability was very low. Mean diagnostic agreement between two clinicians on statements seen as characterizing a case was effectively zero. Mean agreement for one clinician's diagnostic statements on a case and its replicate over time was less than 0.23. In other words, fewer than one quarter of the statements were repeated for the identical case. Five subsequent studies were designed to account for the low findings and to expand the generalizability of the results to learning disabilities specialists and classroom teachers as well. Mean diagnostic agreement between two clinicians remained close to zero across all five studies. Mean diagnostic agreement for one clinician on a case and its replicate remained close to 0.23 across the five cases. The results were consistent across fields (reading, learning disabilities), types of practitioner (clinicians, resource teachers, classroom teachers), and settings (laboratories, classrooms). (Author/FL)