Aims: To estimate the criterion validity of the Research Diagnostic Criteria for Temporomandibular Disorders (RDC/TMD) Axis I TMD diagnoses. Methods: A combined total of 614 TMD community and clinic cases and 91 controls were examined at three study sites. RDC/TMD Axis I diagnoses were algorithmically derived from an examination performed by calibrated dental hygienists. Reference standards ("gold standards") were established by means of consensus diagnoses rendered by two TMD experts using all available clinical data, including imaging findings. Validity of the RDC/TMD Axis I TMD diagnoses was estimated relative to the reference-standard diagnoses (gold standard diagnoses). Target sensitivity and specificity were set a priori at >= 0.70 and >= 0.95, respectively. Results: Target sensitivity and specificity were not observed for any of the eight RDC/TMD diagnoses. The highest validity was achieved for Group Ia myofascial pain (sensitivity 0.65, specificity 0.92) and Group Ib myofascial pain with limited opening (sensitivity 0.79, specificity 0.92). Target sensitivity and specificity were observed only when both Group I diagnoses were combined (0.87 and 0.98, respectively). For Group II (disc displacements) and Group III (arthralgia, arthritis, arthrosis) diagnoses, all estimates for sensitivity were below target (0.03 to 0.53), and specificity ranged from below to on target (0.86 to 0.99). Conclusion: The RDC/TMD Axis I TMD diagnoses did not reach the targets set at sensitivity of >= 0.70 and specificity of >= 0.95. Target validity was obtained only for myofascial pain without differentiation between normal and limited opening. Revision of the current Axis I TMD diagnostic algorithms is warranted to improve their validity. Keywords: diagnostic criteria, gold standard, reference standard, temporomandibular disorders, temporomandibular muscle and joint disorders, validity