This review synthesizes a wide literature on the agreement of treating clinicians’ PD diagnoses with each other and their convergence with common research methods. Median interrater reliability between clinicians was moderate when calculated dimensionally (r = .46) or categorically (κ = .40). The agreement between clinicians’ diagnoses and those from research methods (e.g., self-report questionnaire) was more modest. Median dimensional agreement across 27 studies ranged from .05 to .36, with an overall median of .23. This overall value was moderated by several factors. First, clinicians’ diagnoses agreed more with semi-structured interviews than self-report questionnaires. Second, convergence increased slightly when clinicians utilized more systematic diagnostic methods. Results suggest relatively little overlap between PD diagnoses assigned in research versus naturalistic settings.