Polycystic ovary syndrome (PCOS) is a common disorder but has considerable phenotypic variability and this has led to controversy over its exact definition and diagnosis. The objective of this study was to review the recently proposed diagnostic criteria to determine whether they were sufficiently robust for clinical and research practise. We have reviewed the literature pertaining to clinical measurement and quality of laboratory analysis in relation to clinical and biochemical hyperandrogenism. The recently published statements regarding the diagnosis of PCOS assume that the clinical, laboratory and imaging studies are dichotomous variables, without considering the effect of observer subjectivity or measurement variability on the outcome. The data suggest that there is considerable uncertainty of all measurements and lack of clarity of the definition of the term ‘hyperandrogenaemia’ which can lead to misdiagnosis. The current diagnostic strategies for PCOS are defined too vaguely to ascertain that individuals fit the definition of the syndrome. A pragmatic approach may be taken in the management of an individual depending upon her particular symptoms and needs. However, research into the epidemiology, pathophysiology and treatment of PCOS will require the production of robust definitions of the diagnostic criteria. We propose specific criteria that would answer the criticisms that we have raised.