In Australia, the genetic integrity of dingoes (Canis lupus dingo) is threatened through hybridisation with feral dogs and consequently the identification of the modern ‘pure’ dingo is ambiguous. There are no accurate classification techniques for dingoes and dingo–dog hybrids in the wild. Genetics, skull morphology and visual assessment are methods currently used, but they often yield contrasting results. We tested skull morphological and visual assessment methods for classifying wild canids in south-eastern New South Wales and examined temporal and geographic trends in skull morphology. Published equations based on discriminant functions revealed varying percentages of dingoes, dogs and their hybrids over time, and did not yield similar results to visual assessment methods. Skull characteristics generally became larger over time but have recently stabilised. Changes in the morphology of the molars were consistent with the occurrence of hybridisation with dogs. Geographic variation was apparent and consistent with Bergmann’s Law, with skulls increasing in size with altitude. This study highlights the importance for improved classification methods of wild canids and the importance of considering geographical variation in morphological studies.