Accurately detecting the presence or absence of threatened species is vital for threatened species management, and the detection power of individual sampling methods can vary significantly between species and life stages. The present study compares the detection power of six sampling methods in sampling the endangered Macquarie perch in riverine habitats in south-eastern Australia. In an initial survey in 1998 and 1999, fyke nets captured Macquarie perch at 100% of sites where the species was detected; gill-nets captured the species at 86%; with no other method having >50% detection efficiency. Most Macquarie perch were captured by fyke nets (90% in 1998 and 94% in 1999), followed by gill-nets (7 and 2%). A monitoring program at one of the survey sites over 7 years returned similar results with fyke nets detecting the species in all years. Fyke nets captured primarily young-of-year (YOY) individuals, whereas gill-nets captured adults and subadults. Boat electrofishing returned a high level of false negatives for Macquarie perch. Future sampling for this species should employ fyke and gill-nets to adequately characterise population structure (adults, juveniles, YOY), minimise false negatives and detect the occurrence of successful breeding the previous year.