Electroshocking has long been employed as a survey technique for fish, but has not been directly tested against rock rolling as a survey methodology for stream-dwelling amphibians. Electroshocking has the potential to reduce habitat disturbances that result from surveys, improve abundance estimates, and reduce injuries and effort in collecting data. Furthermore, accurately quantifying species and survey technique-specific estimates of detection probabilities is critical for appropriately interpreting survey results and employing occupancy analyses. We tested the efficiency and sensitivity of rock rolling and electroshocking in detecting Idaho giant salamanders (Dicamptodon aterrimus) in a small stream in northern Idaho, USA, by sampling short (25 m) segments of a stream using both survey techniques. We also conducted multi-pass surveys of 400-m stream segments to estimate detection probabilities for D. aterrimus and Rocky Mountain tailed frog (Ascaphus montanus). Using electroshocking, we detected D. aterrimus 40% more often than by rock rolling and detected 3.5 times as many individuals, with substantially reduced effort. Using electroshocking, detection probabilities were 1.0 for D. aterrimus and 0.79 (95% CI = 0.63–0.88) for A. montanus. Our results show electroshocking to be a much more sensitive and efficient method of detecting stream-dwelling amphibians than the traditional technique of rock rolling. Electroshocking can serve as an important survey technique for secretive stream-dwelling amphibians, allowing managers to quickly and safely acquire valuable data of better quality. © 2012 The Wildlife Society.