A paradox in managing threatened and endangered species is the increased need for documenting population status, which in many instances requires capturing and handling individuals. Electrofishing is a widely used method for sampling fish in small streams, but the potential for detrimental effects call for its careful use. Methods that reduce exposure for individual fish, yet still provide useful abundance estimates, are therefore desirable. Using data from a juvenile steelhead (Oncorhynchus mykiss) monitoring program in Idaho, we quantified variation and bias in capture probability from three-pass depletion electrofishing, and evaluated a method for indexing abundance based on single-pass electrofishing. Capture probability varied primarily at the level of sampling visits, with little spatial variation (i.e. at the level of the study sites). Water temperature, channel depth and, to a lesser degree, discharge best described the variation in capture probability between sampling events. We found no effect of previous capture and handling on individual capture probability, and the capture probability did not differ between subyearling and overyearling fish. Finally, a simple mixed-effects model with study site as a random effect, which related first-pass catch to the associated multiple-pass removal estimate, explained 91% of the variation in our data. The main limitation of the approach is that it does not provide error propagation and confidence intervals to the abundance estimates. However, the approach can be useful where multiple depletion electrofishing data exist, and when a single point estimate is sufficient to monitor major population trends in small streams.