Due to its secretive nature and nocturnal vocalization, multi-species bird monitoring programs are not effective in surveying populations of Yellow Rails (Coturnicops noveboracensis) and, thus, species-specific survey methods should be used. To determine how to optimize nocturnal call-playback surveys of Yellow Rails, we evaluated the effects of survey methods (naïve-estimated vs. detectability-adjusted estimated occupancy, observer, number of surveys, and the use of playbacks) and temporal and environmental conditions (e.g., time, date, temperature, moon phase, seasonality, and cloud cover) on detection probability. In 2010 and 2011, 334 call-broadcast night surveys for Yellow Rail were conducted at 167 survey points within 80 wetlands in south-central Manitoba, Canada. Yellow Rail detection probability was estimated at 0.63 in both years. In 2010, the detectability-adjusted wetland occupancy rate was estimated at 0.63, and in 2011 it was estimated at 0.36. Call-broadcast surveys contributed relatively little to improving Yellow Rail detectability, but repeat surveys at each site increased the number of individuals detected. Detection probability was not correlated with the temporal or environmental variables we studied, or by observer. Surveys where call-broadcasts are not feasible, such as volunteer surveys, are still likely to result in good estimates of Yellow Rail abundances, if surveys are repeated within breeding seasons.