We compared the results of point-count and mist-net surveys during the breeding and post-breeding seasons in four Missouri Ozark habitats: mature upland forest, mature riparian forest, 9- to 10-yr-old upland forest and 3- to 4-yr-old upland forest created by clearcutting. We determined whether differences in abundance estimates among habitats or between breeding and post-breeding seasons varied with survey method (i.e., habitat × method, season × method, or habitat × season × method interaction effects). The habitat × method interaction was significant for 13 of 16 species. The general pattern was for canopy or sub-canopy species to be detected by point counts more often in the mature-forest habitats, and to be detected by mist nets more often in the young-forest habitats. The season × method interaction was significant for 6 species largely because there was a greater decrease from breeding to post breeding season in mist-net captures than point-count detections. Two species had a significant habitat × season × method interaction. Differences in the patterns in abundance among habitats and between seasons, by survey method, are indicative of bias in the survey methods. Individuals planning bird surveys should recognize these biases and select the method that best samples the segment of the community they are surveying and that is least likely to confound treatment effects. Often this may require the use of multiple survey methods.