Nest placement presumably reflects selection for secure sites to minimize failure. Most tests of this hypothesis, however, have failed to support it. We used artificial nests (ARTs) to experimentally evaluate nest-site-choice behavior by an open-cup–nesting bird, the Eastern Kingbird (Tyrannus tyrannus). In 2010 and 2011, we placed ARTs in trees in the riparian zone at Malheur National Wildlife Refuge, Oregon, USA, to test whether (1) characteristics describing the physical location in trees of used and unused ARTs differed, (2) used ART sites more closely resembled naturally chosen sites, (3) successful natural nests (NATs) and successful ARTs were similarly located along the major axis describing nest placement, and (4) unused ARTs resembled failed NATs. Used and unused ART sites differed, but unused ART sites were more similar to NAT sites. The latter unexpected result occurred because (1) unused ARTs were located at sites between more heavily used higher and lower locations and (2) most kingbirds nesting at lower locations used ARTs instead of building their own nest. In both ARTs and NATs, differences between successful and failed nests exhibited the same pattern for most nest-site variables, and the major gradient describing nest location was the same; successful nests tended to be placed on more vertically oriented branches that were placed closer to the top of the tree. Kingbird nest placement was thus selective. However, extensive overlap in the locations in trees of failed NATs and both successful and unused ARTs suggests that other factors, such as macrohabitat characteristics or prior experience of individual birds with particular nest sites, may have influenced success and/or decisions to use or reject nests in particular locations. Thus, consideration of phenomena beyond the nest site itself may be required to fully understand the process of nest-site choice in birds.