We investigated the effects of predation risk on birth-site selection by Rocky Mountain elk (Cervus elaphus nelsoni) during summer in 2002–2004 in northeastern Oregon at macrohabitat (3rd-order selection) and microhabitat (4th-order selection) scales. This study describes vegetative characteristics of birth sites selected by female elk when young <4–5 days old used the hiding strategy and predation sites when most predation events occurred on young >5 days old that used the fleeing strategy. At the macrohabitat scale we observed no evidence that female elk were influenced by predation risk when selecting a birth site on the basis of variables measured in this study. Females chose birth sites with less overhead cover than random sites, suggesting that they might have been influenced more by forage availability than predation risk. At the microhabitat scale females selected birth sites that had more overhead canopy cover and greater visibility at ground level than paired random sites, which suggested that birth-site selection at this scale was influenced by predation risk. Together, these results suggested that female elk selected areas for parturition at the macrohabitat scale that likely had forage to meet high nutritional demands of lactation and at the microhabitat scale selected areas that provided visibility to detect predators and reduce the risk of predation. Predators, mainly cougars (Puma concolor), killed young in areas closer to vegetative edges at the macrohabitat scale and with more visibility at the microhabitat scale. These areas were likely conducive to cougar hunting where sight and cover from forest edges can be important for stalking calves that are traveling with their mothers and family groups.