We estimated survival rates of discarded black sea bass (Centropristis striata) in various release conditions using tag–recapture data. Fish were captured with traps and hook and line from waters 29–34 m deep off coastal North Carolina, USA, marked with internal anchor tags, and observed for release condition. Fish tagged on the bottom using SCUBA served as a control group. Relative return rates for trap-caught fish released at the surface versus bottom provided an estimated survival rate of 0.87 (95% credible interval 0.67–1.18) for surface-released fish. Adjusted for results from the underwater tagging experiment, fish with evidence of external barotrauma had a median survival rate of 0.91 (0.69–1.26) compared with 0.36 (0.17–0.67) for fish with hook trauma and 0.16 (0.08–0.30) for floating or presumably dead fish. Applying these condition-specific estimates of survival to non-tagging fishery data, we estimated a discard survival rate of 0.81 (0.62–1.11) for 11 hook and line data sets from waters 20–35 m deep and 0.86 (0.67–1.17) for 10 trap data sets from waters 11–29 m deep. The tag-return approach using a control group with no fishery-associated trauma represents a method to accurately estimate absolute discard survival of physoclistous reef species.