In captive observation, experimental subjects are retained in captivity for a period of time to monitor their survival. Captive observation can be conducted in field assessments or in the laboratory. Laboratory-based assessments can be used to investigate the effects of isolated variables on the behaviour, physiology, and survival of subjects under controlled conditions. Stressors representative of normal fishing conditions are usually not reproducible in the laboratory and, therefore, these assessments are considered less suitable for generating fishery-specific discard survival estimates. In field assessments, experimental subjects are collected under representative fishing conditions and then transferred to containment facilities for captive observation. Captive-observation techniques may introduce biases to survival estimates. They do not include predation effects on survival of discards, nor some stressors that would otherwise have been experienced by the fish had they been discarded back into the sea, and so may overestimate the true survival rate. In addition, they may add handling and captivity stressors and, therefore, can underestimate the true survival rate.