Voice identification parades can be unreliable, as earwitness responses are error-prone. Here we vary pre-parade instructions, testing performance across serial and sequential procedures to examine ways of reducing errors. The participants listened to a target voice and later attempted to identify it from a parade. They were either warned that the target may or may not be present (standard warning), or encouraged to consider responding ‘not present’ because of the associated risk of a wrongful conviction (strong warning). Overall accuracy was low. Performance varied according to instructions and procedure. False alarms were lower on target-absent serial parades following the strong compared to the standard warning. However, the strong warning was associated with higher false alarms on target-absent sequential parades. We discuss the cognitive processes that might drive this effect. Our novel analyses shed light on these results, highlighting the challenges of directly comparing procedures, and revealing position-related effects.