Summary Animal-borne logging or telemetry devices are widely used for the measurements of physiological and movement data from free-living animals. For such measurements to be relevant, however, it is essential that the devices themselves do not affect the data of interest. A recent meta-analysis (Barron et al. 2010; Methods Ecol Evol. 1:180-187) reported an overall negative effect of these devices on the birds that bear them, i.e. on nesting productivity, clutch size, nest initiation date, offspring quality, body condition, flying ability, foraging behaviours, energy expenditure or survival rate. Method of attachment (Harness, Collar, Glue, Anchor, Implant, Breast-mounted, Tailmount) had no influence on the strength of these effects but anchored and implanted transmitters had the highest reported rates of device-induced mortality. Furthermore, external devices, but not internal devices, caused an increase in 'device-induced behaviour' (comfort behaviours such as preening, fluffing and stretching, and unrest activities including unquantifiable 'active' behaviours). These findings suggest that, with the exception of device-induced behaviour, external attachment is preferable to implantation. In the present study we undertake a meta-analysis of 183 estimates of device impact from 39 studies of 36 species of bird designed to explicitly compare the effects of externally-attached and surgically-implanted devices on a range of traits, including condition, energy expenditure, and reproduction. In contrast to Barron et al., we demonstrate that externally-attached devices have a consistent detrimental effect (i.e., negative influences on body condition, reproduction, metabolism, and survival), whereas implanted devices have no consistent effect. We also show that the magnitude of the negative effect of externally attached devices decreases with time. We therefore conclude that device implantation is preferable to external attachment, providing that the risk of mortality associated with the anaesthesia and surgery required for implantation can be mitigated. We recommend that studies employing external devices use devices that can be borne for long periods, and, wherever possible, deploy devices in advance of the time period of interest.