1. Researchers often attach transmitters and other devices to free-living birds without a clear understanding of potential deleterious consequences to their study organisms, and thus to their data. Studies investigating this topic have generally been limited to a single species or type of device.
2. To achieve a broader understanding we used a meta-analysis of 84 studies to ask: (1) Do devices have an overall effect on birds? (2) Which aspects of avian behaviour and ecology are affected? (3) What attributes of birds influence transmitter effects? (4) What attributes of devices influence their effects? (5) Are effects partially a consequence of capture and restraint?
3. We found a significant negative effect of devices on birds, both overall and for 8 of the 12 specific aspects analysed. The most substantial effects were that birds with devices had markedly increased energy expenditure and were much less likely to nest.
4. Effects were independent of attributes of the birds (sex, age, primary method of locomotion and body mass). We also found no evidence that proportionally heavier devices had greater effects, although researchers generally avoided using heavy devices. Breast-mounted and harness attachments increased device-induced behaviours such as preening, however, and the risk of device-induced mortality differed between attachment methods.
5. Other than foraging behaviours, no effects were a consequence of capture or restraint.
6.Synthesis and applications. We provide the first comprehensive evidence that transmitters and other devices negatively affect birds and may bias resulting data. Researchers should balance the benefits of using these techniques against potential costs to the birds and reliability of the data obtained.