The recent boost in bird migration studies following the development of various tracking devices raised awareness of how detrimental attaching devices can be for animals. Such effects can occur during migration, but also immediately post‐release if the device impairs escape flight performance and, consequently, the bird's ability to evade predators. In this study, we investigated the effect of carrying a device on the escape flight speed and aerodynamic force production in a migratory passerine. We recorded upward‐directed escape flights of 15 male blackcaps. Each individual was tested without a tag, and when equipped with three different leg‐loop dummy tags with masses representing around 3%, 5%, and 7% of their body mass. The experiment was designed such that all individuals passed through all treatments in a randomized order. We found that two factors affected flight speed in roughly equal amounts: first, tagged escape flights had lower flight speeds compared to the control flights, irrespective of tag mass. Second, we found an effect of the total mass, that is, the sum of the masses of the individual bird and of the tag, with heavier birds being slower. In contrast, flight speed was not correlated with relative tag mass in percentage of body mass, the metric commonly used in ethical guidelines for tag attachment. Aerodynamic flight force production also depended on total mass, with heavier birds producing higher forces. But these flight forces did not differ between flights with or without a tag. We conclude that, when tagging birds, it is misleading to choose heavy individuals for tagging in order to minimize the tag mass as a percentage of body mass. This is particularly relevant in species for which body mass is not necessarily related to size, like migratory birds that accumulate large fat reserves. The lower escape speed in “tagged” flights could not be explained by differences in net flight force production, because these did not differ between flights with and without a tag. This suggests that the tag also affected pre‐flight take‐off dynamics, possibly due to a leg harness‐induced reduction in leg push‐off performance.