Two challenges in wildlife telemetry are optimizing the duration of transmitter attachment and minimizing the impacts of radios on the behavior and demography of the study animal. We tested 4 methods of radio attachment for a breeding population of upland sandpipers (Bartramia longicauda) under natural conditions at a tallgrass prairie site in Kansas, USA. To estimate radio retention and weekly survival rates, we used the nest survival model of Program MARK. Radio retention was lowest at the start and the end of the breeding period. The expected duration of radio retention was 1.8 years for a leg-loop harness, 40 days for radios glued to clipped feathers, 26 days for radios glued directly to feathers, and 7 days for radios glued to bare skin. Few radiomarked birds died during our study, but 4 of 8 mortality events were discovered within one week of radiomarking. Both glue and harnesses increased predation risk immediately after radio attachment. The weekly probability of survival was high after a 1-week acclimation period, and the expected survival for a 10-week breeding period was similar in males and females. Attachment of radios with glue had no effect on annual return rates. However, attachment of radios with leg harnesses resulted in lower return rates among radiomarked birds than birds without radios. Radios attached with glue were shed in <1 year but radios attached with harnesses were retained for up to 1–2 years. Our results indicate a tradeoff between optimizing radio retention and minimizing impacts on demography. Glue techniques had retention rates that were suitable for only short-term studies, but attachment with glue had no long-term effect on annual return rates. Leg harnesses provided effective radio retention that had little effect on survival rates during the stationary breeding period, but resulted in lower annual return rates. Robust estimates of radio retention and survival will assist researchers in selecting attachment techniques that best meet the study goals of future telemetry projects.