Evaluating sources of censoring and truncation in telemetry‐based survival data

摘要: Bias in vital rate estimation may come from failing to meet a variety of assumptions during the stages sampling, monitoring, and analysis, though most are not commonly addressed published studies. Here, we pay specific attention various forms censoring truncation that present challenges for telemetry-based monitoring survival. We use simulations assess how uncertainty about times death imperfect detection probabilities affect Kaplan–Meier survival estimates. then treat threatened woodland caribou (Rangifer tarandus caribou) west-central Alberta as case study test potential effects non-random right interval on report frequency (e.g., daily vs. monthly) associated exact time do inherently induce bias point estimates nor variance. Removing individuals at-risk pool intervals which they were detected did negative resulting when probability was independent animals' fates. recommend using subsequent detections impute fates missed such cases all eventually known or permanently censored related changes mortality risk. Although some remain difficult continued concern, find no evidence biases methodology Alberta's program. Our results lend credence recent widespread declines populations across province. © 2015 The Wildlife Society.