An assessment of integrated population models: bias, accuracy, and violation of the assumption of independence

Understanding population dynamics requires accurate estimates of demographicrates. Integrated population models combine demographic and survey data into asingle, comprehensive analysis and provide more coherent estimates of vital rates. Integratedpopulation models rely on the assumption that different data sets are independent, which isfrequently violated in practice. Moreover, the precision that can be gained using integratedmodeling compared to conventional modeling is only known from empirical studies. Thepresent study used simulation methods to assess how the violation of the assumption ofindependence affects the statistical properties of the parameter estimators. Further, the gainsin precision and accuracy from the model were explored under varying sample sizes. Forcapture–recapture, population survey, and reproductive success, we generated independentand dependent data that were analyzed with integrated and conventional models. We foundonly a minimal impact of the violation of the assumption of independence on the parameterestimates. Furthermore, we observed an overall gain in precision and accuracy when all threedata sets were analyzed simultaneously. This was particularly pronounced when the samplesize was small. These findings contribute to clearing the way for the application of integratedpopulation models in practice.
Keywords: accuracy, Bayesian, capture–recapture, dependent data, independent data, individual-based model, Leslie matrix model, population survey data, reproductive success, state-space model