Abstract: Wildlife population models are potentially valuable for conservation planning. Validation is necessary to ensure that models are sufficiently robust for predicting management outcomes consistent with conservation objectives. Sorensen et al. (2008) produced a model of woodland caribou ( Rangifer tarandus ) population growth rate that was recently modified and used as a predictive tool at several scales. We computed confidence intervals and evaluated the performance of this model using novel data. Confidence intervals were wide, and results suggested that the model may have a positive bias, resulting in over‐estimation of population growth rates, as well as low predictive power. Wide confidence intervals mean that current understanding of factors governing woodland caribou herd dynamics is not sufficient for wildlife managers to make reliable projections of responses to management.