Power system expansion models are a widely used tool for planning powersystems, especially considering the integration of large shares of renewableresources. The backbone of these models is an optimization problem, whichdepends on a number of economic and technical parameters. Although theseparameters contain significant uncertainties, the sensitivity of power systemmodels to these uncertainties is barely investigated. In this work, we introduce a novel method to quantify the sensitivity ofpower system models to different model parameters based on measuring theadditional cost arising from misallocating generation capacities. The value ofthis method is proven by three prominent test cases: the definition of capitalcost, different weather periods and different spatial and temporal resolutions.We find that the model is most sensitive to the temporal resolution. Fur-thermore, we explain why the spatial resolution is of minor importance andwhy the underlying weather data should be chosen carefully.