AbstractThis study presents a classification of stratospheric extreme events during northern winter into events with or without a consistent downward propagation of anomalies to the troposphere. Anomalous strong and weak stratospheric polar vortex events are detected from daily time series of the polar cap averaged (60°–90°N) geopotential height anomaly. The method is applied to chemistry‐climate model data (E39CA and WACCM3.5) and reanalyses data (ERA40). The analyses show that in about 80% of all events no significant tropospheric response can be detected. The stratospheric perturbation of both weak and strong events with a significant tropospheric response persists significantly longer throughout the stratosphere compared to the events without a tropospheric response. The strength of the stratospheric perturbation determines the strength of the tropospheric response only to a small degree. Results are consistent across all three data sets.