The rising share of converter-interfaced generation displaces synchronous machines and their grid stabilization characteristics. These leads to the increased need of grid-forming (GFM) converters unlike the today most common used grid-following (GFL) converters. Reliably distinguishing of the two control strategies purely on measurements at the point of connection remains challenging: prior studies offer either measurement-based classification without analytical depth or analytical references without experimental validation. This article closes that gap through an analytical space-vector evaluation of the voltage and current responses to defined grid voltage disturbances, validated by simulations and laboratory measurements across different GFM and GFL control structures. The analysis identifies an inherent decaying DC current component in GFM and ideal voltage source behaviour, which is not present in GFL or ideal current sources. It further exposes the limitations of fundamental-frequency analysis relative to time-domain assessment. The results show that within the first milliseconds of a disturbance, GFL and GFM behave fundamentally differently, and that the proposed space-vector based method enables robust, reliable classification for future testing procedures.