The usefulness of elastic compliance measurements to estimate crack closure in structural steel and the validity of the assumption of a constant compliance value for the fully open crack is examined. Based on considering different issues related to the experimental technique and compliance data processing, local compliance measurements and the compliance offset method recommended by the ASTM standard are selected to be most suitable for structural steel. The compliance data generated in fatigue tests on I 8G2A steel conducted under a variety of loading conditions enabled to choose an optimal strain gauge positioning and appropriate offset criterion values for the original compliance offset method and its modified (normalized) version. The adequacy of the closure measurements is assessed through checking the ability of the resulting effective stress intensity factors to account for the observed effects of the loading conditions on fatigue crack growth rates.