The primary objective of this study was to demonstrate the applicability of a J-based approach to fracture characterization of ductile alloys subjected to dynamic loading. Three-point-bend [SE(B)] fracture specimens of a high-strength steel were subjected to impact loading in a drop tower, and appropriate strains and displacements were measured to allow estimation of J. A finite element analysis of these experiments was conducted to evaluate the applicability of a J-based fracture characterization scheme and to provide guidance for the experimental measurements. The concept of a transition time was introduced to provide an estimate of the time during which the controlling crack-tip field [Hutchinson-Rice-Rosengren (HRR) field] stabilizes. The paper demonstrates that for fracture times greater than the transition time, J can be accurately evaluated by the quasi-static deep-crack formulation. For the case of the steel SE(B) specimen geometry used for this study, the transition time was approximately 300 s. The finite-element analysis indicated that the net moment carried by the remaining ligament of the specimen could be evaluated by measurements of the flexural and stresses at the quarter-span position on the SE(B) specimen. This was verified experimentally in drop tower experiments utilizing strain gages at the quarter-span position. These results were found to compare well with a load estimation scheme based on a full bridge of strain gages attached to the tension and compression surfaces of the SE(B) specimen. The load accuracy obtained using the full bridge was also found to be insensitive to crack extension and plasticity at the crack tip. Methods of determining the crack initiation point in these dynamic experiments were evaluated. It was found that the double capacitance crack-opening displacement (COD) gage used in the experiments provides an accurate estimate of the crack initiation point.