Measurement of the fracture toughness of steel is important for the assurance of the safety of ships and offshore structures, especially when these structures are made of thick sections and/ or applied in cold environments. One key factor that will affect the determination of the fracture toughness is a pop-in, which is a short event in which unstable fracture is initiated and then selfarrests. If the pop-in is large enough, it will be used to calculate the fracture toughness. Pop-ins are believed to be the products of local brittle zones, which occur randomly at crack tips and have finite sizes. Fracture toughness testing codes have ways of determining whether a pop-in is critical (thus, identifying the maximum force and displacement to be used in the determination of the toughness of the material) or not important (thus, allowing for the test to proceed). In an ongoing project on the use of small-scale fracture specimens to predict standard fracture toughness test results, we would like to know how pop-in acceptance criteria should be scaled for specimen size. It is expected that the physical size of the brittle zones that cause pop-ins is invariant of specimen size, meaning that the contribution of the pop-in will be proportionally more important for smaller specimens. An analytical method for relating the pop-ins on one specimen size to another specimen size is developed. This method is partially verified by observations on the size of a local brittle zone observed on a fracture surface and the effect of that pop-in on the force-displacement curve during a CTOD test. The analytical method showed that an equivalent pop-in for a small-scale specimen is indeed larger, but that the effect was subtle.