This methodology study investigated the application of a miniaturized wedge splitting test (WST) accompanied by digital image correlation (DIC) for the novel analysis of fracture mechanisms occurring during fracture of quasi-brittle refractories. Results from representative miniaturized WST were compared to standard WST results. Accuracy was improved by excluding experiments where unstable crack growth occurred or when the crack propagated perpendicular to the notch direction. Besides a sample size effect, a limitation of the fracture process zone (FPZ) size occurred. This FPZ limitation reduced the specific fracture energy G‘F from 112.4 N/m for the standard WST to 55.4 N/m for the miniaturized WST. Due to the sample size effect, the brittleness number was 0.194 and lower than for the standard WST with 0.414. The miniaturized WST enabled a more detailed evaluation of the acting mechanisms by analyzing the load–displacement curves together with the DIC. Grain bridging, microcracking, crack deflection/branching mechanisms and friction effects were evaluable. The analysis of the miniaturized WST could prove the supposed mechanisms acting during the standard WST. However, not all observed mechanisms also occurred in the standard WST due to the difference in the samples' brittleness.