This paper presents an experimental and numerical study of the effect of specimen thickness on the effective notch toughness Kmatρ for cleavage fracture measured using Single Edge Notch Bend (SENB) specimens containing a Unotch instead of a fatigue pre-crack. These specimens are typically used to measure a material’s effective notch toughness Kmatρ and to assess failure of a structure containing a non-sharp defect using the Notch Failure Assessment Diagram (NFAD). Both the experimental data and the Finite Element (FE) failure predictions show a significant influence of specimen thickness on Kmatρ, over and above the microstructural weakest link effect arising from differences in the volume of the plastic zone. Kmatρ is a function of not only the in-plane effect of the notch radius, but also an out-of-plane constraint loss which itself is enhanced by the presence of the notch radius. Significant out-of-plane constraint loss occurred for notched specimens with a ratio of thickness B to width W of 0.5, a geometry that if pre-cracked would have met the minimum thickness requirement mandated by ASTM E1820. Doubling the thickness to B/W=1.0 was sufficient to eliminate the out-of-plane constraint loss. The use of experimentally measured Kmatρ values in an NFAD assessment of a structure may therefore be non-conservative if B/W<1.0