This paper presents an evaluation of overstrength based on an experimental study on large-scale dowelled connections in Cross Laminated Timber (CLT). In order to avoid brittle failure and ensure that ductile system behaviour and energy dissipation can be achieved under seismic loading, the overstrength of specified ductile components needs to be well understood. In timber structures, ductility is often achieved through plastic deformation of steel fasteners in connections. Overstrength is generally defined as the difference between the analytical strength, based on design-codes using characteristic material strengths, and the 95th percentile of the true strength distribution. Two main contributing factors to overstrength are the conservatism of analytical strength predictions, and the overstrength due to variability of material property distribution. In dowelled connections, further overstrength can be introduced if the yield strength of the supplied fasteners exceeds the yield strength of the specified grade. This is often not picked up during supply as the erroneous assumption is made that the stronger material performs better, and therefore acting in the best interest of the client. While this assumption is generally true for most non-seismic load cases, it can cause problems in capacity design as it introduces unexpected overstrength that is rarely accounted for. This paper evaluates the individual contributing factors of overstrength and compares experimental findings to theoretical considerations based on previous studies. It was found that unexpected steel fastener overstrength can contribute significantly to overall connection overstrength. However, the previously derived theoretical overstrength factor of 1.68 was safe in all cases.