Ion exchange membranes (IEMs) are versatile materials relevant to a variety of water and waste treatment, energy production, and industrial separation processes. The defining characteristic of IEMs is their ability to selectively allow positive or negative ions to permeate, which is referred to as permselectivity. Measured values of permselectivity that equal unity (corresponding to a perfectly selective membrane) or exceed unity (theoretically impossible) have been reported for cation exchange membranes (CEMs). Such nonphysical results call into question our ability to correctly measure this crucial membrane property. Because weighing errors, temperature, and measurement uncertainty have been shown to not explain these anomalous permselectivity results, we hypothesized that a possible explanation are junction potentials that occur at the tips of reference electrodes. In this work, we tested this hypothesis by comparing permselectivity values obtained from bare Ag/AgCl wire electrodes (which have no junction) to values obtained from single-junction reference electrodes containing two different electrolytes. We show that permselectivity values obtained using reference electrodes with junctions were greater than unity for CEMs. In contrast, electrodes without junctions always produced permselectivities lower than unity. Electrodes with junctions also resulted in artificially low permselectivity values for AEMs compared to electrodes without junctions. Thus, we conclude that junctions in reference electrodes introduce two biases into results in the IEM literature: (i) permselectivity values larger than unity for CEMs and (ii) lower permselectivity values for AEMs compared to those for CEMs. These biases can be avoided by using electrodes without a junction.