In order for OPV devices to transition from the laboratory to the industrial scale, accurate measurements of device operating stability and lifetime are crucial. This paper compares the degradation of ITO/PEDOT: PSS/P3HT:ICBA/Ca/Al and ITO/MoO3/P3HT:ICBA/Ca/Al devices using the three main ISOS standard testing protocols: (a) ISOS-D-1, (b) ISOS-O-1 and (c) ISOS-L-1. We show that: (1) ITO/MoO3/P3HT:ICBA/Ca/Al devices are more stable than their PEDOT counterparts under the ISOS-D-1 protocol, as has been reported previously. (2) Under the ISOS-O-1 protocol, unencapsulated MoO3 based devices are more stable than the equivalent PEDOT device but, when encapsulated, the degradation rates of the MoO3 and PEDOT devices are the same. (3) By contrast, when measured under the ISOS-L protocol, the MoO3 based devices are either equivalent to (unencapsulated devices) or, indeed, actually degrade faster (encapsulated devices) that their PEDOT counterparts. We demonstrate that these differences arise from the dominant degradation mode changing under the different protocols. As such, this paper highlights that the choice of testing protocol significantly influences the reported stability of OPV devices. In particular, the ISOS-D and ISOS-L protocols do not necessary reflect OPV device performance under actual operating conditions and thus stability measurements using these protocols should be treated with caution. (C) 2016 Elsevier B.V. All rights reserved.