The core self-evaluation scale (CSES) is the most popularly used measure of core self-evaluations. However, recent research has highlighted the potential existence of a negative wording factor (NWF) associated with the negatively-worded items of the CSES. Discounting the effects of a NWF may lead to biased estimates of the reliability and criterion-related validity of the CSES. The current research investigated whether the NWF found in previous research could be replicated in three Australian samples, which consisted of both students and adult community volunteers. Results revealed that the NWF provided a better model fit than the unifactor model but was inferior to an alternative model which measured a positive wording factor (PWF). However, the NWF model possessed equivalent composite reliability to the unifactor model in Study 1, but slightly superior composite reliability in Study 2 and Study 3. In addition, the NWF possessed poorer criterion validity estimates for both life satisfaction and career satisfaction. Contrasts to previous research which has substantiated the claim of a NWF in the CSES, and recommendations for research and practice are discussed.