The item wording (or keying) effect is respondents’ differential response style to positively and negatively worded items. Despite decades of research, the nature of the effect is still unclear. This article proposes a potential reason; namely, that the item wording effect is scale-specific, and thus findings are applicable only to a particular measure involved in an investigation. Using multiple scales and several methods, the present study provides strong and converging support for this hypothesis. In light of the results, we reinterpret major findings of the item wording effect and propose possible future directions for research.