In this study, we investigated the effectiveness of two textual interventions designed to prime people to adopt more privacy-protective behavior. We asked study participants how likely they are to share photo memes in three experimental conditions. In addition to the control condition, we had two experimental conditions where we incorporated textual priming by asking them first to i) imagine themselves as the subjects of the photos, and ii) consider the privacy preference of the subjects in the photos. To our surprise, we found that sharing likelihood was significantly higher in the two experimental conditions than the control condition, even though the experimental manipulation was designed to prime them towards lower sharing. To explain this counterintuitive result, we conducted a follow-up study where we asked the participants to explain why they had made a certain sharing decision. Their responses hint that there may be some disconnection between how people conceptualize privacy and how it is defined in the literature. More research is required to fully understand how people make decisions regarding privacy management and how to intervene effectively to help them make a better decision.