Science-related news stories can have a profound impact on how the public make decisions. The current study presents 4 experiments that examine how participants understand scientific expressions used in news headlines. The expressions concerned causal and correlational relationships between variables (e.g., “being breast fed makes children behave better”). Participants rated or ranked headlines according to the extent that one variable caused the other. Our results suggest that participants differentiate between 3 distinct categories of relationship: direct cause statements (e.g., “makes,” “increases”), which were interpreted as the most causal; can cause statements (e.g., “can make,” “can increase”); and moderate cause statements (e.g., “might cause,” “linked,” “associated with”), but do not consistently distinguish within the last group despite the logical distinction between cause and association. On the basis of this evidence, we make recommendations for appropriately communicating cause and effect in news headlines.