Causal questions require causal answers. Inferring causal knowledge from data is difficult, however. Unless causal assumptions are being made, statistical quantities cannot be interpreted causally. The popular phrase “correlation does not imply causation” summarises this observation and has fostered the use of careful terminology when interpreting the results of statistical analyses. Data science results are often presented in purely associational terms in an attempt to safeguard against making unwarranted (causal) claims. This approach may hinder scientific progress. First, even careful associational terminology may, between the lines, convey an unintended causal claim that misleads the reader, her future actions, and the design of follow-up studies. Second, it may not answer the original research question – a causal question cannot be answered by some answer to a different non-causal question. In this talk, we discuss how statistical causal modeling offers a way to make progress in answering causal questions by explicating causal assumptions, systematically incorporating the available data, and enabling a transparent discourse about the assumptions’ and claim’s validity and scope.