That's right. I just pubmeded. Had a student cite a few sources that referred to a study which was on pubmed. Turns out, unsurprisingly given the sources cited, that the pubmed study's data have been deliberately misinterpreted by the sources my student looked to. Nor did my student check what the original study actually found. Student might be surprised to learn that "study does not support the hypothesis that milk reduces chance of osteoporosis-related bone fractures" does not at all translate to "milk causes increased risk of osteoporosis."
We're going to have to have a day on evaluating sources once everyone gets back from Spring Break.