You can't get causation from an observation study.

You can however get a lot of attention for "finding" a causation.

So to answer your question, it's quacks selling correlation as causation.