> The problem with these studies is that it is exceedingly difficult to correctly account for all the additional differences between people who do get tattoos and those who don't.

I don't disagree to the gist of your post. However, in biomedical and public health research, papers are published even "if you don't find an effect", for good reasons. There could be simply very many unknown factors at the initial phase of some research topics. Researchers find or notice some things, publish them with rigorous discussions, proposed hypothesis with explicitly mentioned assumptions, etc. Other researchers build upon the existing results, add more discoveries, which can be proving or disproving, partially or wholly, etc. It often takes years of multiple teams to get a good enough understanding of a topic. This implicit collaboration is a positive feedback loop to advance the research.

The issue is that the vast majority don't read the detailed discussion in the papers and thus could get a partial thesis, which in many cases lead to incorrect conclusions. Media reports don't help, because they are essentially a simplified version. Otherwise, they can simply refers to the original papers and ask the audience to read them. Also, not many reporters have solid scientific training in the fields they report and don't understand the papers well enough. I don't blame them, since it is hardly their job. Good readers must be aware that the reports can be misleading, or biased, or simply wrong.

edit: grammar