Great graphs—I had to face this issue many decades ago when working on a college paper.
I was easily able to prove correlation between things, but for many reasons I was convinced that no real causation was shown.
Most academic papers using statistics are just wrong—for exactly this reason.
You are one smart cookie. I mean that sincerely. Yes, hence the axiom “correlation does not equal causation”. Studies that look at correlation can be helpful insofar as they might point to an need for further research, though.
For example, if outbreaks of a particular tick-borne disease occur only or mainly where a certain species of mammal is found, that mammal may or may not be a natural reservoir. This correlation does point to the plausible reason for further investigation, though. Researchers go into the field, collect and test these mammals, and find they indeed carry the disease and are also infested with the same species of tick already known to carry the disease. Bingo. Or not — then start over and look for another species.
Proper studies looking at correlation plainly state they prove nothing but only point to a potentially promising area for further research.
Yes, you know what they say about “lies, d*** lies and statistics”. Amateurs and fraudsters using creative math (or faulty data) can sure churn out some whoppers! Even professionals sometimes err. That’s why it’s so important to scrutinize such studies. Statistical analysis can be very useful when properly applied to proper data sets. Otherwise, GIGO. And there’s lots of that floating around the web.
I always approach a study with skepticism, even if it’s in a reputable journal. Everyone should. Some check out, others don’t, while still others point to as-yet-unproven possibilities.