That's not what I said. I think the study they did is suspect.

It would be like a study that shows that 80% of traffic fatalities, drivers use of soft drinks or coffee was present. (I estimated the percent, but would guess it about right)
Then conclude from this that caffeine causes driver impairment.

One can't throw a couple of sets of data together and then claim one affects the other. I believe they even indicated this in the link.

There are 10 kinds of people.
Those that understand binary and those that don't.