Yeah, that's closer to the definition I'm familiar with. It's a species of "correlation does not imply causation".

What the author is talking about here seems to be how to recognize and avoid confirmation bias.