What about the more dangerous roads with high levels of fast traffic vs roads with moderate levels vs roads with low levels vs roads with very low levels vs roads with zero traffic?
There are other distinctions (subcategories, considerations, relevant criteria) when evaluating a road.
What about looking at a person's specific individual choices? -- what is available to the individual cyclist, where the cyclist is living or riding?
To speak of statistics that pertain to bike paths (in general) vs roads (also in general) is a lot less useful and a lot less meaningful, for most people who are facing actual specific choices for their own rides.
There are also many distinctions and subcategories within the broad categories of 'bike lanes' and 'bike paths' -- some of these are useful and relevant, and helpful in making evaluations.