3. We’re almost in 2018, folks
With rare exception will we be interested in stats from more than five years ago. One memorable draft included findings from a research study in 1950 to support the point.
Too limiting, in my view.
Stats from any study remain valid until contested by a more recent study done under sound statistical standards, UNLESS there is reason to dispute them for other than age. (For example, quoting economic statistics from the pre-information age and/or pre-globalization period to back current policy are questionable, simply because we don’t live in that world anymore. )