I take issue with the use of the word "scientific". Surveying marijauna users and asking them if it has been a gateway drug for them is about as scientific as any of the polls on MMAF.
I'm talking about a truly scientific study with a testable hypothesis, exactly one dependent variable, a repeatable design, and peer review.
Everything that lacks those things is pseudo-science and means next to nothing to anyone in the scientific community.
You could say that about all the social sciences.
And you'd be right.