Research suggests we that we trust the contents of secret documents more than the unclassified stuff:
Specifically, “(a) people weigh secret information more heavily than public information when making decisions, (b) people perceive the same information as being of higher quality when it is portrayed as secret rather than public, and (c) people evaluate others’ decisions more favorable when those decisions are based on secret information rather than on public information.”
The authors [of the study] suggest this effect was very much in evidence following the original Wikileaks disclosures, when people who are normally skeptical about the judgments of U.S. officials were suddenly taking as gospel documents written privately by those very same officials. As Dan Drezner wrote at the time, there was a “natural inclination to think that any Wikileaks document will endow it with the totemic value of Absolute Truth. “If it was secret, then it must be true,” goes this logic.” In fact, it’s quite possible for diplomats or military commanders to be as wrong in private as they are in public.
It’s also not hard to imagine the kind of incentive structure that this bias creates. If classified information and secret recommendations are judged to be more credible, there’s going to be a natural tendency for the officials crafting and passing along this information to keep as much of it secret as possible.