Well, if it were true that your beliefs were wildly at odds with evidence, that would be very foolish of you. But that individuals have (effective) agency is so blatantly obvious that we don't bother testing it most of the time. Miming your social group by default still leaves a great deal of agency, and people can choose to execute that agency even if often they don't. For instance, many studies show a sizable shift in support for a statement on the basis of who they think also holds it; but usually those shifts, though large enough to matter considerably, are less than individual variation. Much of that variation can be ascribed to agency.