Researchers analyzed data from three surveys, concluding that direct measures of “trust” in scientists are unclear in what they represent.