I was responding more to your "This test is crap" and subsequent explanations - though I'd consider "children can't have a model of the world that contains a person being wrong about a fact" as a reasonable first approximation for young enough children.