>If this data is taken from exit polls how correct is it when the
>exit polls got the result so wrong ?
THe early exit polls were wrong; the complete ones were closer to accurate. But an error of just a few percentage points (which is pretty good for sampling work, really) in the topline results doesn't invalidate the usefulness of all the other analytic stuff. The patterns are almost certainly correct even if the exact numbers might be off by a couple of points.
Doug