"McGrath, please tell me where to find the research that shows that hand counts are more accurate than machine counts."

Actually there's a methodological problem here. The only way of checking whether a count is accurate is to do a recount. That's either got to be a hand recount or a machine recount. If you do them both, and they come up with a different result, which is the more accurate one? What do you measure it against?

I know that if my life depended on it, I'd sooner trust a careful count carried out by human beings than a machine. And I do not think I am cynical in believing that, whichever candidate in an election like that in Florida had come in a few hundred votes behind, they'd have been asking for a hand recount.

I'm not cynical enough to assume that whichever candidate came in a couple of hundred votes ahead would have been opposed such a count. And I'm not naive enough to assume they would not have opposed it. But either way, I can't see that being opposed to a full manual recount would have been consistent with a proper regard for the democratic process.)