You're right, of course. I realized that after I posted. Not to worry, I stressed in both emails that using exit polls is risky at best because of the margins of error inherent in the sampling.

Thanks for catching my mistake and letting me correct myself.