Of 100 numbers on the list, I was able to speak to 23 people. Of these, 15 said they will vote for Webb, 1 said leaning to Webb, 3 undecided, 3 leaning to Miller and only 1 was a definite Miller vote.
Furthermore, as I worked through the list over the past two days, the statistics seemed to shift. The first part of the list is where the 'leaning to Miller' and 'definite Miller' votes occurred, and the last part of the list was almost completely 'voting for Webb'. My bet is that a formal test would show statistical significance, even though the numbers are small. I have seen this sort of pattern before when doing canvassing -- a lot of people make up their minds on Sunday.