Another poll, big difference

A Zogby Int'l poll; conducted 10/21-23; surveyed 500 likely Dem primary voters; margin of error +/- 4.5% (release, 10/24).

WH '04 Dem Primary Matchup
All Dem Ind Men Wom NH01 NH02 9/25 8/26 6/7 Fav/Unfav
Dean 40% 43% 35% 33% 46% 38% 42% 30% 38% 22% 77%/ 11%
Kerry 17 20 11 17 16 15 17 20 17 25 66 / 26
Edwards 6 5 7 8 4 4 7 2 4 2 50 / 14
Gephardt 4 4 4 5 3 5 3 6 6 7 49 / 32
Lieberman 3 1 7 4 3 3 4 5 6 10 44 / 43
Sharpton 1 - 1 1 - - 1 1 - - 13 / 57
Kucinich - 1 - - 1 - 1 1 1 2 16 / 23
Braun - - - - - - - - - - 21 / 27
Clark 6 8 4 12 2 8 5 10 2 n/a 38 / 23
Oth/Undec. 23 18 31 22 25 27 20 25 24 32

I really don't get this poll, if three other polls have Dean between 35% and 25% and Kerry at or around 18% with Clark in third, why does this one have Clark tied for 3rd/4th with Edwards (who hasn't been on the map really) and Dean at 40?

Seems to be either way off, or the other polls aren't smelling the Dean rout. Wasn't Zogby the guy that did polling for Jeanene Shaheen's ill-fated senate race in 2002, with polls that suggested it was close when in reality it was called at the close of the polls (like over 10 points)? Or was Zogby behind some other terrible numbers for 2002 polling...Maybe it was that "close" Colorado Senate race? All I can say is, I really doubt Dean is up by that much.