New ARG Polls in Iowa, New Hampshire, and South Carolina

American Research Group (ARG) has released new polls in Iowa, New Hampshire, and South Carolina.

Here is a summary of the results from Political Wire...

Iowa:

• Among Democrats, Clinton 30%, Edwards 21%, Obama 15%, Richardson 13%
• Among Republicans, Giuliani 22%, Romney 21%, McCain 17%, Thompson 13%

New Hampshire:

• Among Democrats, Obama 31%, Clinton 31%, Edwards 14%, Richardson 7%
• Among Republicans, Giuliani 27%, Romney 26%, Thompson 13%, McCain 10%

South Carolina:

• Among Democrats, Obama 33%, Clinton 29%, Edwards 18%
• Among Republicans, Giuliani 28%, Thompson 27%, McCain 10%, Romney 7%

• 
  

In Iowa Clinton falls 2% since June, but moves ahead of Edwards, who fell 8%. Obama gained 2%, but the big mover is Richardson who is up 8%. There was big movement in how people who identified themselves as being no party viewed the candidates. Edwards support among those who identified themselves as no party fell from 39% in June to 8% in July. Clinton's support went from 33% to 18%. Obama's support went up from 11% in June to 21% and Richardson's support went from 3% to 25% among people identified as no party. (Here's analysis of the June's poll.)

The poll numbers in Iowa don't mesh well with other polls in Iowa. ARG is the only poll that has had Clinton polling in the low 30's. However, I think it is worthwhile looking at trends in the ARG polls to see how candidates are moving.

The most interesting news in these polls however is Obama moving into basically a tie with Clinton in New Hampshire and leading in South Carolina. In New Hampshire, Obama was down 34% to 25% in June and now they are tied at 31%. In South Carolina, Obama made an even bigger jump. In June Obama was down 37% to 21% and now he is leading 33% to 29%. Though the media praised Hillary after the CNN/YouTube debate in South Carolina, it seems Democrats in South Carolina were more impressed with Obama.

**Update**
Jonathon Singer of MyDD has analysis of the polls here where he talks about how accurate ARG polling is.

I'm a bit reluctant to put up numbers from American Research Group because they don't jibe so well with other polling (my sense is that they're reaching far too wide of a population, not a relatively narrow primary/caucus population), but for whatever it's worth, here are the latest ARG numbers out of the early nominating states (along with June results and Pollster.com averages prior to these latest polls).

	IA		NH		SC
Candidate	ARG	Pollster	ARG	Pollster	ARG	Pollster
Clinton	30 (32)	23.5	31 (34)	34	29 (37)	41.4
Obama	15 (13)	15.6	31 (25)	22.7	33 (21)	26.1
Edwards	21 (29)	27.5	14 (11)	9.4	18 (22)	14
Richardson	13 (5)	12.5	7 (6)	10.1	2 (1)	2.4

Again, let me stress my general sense that ARG casts too far a net in these polls, that it's quite possible that the numbers for the potentially more narrow population that represents likely caucus-goers and primary voters would be different. Indeed, the ARG numbers don't look exactly like the Pollster.com averages posted above (Iowa, in particular, seems off from the general consensus from other polling outfits).