I'm sure with only 600 sampled the MoE is high and the number of junk polls is higher. Any that have moved 3 or more points without good reason (ie, not Pawlenty) are the crap polls.
600 actually isn't bad...4% MoE or so isn't it? Certainly acceptable.
Pawlenty could be a crap poll, but looking at trending, it wouldn't be my top choice.
The sa,ple size & MOE is fine, but when you have regional breakdowns you tend to get some strange results because the sample sizes tend to be quite small for them, they generally tend to even out (one area might be overstated, another area could be ubnderstated)