Questionable new NRSC poll

A new poll released yesterday by the National Republican Senatorial Committee (NRSC) suggests a statistical dead heat in the CO Senate race, far narrower than was found in polling done in mid august. PolitickerCO reports:

In the U.S. Senate race, Democrat Mark Udall held a one-point lead over Republican Bob Schaffer, the poll found. The poll, conducted among 495 "likely" registered voters in Colorado between Sept. 2 and Sept. 3, has a margin of error of +/- 4.5 percent.

It's important to take into account however that the poll was commissioned by the the Republican hill committee working to elect Bob Schaffer, and that unlike any reputable polling agency (such as Rasmussen or PPP) it failed to provide ALL the essential information used to judge such a poll's acuracy:

1. The wording of the questions asked

2. Whether these "likely voters" were actual registered voters

3. Any data about how representative the sample actually was, i.e. about how the sample was proportionally composed in terms of party affiliation

I'm skeptical about how much weight should be given to this poll on these grounds alone. But even if we give the NRSC the benefit of the doubt, I think that the daft conclusion reached by RealClearPolitcs' Reid Wilson (and likely many mainstream news outlets to come) that "despite the free press [provided by the DNC] Rep. Mark Udall didn't benefit at all" is simplistic at best.

Especially given that this poll was conducted smack dab in the middle of the RNC, and 4 days after the Palin veep pick. The temporary bounces that result from both events effect the whole of the republican party, and its effects would inevitably trickle down to this poll to a greater or lesser degree.