Today, the Honolulu Star-Advertiser and Hawaii News Now released a Hawaii Senate poll showing former Gov. Linda Lingle (R) trailing her opponent, U.S. Rep Mazie Hirono (D), by 22 points. The poll was conducted by Ward Research:

If the election for the U.S. Senate were held today, who would you vote for?

I noticed this Hawaii poll and falsely assumed that it was probably one of only a few polls released for this contest. Generally I just ignore and archive data from the contests I'm not maintaining and move on, but today was an exception. I started perusing the release and noticed an annoying ad from the Lingle Campaign that cycled through "Poll Wrong", "Learn More" so I obviously clicked it.

The ad directed me to a Lingle for U.S. Senate press release complaining about the sample Ward Research used in the aforementioned poll. I have no idea if the sample is good or bad, because this information wasn't released. The main issue Lingle's campaign manager raised was that the sample skewed heavily toward Democratic participants. They then attached this comical memo, as evidence:

Memorandum
To: Lingle Senate Campaign
From: Jan R. van Lohuizen
Date: 10/26/12
Re: Star Advertiser polling

The difference between our polling and the Star Advertiser polling is in due [sic] to differences in the partisan composition of the sample. The Star Advertiser sample consists of 60% Democrats, 20% Republicans and 19% Independents. If we applied the Star Advertiser's distribution of partisanship to our data, we would be down by 18%.

We believe that the partisan distribution used by the Star Advertiser's polling is too Democratic. Here is the partisan distribution in the last 3 exit polls conducted in '06, '08 and '10 in Hawaii. As you can see the percent Democrats is nowhere near 60%:

	2006	2008	2010
Democrats	40	45	40
Republicans	23	20	22
Independents	37	34	37

Aside from the grammatical errors in the first and last sentence their entire premise is pointless. They argue that a better sample would have resulted in a 4% improvement in the topline results, which still translates to a crushing 18% loss. Furthermore, Ward Research clarified their 60% figure:

Ward has explained that the partisan distribution in the Hawaii Poll mirrors voting patterns in a state that has overwhelmingly favored Demo­crats. The poll asks voters which party they usually vote with, not whether they consider themselves Demo­crats, so a share of the 60 percent who said they usually vote Demo­cratic would likely describe themselves in other surveys as independents.

If you're Linda Lingle and you're losing by 22%, don't you have more important things to worry about than a sample that may be skewed by 4%.

Updated on October 30, 2012 at 7:12:18 PM CT