Hawaii Senate primary: Which poll to believe?

The matchup for Hawaii’s open Senate seat will officially be set Saturday, when the state holds its Democratic Senate primary.

And depending on which poll you believe, it’s either going to be a barn-burner or a blowout.

Former congressman Ed Case is challenging Rep. Mazie Hirono for the right to face Republican former governor Linda Lingle in the general election, and he’s long been the underdog. Hirono had the unofficial backing of the national Democratic Party, raised tons more money than Case did, leans further left than Case does, and according to some polls, she carries a double-digit lead into Saturday’s election.

As usual, though, polling in the Aloha State has been all over the place — with opposing polls in the last week showing Hirono up by 17 points and Case up 1 point.

Former Hawaii congressman Ed Case (D) holds a sign as he campaigns in a special election for an open House seat. Case is now running for Senate but is an underdog in Saturday’s primary. (Marco Garcia — Associated Press)

So whom to believe?

Here’s the rundown:

* An internal Hirono poll conducted by Benenson Strategy Group (which also polls for President Obama) this week showed her leading by 17 points in the primary.

* An independent poll for the state’s Civil Beat Web site showed Case ahead by a nose, 47 percent to 46 percent. The poll was done by the Merriman River Group, which has also polled for Case.

* A Hono­lulu Star-Advertiser poll, conducted by Ward Research in mid-July, showed Hirono ahead 55 percent to 37 percent — similar to her polling.

* But a poll around the same time conducted for Lingle’s campaign, by Voter/Consumer Research, showed a tight race in the Democratic primary, with Hirono up 1 percent.

So in sum, Hirono’s pollster and Ward Research show her up big, while Civil Beat (using a pollster who has worked for Case) and Lingle’s pollster show the race neck-and-neck.

And really, it’s not a surprise; Hawaii has always been very hard to poll, with certain demographic groups in the state less apt to participate.

Hence, there has been plenty of consternation about these polls.

Hirono advocates note that the Civil Beat poll includes fewer Democrats than other polls (good for Case), more of Oahu (good for Case) and skews old, with 81 percent of respondents being 50 years or older (very good for Case).

This is probably because the poll is automated, which means cell phones are not included. The Washington Post’s policy is to report both types of polls, but traditional live-interview polls (like the other three polls listed above) are generally viewed as more accurate, because many younger voters only have cell phones, for instance. In this case, older voters, who tend to favor Case, may be over-sampled.

In addition, the partisan tilt toward independents may be a result of the pollster’s question, which suggested that the Democratic primary was more competitive than the GOP primary. That’s undoubtedly true, but it may have resulted in more independent respondents opting to pick a candidate in the Democratic primary than the GOP race.

(Note: Civil Beat has a great breakdown of its poll that compares it to Ward’s polling and mentions many of these points. It’s worth a look.)

On the flip side, Case’s campaign argues that Ward has gotten primaries wrong before. In the same matchup in the 2002 Democratic governor’s primary between Case and Hirono, Ward showed Hirono leading him by 20 points three weeks before the primary, and she only won by two points. Ward also polled a five-point lead for now-Gov. Neil Abercrombie the month before the 2010 Democratic governor’s primary, and he wound up defeating former Hono­lulu mayor Mufi Hannemann by 21 points.

In both cases, in a primary between a white candidate and a Asian-American candidate, the white candidate performed much better on primary day than the Ward poll suggested a few weeks prior.

Just before the 2010 primary, meanwhile, a Merriman River Group poll for Civil Beat showed Abercrombie ahead by 17 points — close to the actual margin.

(Now, Abercrombie and Case might well have gained ground in the final weeks of the campaign in 2010 and 2002, respectively, but, to borrow Chris Cillizza’s favorite phrase: It is what it is.)

As far as the other polls go — the internal Benenson poll for Hirono’s campaign and the Voter/Consumer Research poll for Lingle’s — there isn’t as much of a Hawaii track record and the polls don’t include crosstabs, so it’s hard to say where they might be better or worse.

What we do know is that the best polls for Case show him in a tight race, while the best polls for Hirono show her leading by a large margin. If the truth is somewhere in between, that means Hirono wins.

In addition, while Abercrombie might have outgunned Hannemann down the stretch in 2010, it’s harder to see how Case might have done that this year, especially given his lack of funds (Hirono had 10 times as much cash on hand as of July 22). Even if the race is close poll-wise, Hirono has a better organization behind her, which is worth something on primary day.

Either way, this isn’t the end of divergent polling in Hawaii. Republicans have high hopes for Lingle in the general election, but polling there has varied widely as well, with her own poll showing her ahead and Ward and Hirono’s pollster showing the Democrat ahead by nearly 20 points.

We’ll see which pollsters win this round and then re-evaluate the general election.