by Andrew Walden

Because of Hawaii’s three-week mail-in Special Election, political analysts have four mid-vote polls to look over as well as one poll taken just before the ballots were mailed.  When compared to the actual results of the voting, these polls show some very interesting patterns.

The polling numbers (see below) for Djou are very close, off by 3.6% at most.  Likewise for Case the numbers are off by 3.4% at most—all within the margin of error for these polls.  But Hanabusa’s numbers are off by as much as 13% and are always on the low side.  The May 6 & 7 Civil Beat poll come in the closest and the automated May 20 ccAdvertising poll was the furthest off—at least in predicting Hanabusa’s numbers.  

There are some fancy theories about polling discrepancies in Hawaii, for instance Dan Boylan told CQ in April:

Cultural sensitivity when doing surveys in Hawaii is so nuanced that one pollster commented that polling there is more like Japan than in any other part of the United States.

First of all, many survey participants — particularly Japanese-Americans — will say they are undecided when they are questioned about their voting preferences.

“And that’s not true,” said Dan Boylan, a political science professor at the University of Hawaii. “They just won’t tell a person with a disembodied voice on the phone how they’re voting.”

Japanese-American women, especially, tend to be underrepresented in polling because they decline to answer — a circumstance that Boylan argued could give Hanabusa an edge in the race.

“Let’s say there is 15 to 20 [percent] undecided, I would cut that in half in favor of Hanabusa,” Boylan said.

Boylan’s numbers are very close to the discrepancy between polling and the election result, but what is the cause?  What is the “cultural sensitivity”? 

Lets skip the mystical stuff about “Japaneseness” and consider the possibility that Hanabusa supporters—especially union members--are people who see themselves trapped in a web of retaliation.  Pollster ccAdvertising’s extreme undercount of Hanabusa votes indicates that up to 40% of Hanabusa voters clammed up.  The CivilBeat poll showed only 17% of Hanabusa voters feared to speak up.

This should be a logical possibility for anybody familiar with the reality of life in Hawaii.  Such people might be apt to protect themselves by remaining silent when the pollster calls because they do not know who might be setting them up for more punishment.  They are “sensitive” to the “culture of retaliation”—but so unclear about where it comes from that they vote for one of its leading protagonists. 

---30--- 

SPECIAL ELECTION RESULTS COMPARED TO POLLS IN REVERSE CHRONOLOGICAL ORDER

Djou      Case      Hanabusa

39.4       27.6       30.8    (actual voting results)

Poll May 20 CC Advertising

35.78     31.03     17.67

Civil Beat May 6 & 7

39.5       25.5       25.5

National Democrats released May 4

39          31           20

********************************************** Voting Begins May 1

Adv/HNN  April 23-28 

36           28          22

Daily KOS  April 11 - 14

32           29          27

National Democrats (leaked April 8)

32           32          27

Mason-Dixon (January 8-12)

17           37          25