GRIH poll interpretation overstated opposition to Akaka bill

The Hawaii Reporter has a piece today looking in more detail at the poll on the Akaka bill that was commissioned by the conservative Grassroot Institute of Hawaii (GRIH).

It notes an important aspect of how the interpretation of the poll was released, that was picked up on by the Advertiser but I didn't notice until now.

The Honolulu Advertiser also calculated the poll results on a different basis than Grassroot by including the "no response" respondents in the total and writing, "Of the 980 respondents, 198 (20 percent) said they support the Akaka bill, 404 (41 percent) said they do not and 378 (39 percent) gave no response."

The GRIH release states that "67.11 percent of all respondents oppose the Akaka Bill."

Ken Conklin's website with the full poll stats and questions is headlined with the even more generalized "67% of Hawai'i's people oppose the Akaka bill."

In fact, 67% percent of respondents to this particular question said they were opposed to the bill, but only 41% of respondents to the poll overall said they were opposed. 39% responded to the poll but gave no response to this question. (And now having taken this poll myself—see below—I know that there wasn't an explicit option for other or undecided, one had to just remain silent after the question in order to register no response.)

There is a big difference between 67 and 41 percent! And that is not even taking into consideration the issues of how the questions were phrased and arranged.

Now I see that not only was the poll structure misleading, but the interpretation of the results was even more inaccurate and misleading.

Richard Rowland from the GRIH has a letter in the Hawaii Reporter just today saying "two out of three respondents opposed the Akaka bill. It appears that our Governor needs a reality check when it comes to priorities." It appears Mr. Rowland needs a reality check when it comes to interpreting poll results.

A couple other points/questions about this poll.

Why was political party ID slightly higher Republican?
Why was the first question about taxes?

And are these two questions related? The results show that 34.42% of respondents were Republican, and 33.56% were Democrat. I don't have the exact party ID for Hawaii at the moment, but I am quite certain that it is more D than R. I'm just speculating, but the only reason I can think to start a poll on the Akaka bill with a question on taxes is to skew the sample. Maybe those who are Republican are more likely to respond to a question about taxes, and Democrats are more likely to hang up. Only 17% of people called responded to the poll at all, and I think given that the respondents are in fact skewed Republican compared with actual party ID, that could be explained by the opening question. The results released don't provide the breakdown of responses by party ID, but regardless of the reasons why party ID is skewed, I bet if you weighted the poll by party ID so it reflected actual party ID in Hawaii, the opposition would end up being even lower.

For the record, I am not in favor of the Akaka bill, and I agree with the criticism that OHA's poll, and their interpretation, was also misleading, and despite its own biases this poll does throw doubt on claims by OHA of overwhelming support for the Akaka bill. But that doesn't justify GRIH misleading in the other direction. I'd prefer more light and less heat in this discussion. The truth is probably somewhere in between. (I do give Hawaii Reporter credit for at least being willing to point out how the Advertiser provided a different calculation.)

Finally, I wanted to note that last week I got called from this poll! It was pretty interesting to actually listen to the questions and the way there were phrased, and the impression that gave, rather than just reading it in print. But more noteworthy is the fact that the poll supposedly was "conducted by FEC Research between June 29 and July 1." Why in the heck is the poll calling still going on a week after the polling was suppose to be pau? Unless they just forgot to turn off the autodialer, the only explanation I have just reinforces the idea that I originally suggested that this is a push-poll, a poll intended to influence as much as gauge opinion through the way the questions are phrased, and they still had/have it running for that purpose.

On a related note, KHON had an item yesterday about the bill nearing a vote in the Senate.

Posted: Mon - July 11, 2005 at 11:50 AM

getBlogName(); > getCategoryName("C608676235"); >

GRIH poll interpretation overstated opposition to Akaka bill

> >