By Brad Lendon
Advertiser News Editor

In a few months, the first primaries for the Democratic presidential nomination will be upon us, and we'll be inundated, as we are every four years, with opinion polls that supposedly will tell us the results of the voting before it occurs.

A story in the past week tells me that as an editor, I may want to look at those surveys with more skepticism than in the past. And a quick look at the methodology behind one recent poll showed me I need to be even more careful about what I put before Advertiser readers.

A story in Tuesday's paper said that beginning Nov. 24, consumers will be able to transfer their home phone number to their cell phones.

What that means to those of us who watch the political polls is that the sample that pollsters use is changing with the technology.

Nationwide, 7 million people now use cell phones exclusively, an industry expert said in the story. With the new ability to move home numbers to cells, 19 million more people will not have home phones.

Because home phone numbers are not listed, pollsters will find their samples reduced.

In a Chicago Tribune article a year ago, writer Steve Chapman reported that "the problem has gotten so big that Karlyn Bowman, a polling expert at the American Enterprise Institute in Washington, says, 'I expect that in 10 years phone interviews will be a thing of the past, replaced by Internet polling.'"

In the meantime, pollsters may find the remaining sample skewed, as cell-phone users tend to be younger and better educated than the population as a whole.

The issue came up on the Advertiser news desk this month as we talked about whether we should use some of our valuable space for a Pew Research Center poll on political party affiliations.

Copy desk staffers questioned its relevance and importance.

A few told me they would never be contacted by a pollster because they were among those 7 million who don't have a home phone. Others said they wouldn't take a pollster's call at home because their time was too valuable to waste taking a phone survey.

I decided to check the methodology of the Pew survey to see just how they did get their sample, and how they might account for these problems.

Checking their Web site, I found:

• The sample was indeed taken from "residential listings." Cell phones are not mentioned.
• Those residential listings included only numbers in the "continental United States."

Ah, the relevance problem.

At least in this poll, it would seem there wasn't much concern about our political party preference here in the Islands.

That doesn't mean you won't see polls in the Advertiser in the coming presidential year.

It does mean that we need to be certain that when we put those polls before you, you get them in the proper context, with the methodology noted.

We've often heard that numbers don't lie. But now we can see that they don't always give you the whole truth.