The cellphone-only crowd is not yet large enough and their views not different enough to affect the accuracy of traditional political polling, a new study suggests.

Not yet, anyway.

The U.S. report, based on polling by the Pew Research Center and The Associated Press, reaffirmed that those with only a cellphone tend to be younger, less affluent, more likely to be male and more likely to be minorities.

But when the responses to a series of political questions from those with only cellphones were blended in with the rest of those reached on traditional landlines, results were not affected.

Very small differences in the results virtually disappear when the cellphone sample is blended together, and weighted to match the demographics of the U.S. population.

These results are consistent with an AP-Pew poll done earlier this year and other surveys.

About one in 10 adults in the United States have only a cellphone, based on projections of growth from research, including that done by the Centers for Disease Control and Prevention.

Research done during the CDC's National Health Interview Survey suggests the number of cellphone-only households has been growing by about one percentage point every six months over the last few years.

In Canada, there were 16.6 million cellphone subscribers by the end of 2005, Statistics Canada reported. The federal agency said it has observed a growing trend that's seen many Canadians replacing their traditional home telephones with cellphones.

The growth of the cellphone-only group has raised concerns among survey researchers that it would render obsolete the most commonly used polling method of contacting a random sample of the public on traditional landline phones.

Growth of the cellphone-only group may eventually reduce their differences from the overall population, said Scott Keeter, a survey researcher at Pew.

The biggest differences found between landline phone users and cellphone only users involved their levels of political engagement.

Just 49 per cent of the cell-only users in the U.S. were registered voters, while 78 per cent of those in the landline sample were registered voters. That's another reason the cellphone only group has a minimal effect on political polling.

The study was based on a poll of 2,004 adults, including 1,804 with traditional landlines and 200 with only cellphones. The registered voter sample included 1,503 registered voters with landlines and 97 registered voters with only cellphones.

The poll was conducted from Sept. 21 to Oct. 4 and had a margin of error of plus or minus three percentage points for both samples, adults and registered voters.