Friday, July 06, 2007

Cell Phones and Polling

I know that for the most part, no one who (ever read, never mind still) reads this blog cares about polling, but there's a very important post at pollster.com (formerly Mystery Pollster) on the effect of polling in the cell phone era. It's massively long and scholarly, but nevertheless very clear to a layperson.

You should read the whole thing, but here are a couple of the more interesting facts:
  • For the most part, pollsters don't call people on cell phones.
  • By the time the election rolls around in '08, fully 25% of Americans will only use a cell phone (up from 7% in '04).
  • Even this figure may be understated, as many of the people (like me, for example) who have a landline use it almost exclusively for DSL.
  • The people most likely to have a cell phone only are young people, people living in group housing, and poor people.
  • So far, the differences are shaking out so that the effects within age groups are not substantially different between cell/landline users (<2%),
As the cell-phone-only group grows, these subtle differences may become exaggerated. Or, possibly the cell-phone-only group will start to become the normative sample and even more closely resemble landline users. For now, not such a big deal, but worth revisiting in '09.

2 comments:

Unknown said...

I read it, I appreciate it, I care!

The key effort will be to analyze whether cell phone only users as a group respond to surveys differently. My old boss (who now is a mucky muck at Pew Research) pioneered studies on the difference among people who did not have a phone at any point in the last year. There was a measurable "non-phone bias," and particularly where those populations are being specifically targeted, that difference needs to be accounted for.

So the bottom line is that a growing number of people are becoming excluded from phone polling...but it's not yet clear whether they differ significantly as a demographic from those who still have landlines. And of course as the penetration increases, the group becomes more representative of the population at large--so who knows, the "problem" may end before it effectively starts.

cwilcox said...

My land line seems to afford some umbilical cord comfort to me because I just can't give it up. I suppose one day I will... and then look out, I'm goin' radical! If Bush and the court in Ohios have their way we won't have need for polls in the future anyway because the government will already know what we think from listening into our conversations.