Exit Polls: A Look at Cell-Only Voters
Exit polling is notable after Election Day primarily for its massive store of data on voters. Among the unexplored numbers so far this year is new information about those voters who have abandoned their home phones and gone "cell-only."
Cell-only voters have at least doubled their share of the electorate since 2004, according to the network exit poll. They now make up 20 percent of Election Day voters, but as detailed below, the percent of all voters may be lower than that when early and absentee voters are included. Despite the increase, excluding cell-only voters from the Election Day horserace calculation appears to make little difference in the overall vote margin.
As a group, cell-only Election Day voters went for Barack Obama over John McCain, by a 62 to 38 percent margin. Those who have only traditional landline telephones were 53 to 47 percent, and those have both were evenly split, 50-50. (This analysis focuses exclusively on Election Day voters.)
As explored in detail in pre-election surveys, these apparent differences are almost entirely due to the different demographics of these groups, primarily that it's mainly younger adults who have made the switch.
First a basic point: Excluding cell-only respondents from the exit poll, the race among Election Day voters shifts from a six-point margin to a two-point contest. But readjusting the data so younger voters' proportion of Election Day voters stays steady - at 21 percent - takes it back to a five-point race. That's preliminary, but not definitive, confirmation that weighting by age mostly accounts for the lack of cellphones in many polls.
About three-quarters of cell-only voters are under age 45; 44 percent are under 30, and an additional 31 percent are between age 30 and 44. The age skew of these voters has eased somewhat since 2004, when nearly half, 48 percent, were under age 30.
Younger cell-only voters were more apt to vote for Obama than were older ones, just as was the case among all voters. Voters aged 18 to 29 who only have cellphones split 68 to 32 percent for Obama; it was 63 to 37 percent among those with landline service.
There was similarly little difference among voters age 45 and up based on what kind of phone service they have. Cell only voters split 50-50, while those with landline service broke 47 Obama to 53 McCain.
But one difference worth exploring is the significant telephone gap between voters ages 30 to 44: Cell-only voters in this group split 62 to 38 percent for Obama while those with landline service divided by a much slimmer 51 to 49 percent margin.
Unfortunately some of the data that would help understand this difference - e.g., marital status, parenthood and other factors associated with partisanship - were not asked on the same questionnaire in the exit poll. From what we do know, the cell-only members of this age group have less formal education, are more likely to be male and are more Democratic and liberal than their counterparts reachable by landline. None of those variations are apparent by phone usage for voters under age 30.
Overall, age is the biggest difference between cell-only and voters reachable on landlines, and that drives many of the other variations. Cell-only voters are more apt to be Democrats and more likely to consider themselves liberals. Twenty-six percent are black or Latino, compared with 20 percent of those with landline phones.
More than half of cell-phone only voters, 52 percent, report annual family incomes of under $50,000, compared with 33 percent of landline-reachable voters. About six in 10, 62 percent, do not have college degrees, that figure stands at 54 percent for landline voters.
They are also more likely to live in urban areas than other voters.
This analysis ONLY includes those who cast their ballots on Election Day. Since the exit pollsters did not conduct interviews with early and absentee voters by cellphone, only those with landline telephone access were included in pre-election surveys aimed at estimating the early vote.
The telephone poll of early voters did ask about phone usage, and at least 2 percent of respondents took the survey on a landline number they had ported to a cellphone. When these numbers are combined with the estimate of phone usage among Election Day voters, 14 percent of the full electorate said they are cell-only. If cellphones had been included in the early voter poll the number would likely have been higher.
In this year's Washington Post-ABC News tracking poll, we supplemented our landline RDD sample with a random sample of cell-only users - we will be parsing those data more thoroughly in the coming weeks..
Q: What type of telephone service is there in your home that you could use or be reached on?
Among 2004 voters
Among 2008 voters
The comments to this entry are closed.