October 17, 2004
More on Mobile Phones
I want to clarify a few points from Friday's post on mobile phones, based on some of the very thoughtful reader comments.
First, it is tempting to try to project a growth rate for mobile-phone-only adults based on the three different studies I cited, but keep in mind that each used different methodologies and asked different questions. The estimates are not strictly comparable with each other. We do have pretty good estimates of the size of mobile-phone-only population as of last spring and clear evidence that the population is growing. So it is safe to assume the mobile-phone-only population has grown since the spring, but we do not know precisely how much.
Second, keep in mind that we really cannot know for certain (a) how many mobile-phone-only adults will vote and (b) who they will vote for. My point was that the available demographic estimates of the mobile-only population suggest a Democratic orientation, but the same demographic pattern also suggests they are historically low turnout voters. Even if we assume comparable turnout and a plausible pro-Kerry margin, their relatively small size still implies a very small effect. Of course, this inference is a matter of opinion. Yours may differ. Strictly speaking, we do not know for sure.
Third, many agreed with commenter Haim Goldman, who asked, "what about people who have both home landlines and cell phones but rarely answer the landline?"
Good question. We do have some recent data on adults with both mobile and landline phones. It comes from another study reported earlier this year based data from large in-person surveys (approximately 2000 interviews per year) conducted over the last three years by the Roper ASW:
|RoperASW Data* - Percentage of All Adults|
|Percentage who own a cell phone||50%||55%||60%|
|Use cell phone to make…|
|…three quarters of calls||6.0%||7.7%||9.7%|
|Uses only cell phones to make|
|and receive calls||na||1.9%||2.5%|
The last category (those who use only cell phones to make and receive calls) shows a smaller number of mobile-phone-only adults (2.5%) than suggested by the number of households in the same category reported at about the same time by the Current Population Study (6.0%). The difference is likely a function of different question language. The Roper number (2.5%) is a byproduct of a question that asked respondents about the likelihood of abandoning their landline phones. On that question, another 4% of all adults said it was either "almost certain" or "very likely" that they would give up regular phone service.
However, the data in the middle rows are most useful to our discussion. The 6.6% who said they use the cell phone to make "all" calls is close to the CDC estimate of cell phone only adults. The main point is that the Roper data show that another 9.7% who say they use their cell phones to make most of their calls.
What is the difference? A household without a landline is totally out of reach of conventional telephone surveys. A household with a rarely used landline is not. In terms of survey methodology the first is an issue of "coverage," the second of "non-response." There is little a pollster can do about a coverage problem, but there are strategies for reducing non-response. I will discuss this at more length in a few days, but it is mostly about persistence, calling respondents several times over several days to get those who are more difficult to reach, whether it is because they seldom use the landline, are rarely home, sometimes screen their calls with an answering machine or caller ID, or are simply on the phone a lot (yielding a busy signal).
Don't get me wrong: The non-response issue (which also includes those who simply hang up) is definitely a growing problem for polling, and a potentially much bigger problem this year than the relatively small number cell-phone only households. Persistence is not a magic answer. Non-response just a different category of problem, and one that I will take up in a few days.
*Source: Peter Tuckel and Harry O'Neill, "Ownership and Usage Patterns of Cell Phones: 2000-2004. Paper Presented at the Annual Meeting of the American Association for Public Opinion Research, Phoenix, Arizona, May 13-16, 2004
Additional factor for cell-only folks: there may be differing turnout rates between Democractic and Republican cell-phone only constituencies.
Still, if not exclusively college age, wireless only adults are predominantly under age 45 (81%). They also tend to live in large metropolitan areas (82%), earn less than $40,000 annually (66%) and rent rather than own a home (62%; the comparable percentages for adults with a landline are 51% age 18-44, 73% metro area, 39% <$40K and 24% renter).
One sub-population of cell-onlies is young, poor, and mobile - which means it is less likely to vote. Note that the combination of lower income than the national average, renters, AND metro-area homes means that much of the cell-only population are urban poor who, by virtue of higher metro-area costs of living, are substantially poorer than the national average. Their likley party support is strongly Dem, but turnout rates are VERY low.
However, there are other sub-populations: rural cell-onlies, home-owner cell-onlies, and older cell-onlies. Some of these may intersect because poorer rural or older folks will find the same appeal of cell-phones as the young urban poor:
(a) if you have to move (a lot), you can still use the phone without paying exorbitant new-subscriber charges;
(b) your friends and family always know how to reach you even when working two jobs or partying late;
(c) if you're driving in bad neighborhoods, or on late night freeways, or in the middle of a rural nowhere, you can always call for help;
(d) if your landlord evicts you, you can still use the phone;
(e) if you skip too many utility bills the landline companies will cut you off, but the pre-paid and pay-as-you-go nature of cell phones (plus the unlimited night/weekend plans) means that you always know if you can make a call.
But there are also sub-populations who have very different reasons for being cell-only, such as "wired workers" who use Skype and Vonage and have broadband connections or snow-bird retirees who want a year-round phone number. Such users at the richer end of the distribution are more likely to vote Republican and also have higher turnout rates.
In short, even if the cell-only population is 5% of the electorate and leans Dem 2:1, it won't matter if the GOPer leaning sub-population of cell-onlies turns out at twice the rate of the Dem-leaning sub-population.
Posted by: Silent E | Oct 18, 2004 12:42:14 PM
if not exclusively college age, wireless only adults are predominantly under age 45 (81%). They also tend to live in large metropolitan areas (82%), earn less than $40,000 annually (66%) and rent rather than own a home (62%
doesn't that sound like a pretty good description of college students and new graduates in metropolitan areas? 2.5% of the population is wireless only. roughly 5% of the population is in college, metro or non-metro.
Posted by: gabbneb | Oct 18, 2004 10:19:34 PM
If wireless-predominant users are concentrated in coastal cities, wouldn't that suggest that they are more likely to be Kerry supporters? And that Kerry's support is therefore undercounted in blue states? And that Kerry's support is therefore undercounted nationally? And in one battleground state in particular - Florida?
Posted by: gabbneb | Oct 18, 2004 10:22:30 PM
If wireless only voters are concentrated in coastal urban areas, that does suggest that Kerry's strength is being under-reported nationally, but may not be relevant to the outcome of the election, which will be decided mainly in the midwest states. But I agree the the key isn't how many actually don't have a land line, but how many rarely use/answer it.
The key questions which I think would really tell us if there's a problem with the polling are: How many total calls have to be made to get a set number (say 1000) or responses? Has this count gone up significantly since 2000?
If the answer to the second question is yes, that does suggest some real problems. The polling companies certainly have this information; whether they will be willing to share it is another matter.
Posted by: Alex | Oct 19, 2004 12:28:54 PM
If a large % of the cell-only voters are college kids, that does not indicate Kerry's likely vote is undercounted. College kids have very low voting rates and are unlikely to have any impact on this election.
Posted by: lex | Oct 21, 2004 1:30:47 PM
Just to speak up for myself here, I am a mobile-only household:
I also carry the distinction of being a conservative (well libertarian conservative, if you want to get technical) working in the film industry, even living in Hollywood, born here no less (that makes me legal too!). I have never "leaned democratic" (they don't seem to have learned about the Austrian school of economic theory) and I have voted in almost every election since turning 18 (even local only elections). I am now 34.
I am not considered to be a "low income household", I just find having a land line redundant and wasteful (why be saddled with so many extra taxes and fees? twice?) Besides, everyone calls on both lines anyway until they reach you - all of my messages used to be : "You're not there; I'll try you on your cell." A waste of my money and my time.
Posted by: ac | Oct 21, 2004 7:02:07 PM
And...don't forget with caller I.D., many voters, seeing an unfamiliar number, more often than not ignore it.
Posted by: Miriam A. Schaeffer | Oct 30, 2004 5:20:19 PM
The comments to this entry are closed.