Wednesday, September 10, 2008

Landlines and Cell Phones and Poll Bias - Deeper Look

Synopsis:

Overview
  • ISER take on Landlines and Cell Only in surveys
  • There are differences, but one study says not for general surveys

CDC's NHIS Findings - Demographics of cell only vs. landline households

PEW Findings - Implications for surveyors - generally no difference, but significant for some populations.


Overview
A fair number of people have googled their way my post Land-Lines, Cell Phones, and Poll Bias. I really didn't have any hard data when I wrote that post and thought I should see if I could get some. So yesterday I called Virgene Hanna at the Institute for Social and Economic Research (ISER) at the University of Alaska Anchorage. She runs lots of the polling and survey work there. I asked her if there were studies about the differences between people who use cell phones only (no landlines) and landline users and, if there is, what if anything they do about it.

She said that there are differences. Cell phone only users, for example, tend to be:
  • younger
  • in better health
  • and more likely to rent
While there are some significant differences between cell only and landline households, one study (see PEW below) found that in general surveys, these differences had no impact.

So far, she hasn't changed how they do their survey research for a number of reasons:
  1. Things are changing so rapidly, it's hard to get clear cut answers on how to adjust.
  2. Studies of the 2004 election said it didn’t make a difference in the overall outcomes, but use of cell phones only has grown a lot since then.
  3. It's really expensive to reach cell phone only people
  4. Methodologically more difficult to calculate who falls into their sample design when you include cell phone only users
  5. Takes longer to reach them
Then she pointed me to a couple of studies that have looked at this.

The CDC (Center for Disease Control) does its National Health Interview Survey (NHIS).
For many years, NHIS has included questions on residential telephone numbers to permit re-contact of survey participants. Starting in 2003, additional questions determined whether the family's telephone number was a landline telephone. All survey respondents were also asked whether "you or anyone in your family has a working cellular telephone." (from Blumberg)
So, Blumberg and Luke, in "Wireless Substitution: Early Release of Estimates From the National Health Interview Survey, July-December 2007" have taken this data to look at related characteristics of the cell only population. While there is important demographic information, the focus here is on health.
Preliminary results from the July-December 2007 National Health Interview Survey (NHIS) indicate that nearly one out of every six American homes (15.8%) had only wireless telephones during the second half of 2007. In addition, more than one out of every eight American homes (13.1%) received all or almost all calls on wireless telephones despite having a landline telephone in the home. This report presents the most up-to-date estimates available from the federal government concerning the size and characteristics of these populations. Specific findings:

Cell Only Population

In the last 6 months of 2007, nearly one out of every six households (15.8%) did not have a landline telephone, but did have at least one wireless telephone (Table 1). Approximately 14.5% of all adults-more than 32 million adults-lived in households with only wireless telephones; 14.4% of all children-more than 10 million children-lived in households with only wireless telephones...
Approximately 2.2% of households had no telephone service (neither wireless nor landline). Approximately 4 million adults (1.9%) and 1.5 million children (2.1%) lived in these households.

Demographics
The percentage of U.S. civilian noninstitutionalized adults living in wireless-only households is shown by selected demographic characteristics and by survey time period in Table 2. For the period July through December 2007:
  • More than one-half of all adults living with unrelated roommates (56.9%) lived in households with only wireless telephones. This is the highest prevalence rate among the population subgroups examined.
  • Adults renting their home (30.9%) were more likely than adults owning their home (7.3%) to be living in households with only wireless telephones.
  • More than one in three adults aged 25-29 years (34.5%) lived in households with only wireless telephones. Nearly 31% of adults aged 18-24 years lived in households with only wireless telephones.
  • As age increased, the percentage of adults living in households with only wireless telephones decreased: 15.5% for adults aged 30-44 years; 8.0% for adults aged 45-64 years; and 2.2% for adults aged 65 years and over.
  • Men (15.9%) were more likely than women (13.2%) to be living in households with only wireless telephones.
  • Adults living in poverty (27.4%) were more likely than higher income adults to be living in households with only wireless telephones.
  • Adults living in the South (17.1%) and Midwest (15.3%) were more likely than adults living in the Northeast (10.0%) to be living in households with only wireless telephones.
  • Non-Hispanic white adults (12.9%) were less likely than Hispanic adults (19.3%) or non-Hispanic black adults (18.3%) to be living in households with only wireless telephones.

Wireless Mostly Households
Among households with both landline and cellular telephones, 22.3% received all or almost all calls on the cellular telephones, based on data for the period July through December 2007. These wireless-mostly households make up 13.1% of all households. ..
Approximately 31 million adults (14.0%) lived in wireless-mostly households during the last 6 months of 2007, an increase from 28 million (12.6%) during the first 6 months of 2007. Table 3 presents the percentage of adults living in wireless-mostly households by selected demographic characteristics and by survey time period. For the period July through December 2007:

  • Non-Hispanic Asian adults (20.3%) were more likely than Hispanic adults (14.5%), non-Hispanic white adults (13.2%), or non-Hispanic black adults (15.1%) to be living in wireless-mostly households.

  • Adults with college degrees (16.2%) were more likely to be living in wireless-mostly households than were high school graduates (12.7%) or adults with less education (8.7%).

  • Adults living in poverty (8.6%) and adults living near poverty (11.4%) were less likely than higher income adults (15.9%) to be living in wireless-mostly households.

  • Adults living in metropolitan areas (14.7%) were more likely to be living in wireless-mostly households than were adults living in more rural areas (10.9%).


PEW Study

A Pew study presented in a May 2007 paper, "What’s Missing from National RDD Surveys? The Impact of the Growing Cell-Only Population" by Scott Keeter, Courtney Kennedy, April Clark of the Pew Research Center and Trevor Tompson, Mike Mokrzycki of The Associated Press concluded

Analysis of all four studies produce the same conclusion: Although cell-only respondents are different from landline respondents in important ways, they were neither numerous enough nor different enough on the key dependent variables to produce a significant change in overall general population survey estimates when included with the landline samples and weighted according to U.S. Census parameters on basic demographic characteristics. However, certain survey topics and sampling frames may be vulnerable to significant, even dramatic, noncoverage bias if they exclude respondents who only can be reached by cell. This paper concludes with evidence regarding the potential for bias in survey estimates for certain variables among young people.

The four studies were:



The Pew comparison findings were summarized in this chart:


Here are some specific health characteristics of cell only and landline households that the PEW study highlighted:
  • The prevalence of binge drinking (i.e., having five or more alcoholic drinks in 1 day during the past year) among wireless-only adults (37.3%) was twice as high as the prevalence among adults living in landline households (17.7%). Wireless-only adults were also more likely to be current smokers.
  • Compared with adults living in landline households, wireless-only adults were more likely to report that their health status was excellent or very good, and they were more likely to engage in regular leisure-time physical activity.
  • The percentage without health insurance coverage at the time of the interview among wireless-only adults (28.7%) was twice as high as the percentage among adults living in landline households (13.7%).
  • Compared with adults living in landline households, wireless-only adults were more likely to have experienced financial barriers to obtaining needed health care, and they were less likely to have a usual place to go for medical care. Wireless-only adults were also less likely to have received an influenza vaccination during the previous year.
  • Wireless-only adults (47.6%) were more likely than adults living in landline households (34.7%) to have ever been tested for HIV, the virus that causes AIDS.


[Update: June 2, 2009: Here's a December 2008 PEW report on cell and landline polling in the 2008 election.

This study describes the differences between estimates of the horse race and other political measures that Pew reported this fall with those that would have been derived from surveys conducted only by landline. It also addresses the difference between supplementing landline surveys with a sample of people who are "cell only" vs. interviewing all cell respondents even if they also have a landline phone. In this regard there is growing concern that some people have come to rely so heavily on a cell phone that even though they still have a landline telephone they are virtually unreachable on it. Finally, this report describes the operational and cost issues raised by the inclusion of cell phones.

Download the complete report]

No comments:

Post a Comment

Comments will be reviewed, not for content (except ads), but for style. Comments with personal insults, rambling tirades, and significant repetition will be deleted. Ads disguised as comments, unless closely related to the post and of value to readers (my call) will be deleted. Click here to learn to put links in your comment.