By Kevin Ingham
“How are you dealing with the people who only have cell phones?”
I hear this question at every presentation and every meeting. It comes up in social settings when people learn I’m a pollster. Needless to say, I’m a hit at parties.
There’s good reason to ask the question: It poses a challenge for every pollster.
You can go here, here, and here for background on why this is important, but here’s the short version: The number of people without a landline is growing steadily and it is expensive to poll these folks.
So why does the cell-only population matter? Because there is growing evidence that excluding cell-only respondents from a poll will introduce bias that cannot be compensated for by demographic weighting or attributed to the margin of error.
These differences are on full display here in the Evergreen State where the government’s most recent state-level estimates are that 26.4% of Washington adults live in cell-only households.
Below is a table with the results from a poll Strategies 360 completed in November of likely 2012 voters in Washington. Respondents were reached on both landlines and cell phones.
|
DIFFERENCES IN OPINION BY PHONE USAGE |
|||
|
|
|
|
|
| Public Figure Favorability |
Favorable |
Unfavorable |
Margin |
| Barack Obama | |||
|
Cell-only |
63% |
32% |
+31 |
|
Has a landline |
51% |
45% |
+6 |
| Mitt Romney | |||
|
Cell-only |
31% |
41% |
-10 |
|
Has a landline |
36% |
40% |
-4 |
| Maria Cantwell | |||
|
Cell-only |
61% |
12% |
+49 |
|
Has a landline |
49% |
32% |
+17 |
| Chris Gregoire | |||
|
Cell-only |
35% |
49% |
-14 |
|
Has a landline |
41% |
48% |
-8 |
| Presidential Race |
Obama |
Romney |
Margin |
| Obama vs. Romney | |||
|
Cell-only |
51% |
38% |
D+13 |
|
Has a landline |
47% |
43% |
D+5 |
|
|
|||
| Gubernatorial Race |
Inslee |
McKenna |
Margin |
| Inslee vs. McKenna | |||
|
Cell-only |
35% |
45% |
R+9 |
|
Has a landline |
40% |
47% |
R+8 |
|
|
|||
As you can see, there are some clear differences between the two groups including:
- A 25-point difference in the margin of favorability for the president
- A 6-point difference in the margin of favorability for Mitt Romney
- A 32-point difference in the margin of favorability for Maria Cantwell
- A 6-point difference in the margin of favorability for Chris Gregoire
- An 8-point difference in the margin of presidential candidate preference.
Interestingly, at this early point in the race, there don’t appear to be significant differences in gubernatorial candidate preferences between the two groups.
These findings are strikingly consistent with the findings of both SurveyUSA (both locally and nationally) and Marist about the differences between cell-only respondents and those with landlines.
Obviously, there is mounting evidence that the cell phone challenge is one that must be addressed by pollsters if we wish to continue providing high-quality research to the media and our clients.
So, going back to the original question, how is Strategies 360 handling the cell phone challenge?
We are staying ahead of the problem by adapting and innovating. Is your pollster?
Kevin Ingham is Strategies 360’s vice president of polling and research.








