Via the WSJ are the sobering statistics on the share of people who have gone ‘wireless only’.
The article covers this in context of political polling. The law in the USA, where customers pay to receive as well as make mobile phone calls, is that pollsters cannot use autodiallers or robot calls, and must use live interviewers. Autodialers and robots are much cheaper, and faster than using real humans. Meanwhile mobile phone numbers are often initially tagged to a region like the landline numbers, but people generally keep their number when they move regions. This makes it harder for pollsters to build a sample from a particular set of regions based on numbers alone. Overall the impact is that calls to mobile phones are about 10 times more expensive than land lines, and so most pollsters have just ignored the entire mobile-only households.
It doesn’t take much thinking to see how this can skew results. The WSJ reports that the younger and, it seems, poorer and African American or Latinos are undercounted. While in theory this can be corrected for with the right sampling strategy, pollsters are now finding that cell phone users are far more likely to support Democrats than their land-lined peers. All this flows back into the politics of the land, as pollsters represent one force that guides legislation and campaigns. (The other stronger force is, of course, the vast amount of money spent by corporations and lobbyists.)
In New Zealand I suspect much of the same is happening with polls and surveys. I’ve been landline free for years, and I suspect our charts look a lot like the USA ones, where older people are land line lovers, and younger people are more likely to have given up.
Existing data on NZ landline use is, I suspect, highly flawed because of the prevalence of ADSL and the almost universal requirement to have a landline along with it. I, for example, pay for a landline because I have to, yet I have never plugged a phone into that line. Others may plug a phone into that line, but use it for very little. I hear stories of such phones almost exclusively being used for receiving unsolicited calls.
Overall it’s very dangerous to ignore such a segment of the population. The wireless only crowd are more likely, I imagine, to be heavy internet users, to have smart phones like the iPhone and to be reliant almost exclusively on the internet for their content and communication. For pollsters this means getting a lot smarter about crafting the strategy for reaching a statistically sound sample of population. For now a hybrid fixed, mobile and internet-based approach may be best.
In the meantime, like internet-only based polls, take what you read from polls with a healthy grain of scepticism. And think about switching off that land-line of you do have one. It’s a wonderful feeling to be free of polls, surveys and robot calls.