Reply #19: A less conspiratorial explanation: cell-phone-only bias [View All]

[ProgressiveEconomist]

The "Fortunate Son" may be benefiting from a trend in telephone survey research that Gallup alone among major polling companies is ignoring. To carry out a poll, companies like Gallup most often use "random digit dialing" in landline telephone exchanges.

But cell-phone exchanges have to be excluded. Would you want large corporations calling you at your expense for half an hour at a stretch? Thus those who are more likely to have only cell-phones are less likely to get into telephone samples. This trend started to affect telephone surveys at the end of the 90s. The proportion of adults reachable only by cell-phone is now up to about 7 percent and may reach 10 percent by the 2008 campaign season.

Who's more likely to be reachable only by cellphone? Why, the young, the mobile, and those more open to new ways of living. In other words, Democrats.

Virtually all survey companies recognize this problem with their "sampling frame", and apply statistical adjustments for "party identification" to remove any biases arising from it. Gallup stands alone in refusing to do this, ostensibly for the purest of purely "technical" reasons. But guess who may have benefited from this new wrinkle in the ability to carry out accurate survey research?