Monday, October 18, 2010

Not polling cellphone users

Nate Silver

We saw a interesting note from Nate Silver, a baseball guy, a statistician and pollster that the Clarion Content has been following for some time. Silver, once an independent blogger, has recently been hired by the New York Times.

Try not to hold it against him, he still unearths some fascinating data. For example, last week he held forth on the effect on polling of the exclusively cellphone demographic. At least 25% of Americans no longer have landlines; of course, these folks tend to skew tech savvy and younger. Strangely enough, many American polling companies do not do any surveying of the only cellphone using set at all, dialing up only landlines.

None! Egad, a remarkable oversight.

Good catch, Nate. The answer of the pollsters is that they use demographic weighting to make up for the underrepresented. As Silver points out, "It’s always been harder to get men on the phone then women, younger people than older people, blacks and Hispanics than whites." However, Silver points to new study by the renown Pew Research suggesting that bias remains even after demographic weighting is applied. According to Pew the effect of the failure to include cellphones may result in a four percentage point bias against Democrats on the generic ballot.

Read the whole article here.

No comments: