This may seem like a fairly arcane subject, but in the 2008 election it is not.
The percentage of households which only have cell phones has risen from 6% in 2004 to 25% now, in time to scramble the accuracy of polling done for this election. Importantly, those households tend to be younger people, many in the 18 to 34 year old range. I heard repeatedly that pollsters only call households with land line phones, at random. The reason seems to be that there are regulations prohibiting automated calls to cell phones and pollsters use automated calling equipment. The net result? Polls tend to under- report younger voters. By how much? In the Wisconsin Primary Obama was polling 3 points over Clinton. The actual vote? Obama by 17 percent. 14 percent seems to have been hidden from pollsters by those voters only having cell phones... or else Obama had a really strong finish? Some of that 14 points could be put down to a good final day or two, but I am inclined to think that the cell phone effect also played a big part.
When you read this (click here), scroll down and look at the two graphs, one showing increased percentage of households with only cell phones, the other showing the percentage of younger households. In some polls (the Gallup Daily Tracking poll is the most notorious) the weight given to younger voters is 10%. But here, we see that age group looking more like 30%. Other polls weight that age group as 20% of their adjusted sample. That is, they call a bunch of people, ask them their age, ask them who they are voting for, then when they gather the 1200 or 1500 responses they group them by age range ("weighting") and pull out 20% of that age group and count them toward their final tally. One argument is that younger voters don't really vote. If Joe Biden were the candidate, that might be true, but he is not.
The other interesting thing about cell phone only households is that they tend to be minority households. If those individuals get out and vote this year, that would have a major impact on the outcome.
If these effects even gave as little as a 2 or 3 percent shift in the actual voting pattern, that could well determine the election. Notice that this phenomena is a new and undocumented effect, that will undoubtedly be a big news item. And note that the outcome of the Wisconsin Primary suggests it could make the difference, all else being equal.
And here is your daily dose of Sarah Palin, whose towering popularity in her home state seems to be taking a bit of a hit. The Anchorage Daily News (the paper of record in AK) is not happy. In fact they are coming perilously close to saying that Palin has become what she has railed against, a typical politician. It is probably too much to hope to be able to flip Alaska into the Obama column, but it will be fun to watch the Obama campaign try.
The percentage of households which only have cell phones has risen from 6% in 2004 to 25% now, in time to scramble the accuracy of polling done for this election. Importantly, those households tend to be younger people, many in the 18 to 34 year old range. I heard repeatedly that pollsters only call households with land line phones, at random. The reason seems to be that there are regulations prohibiting automated calls to cell phones and pollsters use automated calling equipment. The net result? Polls tend to under- report younger voters. By how much? In the Wisconsin Primary Obama was polling 3 points over Clinton. The actual vote? Obama by 17 percent. 14 percent seems to have been hidden from pollsters by those voters only having cell phones... or else Obama had a really strong finish? Some of that 14 points could be put down to a good final day or two, but I am inclined to think that the cell phone effect also played a big part.
When you read this (click here), scroll down and look at the two graphs, one showing increased percentage of households with only cell phones, the other showing the percentage of younger households. In some polls (the Gallup Daily Tracking poll is the most notorious) the weight given to younger voters is 10%. But here, we see that age group looking more like 30%. Other polls weight that age group as 20% of their adjusted sample. That is, they call a bunch of people, ask them their age, ask them who they are voting for, then when they gather the 1200 or 1500 responses they group them by age range ("weighting") and pull out 20% of that age group and count them toward their final tally. One argument is that younger voters don't really vote. If Joe Biden were the candidate, that might be true, but he is not.
The other interesting thing about cell phone only households is that they tend to be minority households. If those individuals get out and vote this year, that would have a major impact on the outcome.
If these effects even gave as little as a 2 or 3 percent shift in the actual voting pattern, that could well determine the election. Notice that this phenomena is a new and undocumented effect, that will undoubtedly be a big news item. And note that the outcome of the Wisconsin Primary suggests it could make the difference, all else being equal.
And here is your daily dose of Sarah Palin, whose towering popularity in her home state seems to be taking a bit of a hit. The Anchorage Daily News (the paper of record in AK) is not happy. In fact they are coming perilously close to saying that Palin has become what she has railed against, a typical politician. It is probably too much to hope to be able to flip Alaska into the Obama column, but it will be fun to watch the Obama campaign try.
Arthur
No comments:
Post a Comment