As the 2010 election draws near, many political pundits have predicted that the Democrats will face significant losses in House elections (for a nice summary of election forecasts see here and here). Some have even suggested that the losses will be enough for Democrats to lose majority control over the House. Commentators on CNN have even began discussing how Republicans and Democrats will share power in Washington after the election.
Although it seems highly likely that Democrats will face losses in both chambers in the midterm election (as this is usually the case in midterm elections for members of the President's party), the extent of the losses may be slightly less than expected. A new study by the Pew Research Center for the People & the Press suggests that polls that do not call cell phones may be misrepresenting how well Republicans will do in this year's upcoming election. While most national polls do call cell phones, the state and local polls generally do not. Given the study's results, predictions that use state and local polls in predicting this year's election results may be overestimating how well the Republicans will do.
More after the jump…
According to the National Center for Health Statistics, 25% of Americans have no landline service and only have a cell phone. Despite the growing numbers of cell phone only use, many pollsters who conduct telephone surveys at the state and local level still do not call cell phone numbers because it is more costly than just calling land line numbers. Federal law requires any call placed to a cell phone to be hand dialed, which increases costs and is more time consuming than automated dialing.
In order to assess whether there is a bias in surveys that are conducted with land line only numbers, the Pew Research Center for the People & the Press conducts two polls: one using land line only numbers and another using land line and cell phone users. The results suggest a Republican bias in the land line only numbers poll:
"In three of four election polls conducted since the spring of this year, estimates from the landline samples alone produced slightly more support for Republican candidates and less support for Democratic candidates, resulting in differences of four to six points in the margin. One poll showed no difference between the landline and combined samples…
Limiting the analysis in the survey to those considered most likely to vote in this year's elections, a similar bias is evident. The combined landline and cell estimate produced a seven-point Republican advantage: 50% supported the GOP candidate for Congress in their district while 43% backed the Democratic candidate. The Republican lead would have been 12 points if only the landline sample had been interviewed, a significant difference from the combined sample of five points in the margin."
Given these results, it is possible that many of the state wide and local polls may be misrepresenting Republican strength in this upcoming election. How much is uncertain. Even if there is slight misrepresentation, it will only make a difference in tight races. Regardless of the outcome, it appears that with the increased use of cell phone only use, pollsters at the state and local level may have to soon start including cell phone users in their polls in order to get more accurate results.