Election Issues Today

Are Pollsters Overlooking the Obvious?

Mark Blumenthal recently wrote in the National Journal about issues pollsters are facing (and fearing) in predicting voting patterns and outcome in this general election for President. (He mentions that George Stephanopoulos, of ABC, also recently summed up pollsters’ concerns on ABC’s “This Week ” .)

According to Blumenthal and Stephanopoulos, the three issues that pollsters fear could “throw off” their pre-vote polling include:

1.  The number of cell phone-only households and voters, a VERY large number of whom are in the 18 to 29 year-old age demographic. These people do not get polled during landline telephone surveys.  They seem to be in the Obama camp in large numbers. Question? Could this throw off the estimated strength Obama may demonstrate in “too close-to-call States? In the Weaker pro-McCain States?

2.  The number of African-Americans who may come out on Election Day, expected to be largely in support of Obama, while currently being disproportionately underestimated as “likely voters” in polls.

3.  The “Bradley Effect” also called the Bradley-Wilder Effect, in which White voters do not truthfully indicate their actual preference, in that they often change their minds at the last minute, in the voting booth, as they are simply not able to bring themselves to vote for a Black man. The Bradley Effect dates from the 1982 Gubernatorial race in California, when Los Angeles Mayor Tom Bradley had a good lead in the polls right up until election day, and then lost. Post election research showed that White people even lied to pollsters as they EXITED the polling place, especially if the pollster was Black. Also see my previous article: White People Lie to Pollsters.

All-in-all, the first two “unknowns” would appear to favor Senator Obama, and the third variable, the so-called “Bradley Effect,” would benefit Senator McCain.  So, could these variables cancel each other out?  We must also consider that recent elections have tended to reflect a very small, if any, “Bradley Effect,” although those elections were not for President.

Of course, since pollsters are aware of these possible problems in polling, they are now conducting research to validate or invalidate, so far as possible, each of these variables.

We shall see how it all turns out on Election Night. Fun.