Don't Forget the Undecided Voters

Apparently it has become my mission in life to try to persuade my fellow Republicans and conservatives, not to mention the voting public at large, to stop believing polls that supposedly show Barack Obama "winning" the race for president over John McCain, despite the fact that the actual election has not yet taken place. 

In a recent article, I pointed out the fundamental inaccuracy (and bias) that results from how the polls are constructed (and, I should have added, implemented), how the polls are "weighted" (e.g., by assuming that Democrats or blacks or younger voters will vote in disproportionately large numbers this year), and how the media accounts -- or rather, fails to account -- for the polls' "margin of error."  This last point is critical because most polls show that the "gap" between the candidates is within the reported margin of error, meaning that "in reality" McCain might be "winning" the race.


But there is another crucial issue that the media routinely overlooks when reporting polls:  undecided voters.  For example, consider the Fox News/Opinion Dynamics poll  from October 10.  The Fox News headline reads:  "Obama Maintains Lead Over McCain."  The poll shows Obama leading McCain by 46% to 39% among registered voters nationwide.  Moreover, the margin of error for this polls is +/- 3 points, which means that Obama's lead is larger than the poll's margin of error. 

But note the real message of this poll:  Neither Obama nor McCain garners majority support, and 15% of voters remain undecided.  That's huge.

Consider another example, the Reuters/C-Span/Zogby poll  from October 19.  This poll shows Obama leading McCain by 47.8% to 45.1% among likely voters nationwide.  The reported margin of error for the poll is +/- 2.9 points, larger than the gap between the candidates.  But, again, note that the poll includes 7.1% undecided voters.

(Of course, the wide variability in the many published polls should be another clue that they should not taken too seriously.)  

The race between Obama and McCain will be determined by how these undecided voters (not the core supporters of the Democratic and Republican candidates) decide to "break."  The polls, by definition, cannot tell us how these voters are going to vote. As one of this election's leading clichés has it, Obama seems to have a hard time closing the deal.

The bottom line, once more, is that the election will not be decided until the American people go to the polls on November 4 and actually cast their votes.  For the reasons I've stated previously, I predict  that most voters, ultimately, will pull the lever for McCain.

Steven M. Warshawsky is an attorney  in New York City. 
Apparently it has become my mission in life to try to persuade my fellow Republicans and conservatives, not to mention the voting public at large, to stop believing polls that supposedly show Barack Obama "winning" the race for president over John McCain, despite the fact that the actual election has not yet taken place. 

In a recent article, I pointed out the fundamental inaccuracy (and bias) that results from how the polls are constructed (and, I should have added, implemented), how the polls are "weighted" (e.g., by assuming that Democrats or blacks or younger voters will vote in disproportionately large numbers this year), and how the media accounts -- or rather, fails to account -- for the polls' "margin of error."  This last point is critical because most polls show that the "gap" between the candidates is within the reported margin of error, meaning that "in reality" McCain might be "winning" the race.


But there is another crucial issue that the media routinely overlooks when reporting polls:  undecided voters.  For example, consider the Fox News/Opinion Dynamics poll  from October 10.  The Fox News headline reads:  "Obama Maintains Lead Over McCain."  The poll shows Obama leading McCain by 46% to 39% among registered voters nationwide.  Moreover, the margin of error for this polls is +/- 3 points, which means that Obama's lead is larger than the poll's margin of error. 

But note the real message of this poll:  Neither Obama nor McCain garners majority support, and 15% of voters remain undecided.  That's huge.

Consider another example, the Reuters/C-Span/Zogby poll  from October 19.  This poll shows Obama leading McCain by 47.8% to 45.1% among likely voters nationwide.  The reported margin of error for the poll is +/- 2.9 points, larger than the gap between the candidates.  But, again, note that the poll includes 7.1% undecided voters.

(Of course, the wide variability in the many published polls should be another clue that they should not taken too seriously.)  

The race between Obama and McCain will be determined by how these undecided voters (not the core supporters of the Democratic and Republican candidates) decide to "break."  The polls, by definition, cannot tell us how these voters are going to vote. As one of this election's leading clichés has it, Obama seems to have a hard time closing the deal.

The bottom line, once more, is that the election will not be decided until the American people go to the polls on November 4 and actually cast their votes.  For the reasons I've stated previously, I predict  that most voters, ultimately, will pull the lever for McCain.

Steven M. Warshawsky is an attorney  in New York City.