Pollster.com

Articles and Analysis

 

"I'm Competing with the Margin of Error"

Topics: 2008 , The 2008 Race

Today, the Hartford Courant (via Political Wire) brings us this perennial bit of political humor, the long shot presidential candidate whose total support fails to exceed the margin of error:

When pollster John Zogby asked a group of 339 likely Democratic voters earlier this month whom they wanted for president in 2008, one or two mentioned Connecticut's senior senator.

One or 2 percent mentioned Christopher Dodd?

No. One or two people, Zogby said.

"I'm competing with the margin of error in most polls," Chris Dodd grinned as he described his underdog status recently to a New Hampshire audience.

He should be so lucky. The margin of error was 5.4 percentage points.

There is one more thing perennial about long shots in early presidential trial heat poll questions. Consider these numbers, each the standing of the respective candidate from the Gallup Poll trial heat questions asked on surveys at about this time in each election cycle.**

  • 6% - George McGovern, August 1971
  • 3.5% - Jimmy Carter, February 1975
  • 2.9% - George Bush, January 1979
  • 1.6% - Gary Hart, December 1982
  • 3.0% - Bill Clinton, February 1991
  • 3.8% - John McCain, March 1999
  • 4.0% - Howard Dean, January 2003

Of course, Jimmy Carter and Bill Clinton went on to be elected President, and George McGovern won his party's nomination. George Bush, Gary Hart and John McCain all emerged as the principal challengers to the respective front-runners in each race, with Hart falling just a hundred or so delegates short of defeating Walter Mondale in 1984. While Howard Dean's campaign tanked early, he rose to 25% in the Gallup poll by December 2003, while John Kerry, the eventual Democratic nominee, had fallen to just 7%.

The point is not that front-runners always lose. In fact, the early front runners probably win more often than not (an interesting empirical test for another post). The point is that writing off candidates now because they begin with support in the single digits is foolish. In almost every election, a long-shot candidacy that no one saw coming emerges from the single digits to compete with the front runners.

**I obtained the numbers from the subscription-only "Gallup Brain" search engine, which does not clearly display the respondent base (i.e. Democrats, Democratic primary voters) used to calculate the results. So it is possible these numbers may differ slightly from those published at the time.

 

Comments
Nick Panagakis:

Competing with the margin error is clever - but not accurate.

The stated MOE of 5.4% for 339 is for estimates or poll findings near 50%.

The MOE declines for findings less than or greater than 50%. Absolute error is greatest at 50% so this has become the conventional statement of potential statistical error used by pollsters.

The formula for calculating MOE makes it impossible for an MOE to exceed the estimate.

____________________

Nick Panagaakis:

Competing with the margin error is clever - but not accurate.

The stated MOE of 5.4% for 339 is for estimates or poll findings near 50%.

The MOE declines for findings less than or greater than 50%. Absolute error is greatest at 50% so this has become the conventional statement of potential statistical error used by pollsters.

The formula for calculating MOE makes it impossible for an MOE to exceed the estimate.

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR