Bookmark and Share

Alex Bäcker's Wiki / Most election polls aren't worth the paper they're printed on
  • If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • You already know Dokkio is an AI-powered assistant to organize & manage your digital files & messages. Very soon, Dokkio will support Outlook as well as One Drive. Check it out today!


Most election polls aren't worth the paper they're printed on

Page history last edited by Alex Backer, Ph.D. 15 years, 8 months ago

Check out two polls announced by two leading news agencies today:

Are these people polling the same country?


The article by Josh Drobnyk in Saturday's edition of the Times is irresponsible. Entitled "Fence-sitters inch to McCain", it is based on polling a total of 14 voters, with McCain edging Obama by 2. While that could make a good comic article in The Onion, it does not fit in the pages of serious journalists --it's the equivalent of me polling my parents and reporting it with that title. Even the director of the institute that conducted the focus group said it was unscientific.


But Drobnyk is by no means alone: earlier last month, the Times-Bloomberg poll announced that "Obama leads McCain, 49% to 45%, in a survey of likely voters that has a margin of error of 3 percentage points." Based on the responses of 838 likely voters, this slim difference is not statistically significant --there is a probability of more than 10% that identical vote intent for both candidates would have yielded a difference of 4% or greater with a sample size that small. Statistically insignificant polls in turn can affect the very population they intend to measure, so bad math here can have repercussions beyond bad reporting.


This is not surprising, as the level of mathematics taught to our kids in schools is far from challenging, even in some of our best schools.


Shallowness of mathematical analysis seems to underlie many of our nation's current debacles. If we want to prevent the next bailout, I say some of the 700 billion dollars in the bailout get used to improve mathematics education in our nation’s schools. It is time for a committee of our nation's top mathematicians and mathematical educators to re-calibrate the standards of mathematics taught in our schools. Otherwise, we can count on more problems down the line. Although it does not take a rocket scientist to tell that it is a stretch to generalize from two voters to any meaningful number of fence-sitters, perhaps The Times should lead by example by hiring a statistics consultant to verify the mathematical soundness of any articles that rely on math. These days, more and more things --more than 700 billion things-- do.


--Alex Backer, Oct. 22, 2008


Comments (0)

You don't have permission to comment on this page.