Tuesday, November 4, 2008

Ten Reasons Why You Should Ignore Exit Polls

Posted by Randy Barnett:

http://volokh.com/archives/archive_2008_11_02-2008_11_08.shtml#1225814335


   From [1]FiveThirtyEight:

     1. Exit polls have a much larger intrinsic margin for error than
     regular polls. This is because of what are known as cluster
     sampling techniques. Exit polls are not conducted at all precincts,
     but only at some fraction thereof. Although these precincts are
     selected at random and are supposed to be reflective of their
     states as a whole, this introduces another opportunity for error to
     occur (say, for instance, that a particular precinct has been
     canvassed especially heavily by one of the campaigns). This makes
     the margins for error somewhere between 50-90% higher than they
     would be for comparable telephone surveys.
     2. Exit polls have consistently overstated the Democratic share of
     the vote. Many of you will recall this happening in 2004, when
     leaked exit polls suggested that John Kerry would have a much
     better day than he actually had. But this phenomenon was hardly
     unique to 2004. In 2000, for instance, exit polls had Al Gore
     winning states like Alabama and Georgia (!). If you go back and
     watch The War Room, you'll find George Stephanopolous and James
     Carville gloating over exit polls showing Bill Clinton winning
     states like Indiana and Texas, which of course he did not win.
     3. Exit polls were particularly bad in this year's primaries. They
     overstated Barack Obama's performance by an average of about 7
     points.
     4. Exit polls challenge the definition of a random sample. Although
     the exit polls have theoretically established procedures to collect
     a random sample -- essentially, having the interviewer approach
     every nth person who leaves the polling place -- in practice this
     is hard to execute at a busy polling place, particularly when the
     pollster may be standing many yards away from the polling place
     itself because of electioneering laws.
     5. Democrats may be more likely to participate in exit polls.
     Related to items #1 and #4 above, Scott Rasmussen has found that
     Democrats supporters are more likely to agree to participate in
     exit polls, probably because they are more enthusiastic about this
     election.
     6. Exit polls may have problems calibrating results from early
     voting. Contrary to the conventional wisdom, exit polls will
     attempt account for people who voted before election day in most
     (although not all) states by means of a random telephone sample of
     such voters. However, this requires the polling firms to guess at
     the ratio of early voters to regular ones, and sometimes they do
     not guess correctly. In Florida in 2000, for instance, there was a
     significant underestimation of the absentee vote, which that year
     was a substantially Republican vote, leading to an overestimation
     of Al Gore's share of the vote, and contributing to the infamous
     miscall of the state.
     7. Exit polls may also miss late voters. By "late" voters I mean
     persons who come to their polling place in the last couple of hours
     of the day, after the exit polls are out of the field. Although
     there is no clear consensus about which types of voters tend to
     vote later rather than earlier, this adds another way in which the
     sample may be nonrandom, particularly in precincts with long lines
     or extended voting hours.
     8. "Leaked" exit poll results may not be the genuine article.
     Sometimes, sources like Matt Drudge and Jim Geraghty have gotten
     their hands on the actual exit polls collected by the network
     pools. At other times, they may be reporting data from "first-wave"
     exit polls, which contain extremely small sample sizes and are not
     calibrated for their demographics. And at other places on the
     Internet (though likely not from Gergahty and Drudge, who actually
     have reasonably good track records), you may see numbers that are
     completely fabricated.
     9. A high-turnout election may make demographic weighting
     difficult. Just as regular, telephone polls are having difficulty
     this cycle estimating turnout demographics -- will younger voters
     and minorities show up in greater numbers? -- the same challenges
     await exit pollsters. Remember, an exit poll is not a definitive
     record of what happened at the polling place; it is at best a
     random sampling.
     10. You'll know the actual results soon enough anyway. Have
     patience, my friends, and consider yourselves lucky: in France, it
     is illegal to conduct a poll of any kind within 48 hours of the
     election. But exit polls are really more trouble than they're
     worth, at least as a predictive tool. An independent panel created
     by CNN in the wake of the Florida disaster in 2000 recommended that
     the network completely ignore exit polls when calling particular
     states. I suggest that you do the same.

References

   1. http://www.fivethirtyeight.com/2008/11/ten-reasons-why-you-should-ignore-exit.html

No comments: