|
|
|
|
|
|
|
Click here to see our elections web page
*
Why it's not a free country anyway
*
Examples where the united $tates has not respected majority rule before
*
See our article refuting a certain line of approach to seeing Florida as a fraud
*
See our article on Zogby as a case of how capitalism influences polling
*
About Republican conspiracy theories of exit poll rigging
*
Banana Republic III: Did Kerry win?
by mim3@mim.org with a contribution from MC12
"We have decided to begin with
statistics, fully aware of course that
statistics are deeply antipathetic to certain
readers, who prefer `flattering deception'
to `base truths,' and to certain authors,
who are prone to smuggle in political
contraband under cover of `general'
disquisitions about internationalism,
cosmopolitanism, nationalism,
patriotism, etc."
--V.I. Lenin, "Statistics and Sociology"
"We must carry statistics to the people and make them popular."
--V.I. Lenin, "Immediate Tasks of the Soviet Government"
In theory, election results should be completely unconnected to statistics. Ideally we would like to think that statistics does not help in counting anything. "Sesame Street" education expanded to nine digit numbers should be enough, but it's not. We should all get used to living in a world of statistical complication.
In practice, with the hanging chads that did exist in Ohio in 2004, with machines that subtracted one vote for every vote added, with some machines completely turned off and never recovering their votes, with Democratic straight-party votes being counted as Libertarian, and places with more votes than voters, and worst of all--with some errors never caught because they were not obvious enough to voting precinct workers--statistics is relevant even to elections. Elections in the United $tates are huge operations fraught with humyn error and guile. They have yet to issue their report, but European monitors of U.$. elections will criticize procedures and follow-through in the November 2 2004 election.
Most people have heard about all that and the shame of Florida in 2000. Despite this, many would rather go shopping than consider that maybe someone stole the election. Very few people will put in the effort to check on the vote results and the question is so large and involving so many millions of people, one result will be political burnout and frustration. This article is for the brave who are going to put in the effort. Many will leave the election result as is, because "they don't want to go there."
If the united $tates is going to brag around the world about its "free elections," then the people are right to think that they should have an independent means of monitoring the elections and checking the results. When Kathy Dopp saw heavily Democratic counties voting heavily Republican, she was right to blow the whistle, based strictly on prediction, not based on her individual meetings with voters, which she did not have and no one had anyway. We don't agree with her conclusion, but she definitely set about checking the results the right way.
MIM gives credence to the concerns about the elections based on exit polls as well. The two methods of external validation of the election results are 1) Use of prediction, as in Kathy Dopp's case; 2) Completely separate data as in exit polls.
Statistics consumers should insist on a validation method separate from the
vote results. There is no other way to confirm the big picture.
In discussion of the Great Leap Forward population projections, MIM has already reported about past instances where statisticians believed it would be better to use statistics than rely on an imperfect Census count of the U.$. population. For many decades there have been questions about the cultural diversity and thoroughness that would be necessary to find all the various poor and ethnically diverse people in the united $tates. The 2004 election is a very similar problem.
If anything, the threat of political bias hangs over elections more than in the Census. In many Dixiecrat u.$. counties, not just the Republican chair but also the chair of the Democratic Party supported Bu$h for president. In that kind of atmosphere, we might question poll workers.
As it turns out, as in Florida 2000, the Supreme Court decided the Census of 1990. An argument went to court that the Census Bureau missed 1.58% of U.$. people. With the thin margin of Bu$h's victory, a 1.58% mistake in the elections would be fatal to the results.
In theory, a perfect exit poll could provide a more accurate result than a flawed vote result. We should not dismiss either the vote results or the poll results out-of-hand, but instead we should use our reasoning ability. That is why MIM is calling for release of the exit poll data and documentation to the public. The most difficult part in consuming statistics for the 2004 election dispute is that an argument referring to the voting results does not invalidate the exit polls. The whole point in the dispute is to see if there is some way to check on the results.
Statistics consumers should know that exit polls are better than pre-election
polls.
Exit polls are far better than pre-election polls. Gone is the indecision of the voter. It's much better to ask about what someone just did than to ask about a mood about a future choice.
Pre-election polls have to ask about "likely voters" and distinguish them from other voters.
Even many questions of "weighting" in a sample are easier to resolve in an exit poll than a pre-election poll.
Source on pre-election poll difficulties and concerns: http://www.aapor.org/pdfs/varsource.pdf
Weighting the exit polls in 2004 based on the 2004 vote result is a violation
of advice above.
We never said validating the election results is going to be easy. When the
question is how do voters in a particular
state feel about the Iraq War, weighting according to exit polls is correct.
When the question at stake is whether the
vote results of 2004 themselves are accurate, it makes no sense to weight according to
vote results from 2004. The mistake would be "assuming that which is to be proved."
In 2002, a polling organization said the following:
"The exit poll results are weighted to reflect the complexity of the sampling design. That is, the weighting takes into account the different probabilities of selecting a precinct and of selecting a voter within each precinct. For example, states that were selected at a higher rate receive a smaller weight than other precincts of the same size. There is also an adjustment for voters who were missed or refused to be interviewed, which is based on their observed age, race and sex."NOTE: THE EXIT POLL RESULTS ARE ONLY MEANINGFUL IF THEY ARE WEIGHTED. UNWEIGHTED TABULATIONS MAY BE SERIOUSLY MISLEADING AND SHOULD NOT BE USED FOR ANY REASON."
Source: http://www.ropercenter.uconn.edu/usvns2002_2.html#method
We ask our readers to note that the correct way to read the above is that some polls require weights to be used. It does not mean the 2004 vote results should determine the weights for the 2004 exit poll--if those 2004 vote results are the very subject of dispute. Some are alleging Republican county vote inflation and one county in Ohio has already had a proven case. The question is what is historically believable, something raised by the Kathy Dopp paper for example. There is also no guarantee that there is an easy answer to find in the exit polling data once released. We can only say that those who look at the exit polls now and see them so far off have a legitimate question with no decisive rebuttal yet.
Amerikan individualism is so very bitter, because its source is the settler history of the united $tates when each white stole a piece of land and then shot the indigenous people in his/her way. A series of ideological excuses came together to justify the white man's systematic crimes as "every man for himself." This bitter individualism affects even the whites' dealings with other whites and no where is this more evident than the fact that a very wealthy country resists statistical education like the plague.
In contrast, we of the more numerous exploited classes and oppressed nations of 6 billion people opposing U.$. imperialism have an interest in obtaining the overall picture in all questions. We are heartened that one of the most popular topics on the U.$. Internet right now is the election statistics and vote-rigging conspiracies.
The rulers want to spread individualism, so that no one unites to challenge their rule. Their preferred data is the anecdote, the tabloid story on the Hollywood stars. In contrast, a strong role for statistics is part of any Marxist-Leninist-Maoist ideology. While we Maoists also use anecdotes, we'd rather have a poll. Being humyn and not having developed science far enough yet, our polls have their flaws too, but there is no perfect answer yet to large questions such as how tens of millions of people voted, so we should not "make the perfect the enemy of the good."