DEMOCRACY: Government & Politics | November 15, 2004

RESPONDING TO FEEDBACK ABOUT

"DO THE MATH: SOMETHING STINKS"

A Posting in "Comments From Left Field",

Summarizing Several E-Mail Exchanges with Political Activists

Prompted by My Previous, Widely Circulated Statistical Analysis

Note that I've included in brackets [] information that I have subsequently learned -- information that tends to make my case even stronger...

The feedback I am getting to my statistical analysis of the significant discrepancies between the exit polls and the reported votes falls into two general categories. Please allow me to address those concerns and perhaps clarify things a bit.

First, there is the political aspect. If there were actually such irresolvable differences between the exit polls and the reported votes, why is the Democratic leadership conceding the election and not pursuing this matter? Well, Rep. John Conyers Jr. (D-Michigan), of the House Judiciary Committee, and a growing number of his colleagues are actively pushing for investigations into the numerous reports of voting irregularities, in Florida, Ohio, and elsewhere.

But the DNC, Kerry, and other top Democratic leaders are pressing for us to "move on" (although moveon.org seems to be pressing the other way). Politically, no one wants to be seen as a "sore loser"...particularly when they don't think they can prove something was (at least intentionally) wrong.

Note that a statistical analysis as I have done indicates only what has happened, not how it has happened. I just crunch the numbers that I know to be factual and report my findings. If there are different numbers to work with, then they should also be crunched and any discrepancies also accounted for.

Statistically speaking, there are severe problems, in the exit polls and/or the election results. How such problems arose -- whether they were intentional or otherwise -- and how anyone deals with them are, of course, separate, non-mathematical issues.

Fundamentally, however, it is irrelevant how any political leaders or parties deal with voting issues: Each vote belongs to one citizen, no one else, so this is ultimately a matter for the PEOPLE to press, or not.

The second general area of feedback I have gotten concerns the math. I'm sorry it was so long and involved, but that's just the way it is -- if a case is to be made mathematically, it must be exhaustive and detailed. Cheating is easy; truth is hard.

But perhaps I can better sum up my findings, in everyday terms -- after all, statistics is not science fiction; it is ultimately the application of common sense to make sense of otherwise complex and potentially deceptive problems.

In this case, my central finding is that there is little chance of having exit poll results for one candidate -- let alone two -- in one [national] poll -- let alone many [state polls] -- be so far off just due to "random" factors.

Based on the national exit poll data I personally collected from CNN on Election Night and yesterday, I am forced to conclude that there were other, "non-random" factors -- intentional or otherwise -- affecting the outcome of the exit polls and/or the reported votes.

That is exactly the same conclusion I would have to come to if this were a scientific study, say, comparing the average weights of two breeds of cattle: If the average weight of a large sample of one breed was significantly larger than the average weight of a large sample of the other breed, I would have to conclude, with a high degree of statistical confidence, that the first breed was on average heavier than the second. That's how virtually all objective scientific experiments are analyzed.

Subjectively, I disdain "conspiracy theories" in general; objectively, most prove either unprovable or highly unlikely...like the proposition that the exit polls and/or reported voting results in this election were not in significant error. And I have yet to here anyone explain this, other than to give unproven or unprovable hypotheses or other, often biased speculation.

The thing that causes me the most concern is that the exit polls were numerous [Actually, one organization was contracted to supply exit poll results for all the major media] and the CNN national exit poll was (apparently, undoubtedly) composed of numerous state exit polls and that the actual election returns were from numerous polling stations and aggregating centers across the country: For such significant error to occur in the overall results, there had to be either numerous errors on down the line, in the exit polls and/or the voting returns, or -- more reasonably, more statistically likely -- much fewer but larger errors in central reporting locations.

A democracy having such significant problems in voting is not unlike a vertebrate having significant problems in its central nervous system. Left unchecked, they can lead to paralysis, insanity, or death.

Once again, I will happily crunch other verifiable numbers, which I have not seen (and I admit to not being "omniscient"); but any other results would have to be reconciled with the analysis of this other, factual data.

Nevertheless, although we must always courageously seek the truth, we Democrats cannot let ourselves get bogged down in the past; we have important political considerations for the future to pursue, in making our message more acceptable to more Americans. Conspiracies or no, I doubt we can win close elections; we shouldn't have to.