While researching a forthcoming blog entry of the American (effectively) two-party political system I heard an interesting radio program about how error prone the US election system is. Citing a the Voting Technology Study by Caltech/MIT, the organization Election Watch is saying that in the 2000 election an estimated 4 to 6 million votes were simply lost for one reason or another.
This is indeed a pretty shocking number considering that the number of votes cast in 2000 was 100 million, so 4 to 6 million "lost" votes means an error of 4 to 6%. Think about it - Gore won the peoples (popular) vote by a margin of only 0.5%, and Bush won the electoral college vote by 0.9%. Both figures being well below the 4 to 6% of lost votes, so one could argue, as Election Watch has, that doing recounts to settle the issue of whether Bush really won by a few hundred votes in Florida is largely irrelevant compared to the larger problem of 4 to 6% of all votes being lost.
Add to this the impact of the electoral college system which converts even the slightest majority in a state's popular vote to a 100% win of the electoral college votes for that state. Not to mention the inbuilt bias of the electoral college system that gives many states a great say in the electoral college than their percentage of the nations population. For instance Wyoming gets 0.56% of the electoral college votes, but has only 0.19% of the voting age population. Whereas California gets 10% of votes for its 10% of voting age population.
In Florida 2000 the "win" of the state's popular vote by 537 votes, or just 0.009% converted into a winner-takes-all victory of Florida's 25 electoral college votes which is 4.7% of the total. That's quite some leverage achieved through the electoral college system: 0.009% to 4.7% - a more than five hundred fold amplification of the majority in Florida.
All these figures should serve to increase our alarm at a figure of 4 to 6% of all votes being lost in a national election. Now a statistician will tell you that even if not all votes are being counted then the ones counted are basically a large enough representative sample of all votes that would have been counted. Within the margin of error the vote result would have been the same. This is the same principle by which opinion polls and exit polls purport to determine the result of the election before any actual vote counting takes place - the just count a sample and a tiny one at that. Basic statistics will then predict the level of confidence in the result (probability of being right) for a given sample size and pollsters pick a sample large enough to give them a high enough confidence in being right - 95%, 99%, 99.9%... you name it. All the way to 100% by counting every single vote.
Of course pollsters don't count actual votes - people lie about whether they will really vote and who they will or have voted for. Also the sample of people they select is frequently biased - even calling people by telephone builds in a bias because democrats tend to hang up on pollsters, it takes a certain degree of wealth and credit worthiness to even have a phone, less affluent families are more likely to have a listed phone number (as opposed to paying for an unlisted one not available to pollsters). Plus its any guess as to how employment status affects the chance of being home at any particular time of day when the pollsters might ring. All these have to be considered and somehow "corrected" for with plenty of room for error.
Now the Voting Technology Project estimated that in the 2000 election 1.5 to 2 million votes were lost due to faulty equipment and confusing ballots, 1.5 to 3 million were lost because of voter registration problems (I know at least one person who had that problem), and one million were lost due to problems at the polling place like long lines or poor access to polls. In addition there was an unknown loss due to problems with absentee ballots.
So the question is if 100 million votes are counted, and another 4 to 6 million are not then is a win by 0.9% of the electoral college, or (loss by) 0.5% of the popular vote valid? Is the actually number of votes counted a large enough sample to ensure a very high degree of confidence? Indeed what should be the level of confidence attached to a presidential election result? 99%, 99.9% or 100%?
Even more importantly, what is the distribution of intended votes in those 4 to 6 million lost in the 2000 election? What if it had even a 10% bias to Democrats? We've already learned that in Florida an estimated 40,000 people were incorrectly purged from the electoral role, a group that had a strong bias towards democratic voting African-American voters. What if that bias were replicated elsewhere? What if its mostly poorer voters who are not able to find time to register to vote, or fill out the form wrongly due to literacy problems? What if it is people working the longest hours for least money who can't afford time off work to stand in line to vote?
Yes I'm sure there are biases in the other direction like too many rich republican voting white guys who said they wanted to vote but just couldn't be bothered to (in this case voter apathy probably goes both ways), but my strong feeling is biases probably lean toward excluding more DNC votes than GOP votes, and by a healthy margin too.
Think about it - if the 4 to 6% rate of lost votes were applied to Florida 2000 that means between 238,000 and 358,000 lost votes. So the Bush margin of "victory" in Florida by 537 votes could have been wiped out by counting the lost votes if they contained just 0.24% to 0.15% more Gore votes than Bush votes (depending on how many votes were actually lost). No wonder people wanted to recount the votes and no wonder people are getting excited and indeed very angry about the laughable status of our US voting system.
Yes it may be better than many third world nations, but really is it befitting a nation that claims to be #1 of everything in the world and the squeaky clean image for democracy that everyone should aspire to? Election results as in 2000 determined by tiny margins are hardly victories, they are more lucky draws plucked out from the political noise of closely divided nation.
Indeed what is up with a country that hands such power to a single person when he is elected by such a tiny margin? Why should one gang of politicians get supreme power for getting less than a 1% more votes in the electoral college? Why should that be converted to such a radical shift in power in the country? Should we insist on even a 10% majority or even larger before one gang gets to lord it over the other?
Remember if you have 100 people evenly split at 50:50 and five change their minds that results in a 55:45 split - a 10% lead even though only 5% of the voters changed their minds. Even with a 10% majority would you like to feel the nations future was at the whim of just one in twenty people changing their minds? Should we require something like a 2/3 majority before one party gets some massive advantage over the other like having their gangs leader being President of the whole nation? At least with the 2/3 requirement you have to change the minds of one in three people to reverse the result - a significant achievement.
Ultimately I believe this amplification of political power achieved by the presidential system, the electoral college and the effective two party system in the USA has lead to all parties dumming down their manifestos to muddled and misleading mediocrity. Likewise voters have similarly dummed down their personal ambitions and expectations for how the country should be run. In effect we have a mediocracy of government serving a mediocrity of citizens. But more on that in a future blog entry...
No comments:
Post a Comment