Tag: Al Gore

USA Today reported that the Supreme Court unanimously agreed to vacate the Florida Supreme Court decision and ask it to clarify its reasoning. Not true. The ruling was issued per curiam (in the name of the court), and does not break out how individual justices voted.

As for the ruling itself, this was a pretty clever way to deal with the situation — essentially a warning shot to the Florida Supreme Court to stop creating law rather than interpreting it, but at the same time trying to sidestep the issue of whether the Supreme Court should substitute it’s own judgment for that of a state court.

Attorney Dennis Newman is currently leading the Gore team responsible for handling legal challenges to votes in Florida. Among the efforts Newman is spearheading is a lawsuit aimed at forcing some counties to consider so-called “dimpled” ballots — where there is a slight indentation on the ballot — as actual votes.

The Palm Beach Post reports, however, that in 1996 Newman represented Democrat Philip Johnston who was declared the winner in a Boston race on election day by 266 votes, only to end up losing the election by 108 votes after a recount. Amazingly enough, when it became clear that counting dimpled ballots favored the Republican candidate, Newman argued that considering such ballots was wrong and that the so-called chads might fall out on their own from repeated handling.

Ah, the fresh smell of hypocrisy wafting through the land.

Could Pat Buchanan really have received 3,000+ votes in Palm Beach, Flordia? Of course not, scoff the media and Democrats. To prove it a chart plotting total votes received by Buchanan in Palm Beach has been repeatedly shown on numerous television stations and printed in quite a few newspapers — the chart is very effective since the Palm Beach vote totals dwarf Buchanan’s vote totals in other counties.

In fact the chart is a statistical lie. As the Independent Institute points out, the important thing to know when deciding if Buchanan received a disproportionate number of votes in Palm Beach is to compare the percentage of votes he received in that county compared to the rest of Florida. Plot the chart like that, and the result is unspectacular.

The full press release and explanation behind the plot is at the Independent Institute’s web site, but here’s Buchanan’s votes plotted as a percentage of total votes cast,

In an article for Salon, Alicia Montgomery writes about something that had me laughing out loud while watching the news. There was a spokesman for Al Gore saying that the Florida mess should be left to the states, while George W. Bush’s folks were running around saying no, we need federal intervention here.

Expect Bush to start running ads featuring a Florida voter saying that because Gore opposed federal remedies he feel like his vote was taken away from him a second time, while Gore will respond with promises to return power back to the people.

I keep running across different statistical analyses of the Palm Beach voting and hand re-count that try to make the case that Bush or Gore is trying to steal the election. Unfortunately I have yet to see one that wasn’t just garbage.

A few days ago, for example, Slashdot linked to an MIT statistical analysis that compared the ratio of Buchanan voters to Bush voters by county. Of course the effect of the resulting graph made it look “obvious” that the Buchanan votes had to be in error. The only problem is, as one Slashdotter noted, there is no scientific reason for choosing to compare Buchanan’s votes to Bush’s votes and comparing say Buchanan’s vote to average rainfall or the number of Britney Spears songs played on local radio stations. The Buchanan:Bush ratio appeared to have been chosen because it would produce the most dramatic graph.

Similarly, a statistical analysis of the hand re-count by Bruce Borders on the conservative FrontPage.Com site argues that re-count disproportionately favored Gore. I was very intrigued by this until I read the following paragraph,

In large democratic leaning counties, the percentage change in votes was significantly larger for Al Gore than for George W. Bush. In these counties (Broward, Miami-Dade, Orange, Palm Beach and Pinellas) the percentage increase in votes for Al Gore is approximately 9 times greater than the percentage increase in votes for George W. Bush (0.068% increase for Gore vs 0.0077% for Bush). This is a statistically significant difference at the 0.108 level of probability – this means that we are approximately 90% confident that we are drawing the correct conclusion.

The last sentence is what makes me think this analysis is bunk. Although I have no probability data, for the most part people who report statistically signficant results at 90% confidence level are doing so because they know that if they report results for the 95% confidence level, which is a widely accepted level of confidence, the results no longer are meaningful.

Pharmaceutical companies are notorious for this. Got a drug that doesn’t seem to work all that well when work out the results at a 95% confidence level? No problem — just redo the statistics for a 90% confidence level and hope nobody notices.

A 1 in 10 chance that the results are due entirely to chance is pretty darn high given the seriousness of the claim.