Original URL: http://www.theregister.co.uk/2004/11/12/florida_e-vote_conspiracy_theories_grow/

Florida e-vote conspiracy theories grow

Touch screens OK, optical scanners a puzzle

By Thomas C Greene

Posted in Government, 12th November 2004 21:08 GMT

There's been a good deal of Internet chatter about the possibility that the Florida vote was rigged, although optical scan machines, not touch-screen systems, are suspected. A pair of charts provided by Kathy Dopp breaks down the county vote in relation to registration rolls.

Before the election, everyone in the tech press (El Reg included) assumed that touch-screen machines were destined lead us all to months of confusion, doubt and frustration. Ironically, the Florida counties using touch-screen machines showed the fewest surprises, as reflected in the first chart.

The real puzzles come in the counties using paper ballots read by optical scanners, as shown in the second chart. It is a dramatic difference, one has got to admit. And it is odd that Kerry lost Florida precincts in which he polled high, and in which Al Gore previously did well, then lost them decisively.

Strictly speaking, this is not evidence of vote rigging: it is evidence only of the fact that the election results were surprising. There are many innocent reasons why Bush might have done better than expected: Democrats and Independents might have voted for him in significant numbers, for example.

It gets weird only in view of other factors. For example, it is undeniably strange that counties using touch-screen machines should have posted results that conform closely to registration rolls and exit polls, while counties using a different type of electronic tabulation should be far off the mark. However, we need to balance that with the fact that both Diebold's and ES&S's optical scan machines show the same peculiar results, which means that any rigging would have to be pulled off either 1. without cooperation from either of the vendors, or 2. with cooperation from both vendors together. Neither is terribly likely.

But there are other factors. While a departure from the registration rolls can be explained easily by cross-party voting, it is more difficult to dismiss departures from exit polls, which tend to be fairly accurate. When they're off, they're usually off by a relatively small margin, and they tend to break for both candidates in different regions. But in this case, we find only one candidate, John Kerry, over-reported by exit polls. That's not quite what one would expect.

Republicans have offered a few incredibly lame explanations, such as the one holding that they, being a naturally modest folk, were less likely to consent to be interviewed. But if exit polls had been reliable in the past, this factor must have been calculated into them, or it is an incredibly recent phenomenon. (Where they might have got a sudden blush of modesty after screaming bloody murder about the Liberal Gomorrah that mysteriously sprang up about them while they were in possession of the federal bureaucracy, both chambers of Congress, and the Supreme Court is difficult to pinpoint, but we're looking into it.)

Another issue in Florida is the simple fact that its Governor is the President's brother. This is hardly sufficient to cast suspicion on the process, but it is, perhaps, enough to withhold an automatic presumption of innocence. However, this needs to be balanced with the fact that Governor Bush has, over the past four years, grown significantly more popular among Floridians than he was in 2000. His lobbying for the President may well have had an effect among Democrats and Independents. Still, his close relationship with the President is reason to approach the Florida election results with a measure of caution. Not cynicism, mind, just healthy skepticism.

The wind up here is that there are a few intriguing questions about the Florida vote that deserve good, straight answers. The anomalies should be investigated, although the findings are unlikely to be decisive. But we will learn what we need to know about the reliability of optical scan machines, and the vendors who make them, whether the news should be good or bad.

The President's margin of victory is in the range of 3.6 million votes nationwide. So even if there has been mischief in a few counties here and there, the margin of victory is quite likely to exceed the margin of doubt. Thus, it appears that a failed Texas oilman and cowboy wannabe really is President of the single most influential nation on Earth.

This time around, anyway. ®

Related Stories

E-voting promises US election tragicomedy E-voting security: getting it right