Log in

No account? Create an account
color cycle (slow)

Kistaro Windrider, Reptillian Situation Assessor

Unfortunately, I Really Am That Nerdy

Previous Entry Share Next Entry
Ack Basswards
color cycle (slow)
Does it seem off to you- maybe even just a little- that for small Florida counties, Democrat registration was inversely proportional to Kerry votes? That the more people who registered as Democrat, the better Bush did?

Also see the diagram at this link, which compares exit polls to results. They seem to be pretty accurate- when the state isn't using electronic voting macines. But when they do, the exit polls are very consistently showing a better response for Kerry than was actually voted.

Imagine that. What a coincidence.

Yes, there is a conspiracy theory here, and I'm giving it some serious credibility.

  • 1
Um, er, or maybe exit polls suck.

But only in locations handled by electronic centralized vote counting, and they're pretty accurate in other places. Interesting coincidence, then.

C'mon! You took probability. I'm sure the confidence level is pretty high before they release these things..


I did *not* take probability!

The data presented in those graphs are unconvincing for a few reasons. Only 3 paper ballot graphs and 6 electronic ballot graphs are presented. Showing us a wider range of states to prove the paper ballot and electronic voting graphs shown were not selected for their disparity. The graphs don't specify if they are statewide results or only results in select precincts. For example, Ohio is shown as an electronic voting state, and while much of Ohio did indeed use the voting machines, a significant portion used paper ballots (and if I remember correctly I read that more precincts used paper ballots than voting machines). The graphs do not make that distinction. Lastly, the margin of Kerry's victory in the exit polls in New Hampshire and Pennsylvania (~15% and 20% respectively) are seriously out of line with numerous and independent professional polls taken in the days leading up to the election in those states. Without citing a source the data in those graphs are particularly suspect.

The selective use of statistical data by various bloggers as "evidence" of vote tampering is not enough to sway me. While the link you provided regarding the inverse relationship between registration and votes in Florida is interesting, until they have been through a rigorous statistical analysis by a reputable and unbiased source I must continue to consider it merely interesting as opposed to compelling. The very fact that the smaller counties seemed to have a larger divergence between registration and actual votes can also be attributed to the fact that the smaller set of numbers involved is more prone to error - statistically speaking, larger samples are inherently less likely to show false trends in the data.

All in all, while I won't discount the idea that there were irregularities in the useage of voting machines, I am still at this point unwilling to attribute them to a willful conspiracy to make sure that Bush retained office. I am more likely to subscribe to the theory that our country is increasingly populated by morons and thus more likely to vote for morons :p

I agree that the diagram I showed has far too little data to have meaning; it's just more an interesting eyebrow-raising observation than anything major. I would like to point out http://www.blackboxvoting.org/ for more information on possible ballot issues, though.

I haven't wholeheartedly thrown myself behind the "there's fraud here" camp, I just suspect they may be on to something.

Maybe even just a little, yes :)

As for exit polls, I was using Google and came across this article from the Khaleej Times (in the U.A.E.).

As far as overall election results of predicted versus actual... I did some number crunching...
State%Change Dem%Change Rep
Florida isn't too far a stretch there in difference between "predicted" (based on party registration) and "actual" (based on reported results). (Admittedly my sample size is ridiculously small, but already Florida's numbers look relatively tame compared to Massachusetts...)

I suspect that if there really is some truth to the conspiracy theory that the popular newsmedia would report on it pretty quickly, as such a controversy would be big ratings (== big money).

There are three kinds of lies: lies, damned lies, and statistics.
--Mark Twain*

The popular media never reported on the major flaws in the Diebold voting machines that would allow vote tampering.

As best I can tell, the "major flaws" consist of someone breaching physical security of the machines. If that can be accomplished, there is a significant chance that physical security of the paper ballots can be breached as well, no?

It's far, far easier to do one, than the other.

Perhaps, yes. But it's not like they hook up the voting machines to the internet (I hope!). As long as physical security is maintained, then the only extra problem is bad programming (which is a big issue, yes). If you can't trust your election officials, you can't trust them. It doesn't matter what your voting method is...right?

These results being suspected here were actually based on optically scanned paper ballots. If it's a problem, let's look at the ballots themselves by hand recount. We did it before...

Guess what? They hook up the voting machines to the Internet.

Well, not quite. The central tabulators are connected by modem to a phone line; it's direct phone communication using Internet-like protocols, but it's not directly connected to the Internet.

Oh, and there is a known hack that works even if the modem is set to "inactive." It has to be unplugged. It's got the front page at http://blackboxvoting.org if you want more information...

You're right... Allowing the vote-counting machine(s) to accept incoming calls isn't too bright. Yeah...

I downloaded the 1988 election results from Florida... <th>Country</th><th>Dukakis</th><th>Bush</th></tr>

I chose these counties because they had particularly high %Change Rep and "high" negative %change Dem. on this page. Unless these counties have drastically changed party association in the last few years, I would take it as a sign that, despite their Democratic registration, they vote Republican for President. All the other years from 1980 to present are available from the Florida Election Results page, but I don't feel up to trudging through spreadsheets from the other years at the moment. It does weaken the conspiracy theory a bit, though.

And of course by "Country" in my table I mean "County"... blargh.

  • 1