Spot checks appear to be in order.
Nov. 6th, 2004 02:23 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
asherisThis would certainly explain a lot.
(The description of how to hack the vote... I knew that, in theory, for someone with a reasonably good knowledge of the system it wouldn't be that hard. I didn't suspect it would be so bloody easy, though! Odd to realize that with a little time to check around on such a machine, I could have fairly easily figured out how to do it for myself.)
Check out the chart of voting results in Florida put together by Kathy Dopp, from the 2004 election data made public by the Florida Secretary of State's Office: voting totals, with county and seat breakdowns (and a link to download all the data yourself), and voter registration numbers.
They have also completed an initial analysis of the Pennsylvania voting data. Notice the extreme difference between the two states. In PA, there is no county with a deviation of more than 100% greater or less than expected. In Florida, 26 of the state's 67 counties showed greater than 100% deviation from expected numbers. (11 of those were over 200%; 4 of them over 300%; 3 over 400%; and one over 700%! - as tabulated from data available with 98.6% of precincts reporting; numbers will have changed very, very slightly since the updated, 100% reporting is now available. I expect Ms. Dopp will have updated charts up in the near future.)
Expected votes will normally vary somewhat from the actual votes due to increased/decreased turnout and assorted other factors. Nationally, on average, registered voters chose their party's candidate about 90% of the time in 2004, leaving a 10% (or so) "fudge factor".
However, the numbers being reported in Florida jump out not only because of their extreme deviation from the expected, but because there is a distinct pattern of these deviations always being in favor of the Republicans, in heavily Democratic counties, only in the counties using optical scan voting machines. (The larger counties, many of which were using the new electronic/touch-screen voting systems, were under intense press scrutiny. The smaller counties, mainly using the optical scan ballots, were not being watched as much or as closely.)
In the larger FL counties, the Dep/Rep percentages of registered voters fairly accurately predicted the results. In many of the smaller counties, the Dem/Rep percentages of registered voters were almost the exact opposite of the votes reported! (ex: county with registration numbers of 70% D, 25% R, showed election results of Bush getting 65% of the vote, and Kerry getting 30%)
For there to be such a pattern, in 26 of 67 counties, suggests that something needs to be checked! Hand counting in some random precints, in the couties with the most egregious differences, would help show whether there was a corruption of the totals in the tabulator machine, or if the results really are (not likely) accurate.
For example, in the predominantly white Liberty County where over 80% of registered voters are Democrats, Bush received 699.59% more votes than expected:
When Ms. Dopp created her preliminary tables, 98.6% of the vote was in and voter turnout for Liberty county was reported as 74.1%. I updated the numbers for this one county, based on the 100% count now being reported by the FL SoS, with an updated voter turnout for Liberty county of 74.9%. Hence, Ms. Dopp's report shows a 712.3% increase above expected, while mine shows a slight downward correction to 699.59%.
(The description of how to hack the vote... I knew that, in theory, for someone with a reasonably good knowledge of the system it wouldn't be that hard. I didn't suspect it would be so bloody easy, though! Odd to realize that with a little time to check around on such a machine, I could have fairly easily figured out how to do it for myself.)
Check out the chart of voting results in Florida put together by Kathy Dopp, from the 2004 election data made public by the Florida Secretary of State's Office: voting totals, with county and seat breakdowns (and a link to download all the data yourself), and voter registration numbers.
They have also completed an initial analysis of the Pennsylvania voting data. Notice the extreme difference between the two states. In PA, there is no county with a deviation of more than 100% greater or less than expected. In Florida, 26 of the state's 67 counties showed greater than 100% deviation from expected numbers. (11 of those were over 200%; 4 of them over 300%; 3 over 400%; and one over 700%! - as tabulated from data available with 98.6% of precincts reporting; numbers will have changed very, very slightly since the updated, 100% reporting is now available. I expect Ms. Dopp will have updated charts up in the near future.)
Expected votes will normally vary somewhat from the actual votes due to increased/decreased turnout and assorted other factors. Nationally, on average, registered voters chose their party's candidate about 90% of the time in 2004, leaving a 10% (or so) "fudge factor".
However, the numbers being reported in Florida jump out not only because of their extreme deviation from the expected, but because there is a distinct pattern of these deviations always being in favor of the Republicans, in heavily Democratic counties, only in the counties using optical scan voting machines. (The larger counties, many of which were using the new electronic/touch-screen voting systems, were under intense press scrutiny. The smaller counties, mainly using the optical scan ballots, were not being watched as much or as closely.)
In the larger FL counties, the Dep/Rep percentages of registered voters fairly accurately predicted the results. In many of the smaller counties, the Dem/Rep percentages of registered voters were almost the exact opposite of the votes reported! (ex: county with registration numbers of 70% D, 25% R, showed election results of Bush getting 65% of the vote, and Kerry getting 30%)
For there to be such a pattern, in 26 of 67 counties, suggests that something needs to be checked! Hand counting in some random precints, in the couties with the most egregious differences, would help show whether there was a corruption of the totals in the tabulator machine, or if the results really are (not likely) accurate.
For example, in the predominantly white Liberty County where over 80% of registered voters are Democrats, Bush received 699.59% more votes than expected:
| Reg Voters by Party Affil. |
Probable (est) results, based on turnout % |
Election results, from FL SOS Office |
Deviation from probable (est) results | |||||||||
| Rep | Dem | NA | Total | Bush | Kerry | Total | Bush | Kerry | Other | Total | ||
| 7.9% | 88.3% | 3.8% | 4075 | 241 | 2695 | 74.9% | 1927 | 1070 | 24 | 3021 | ||
When Ms. Dopp created her preliminary tables, 98.6% of the vote was in and voter turnout for Liberty county was reported as 74.1%. I updated the numbers for this one county, based on the 100% count now being reported by the FL SoS, with an updated voter turnout for Liberty county of 74.9%. Hence, Ms. Dopp's report shows a 712.3% increase above expected, while mine shows a slight downward correction to 699.59%.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
Addendum:
Date: 2004-11-06 10:57 pm (UTC)Margin
Total equals Dem + Repub + undeclared + other.
This county may have been a bad example to use. The people there seem to have a definite disconnect between declared party affiliation, and voting patterns!