FL - Pinellas County Precinct Analysis (Bubble Chart)

[lostnfound]

A previous thread of mine pointed out that the only 4 counties in Florida which had Sequoia voting machines (with WinEds 2.6.220) had a remarkably consistent increase in growth rates (9%-11%) between 2000 and 2004.

I wanted to look at whether any consistent patterns flowed down to the precinct level. But between 2000 and 2004 the shapes of many Pinellas precincts were redrawn, so it is difficult to compare 2000 to 2004 at the precinct level. Using redistricting maps and files from the Florida Secretary of State and the Pinellas County Department of Elections websites, I was able to identify 65 precincts which maintained a constant shape during this time. Plotting the data for vote growth by party between elections, you can see that the gains by Bush exceed those by Democrats in 64 out of 65 precincts.

Note that total votes cast do not include early votes (which favored Kerry, worth 15% of total votes) nor absentee ballots (which favored Bush, worth 19% of total votes).

While I can't draw any conclusions from this yet, 64 out of 65 precincts seems to me to be an unlikely "success" record in a diverse metropolitan county like Pinellas. I also checked voter registration changes for each precinct (not shown); growth in registrations strongly favored the Democrats.

I intend to do more work on this and on the 3 other Sequoia counties, but wanted to post this as food for thought, FWIW.

[intheflow]

1. Thanks, lostnfound.

You number-crunching people amaze me. This is an intersting analysis. Keep it up!

[illuminaughty]

2. Can you check these counties?

Pinellas (already stated)
Hillsborough
Indian River
Citrus
Marion

This is 22% of FLorida voting population.
ALso google Marc Fratello

Interesting bio to say the least

[lostnfound]

7. I was planning to do 3 of those..

(Pinellas, Hillsborough, Indian River) and I will try to do them all.

[MrUnderhill]

3. But what does this prove?

If the election went as advertised (i.e. much higher turnout and a Bush "win") wouldn't you expect him to outperform?

Kerry got more votes than Gore did in virtually every precinct in the state. Had Bush merely repeated his 2000 numbers (assuming THEY were accurate), it would be Republicans looking at these kinds of charts wondering how Kerry had magically done so well.

[lostnfound]

6. Doesn't prove anything, except that *'s magic seems to have worked across

all demographics -- some of these precincts are WASP, others minority.

The purpose of doing this plot was to see if precincts lined up neatly along a 10% offset the way the 4 counties did. IF they had lined up neatly along a 10% line, it would have been strong circumstancial evidence of a skewed tabulator program.

Also by breaking it down to smaller groups(precincts) it might be possible to spot outliers that indicate machine "malfunctions".

The fact that *'s "magic" worked across all demographics could be indicative that
1) *'s campaign was incredibly effective and he is simply very popular
OR
2) what was a 50-50 race was systematically skewed using simple edits to the central tabulator software. Take a normal distribution, sprinkled around the 1:1 line shown on the chart, add 10%, and you end up with nearly-zero below the line.

This is a work in progress; I'm sharing it just for those who are interested.

[Critical Thinker]

4. Good stuff. I wonder why

1) why do so many precincts fall into the 3rd quadrant, i.e. NEGATIVE growth for BOTH repub and democrat? Was this because more folks took advantage of the early voting and election day turnout was down as a consequence?

2) IMO, the real oddities brought to light by the plot are the precincts FARTHEST from the origin - e.g. what's up with the precinct in the lower left corner (-57%, -36%)... Did half their population die since 2000, or did the "missing half" just vote early?

[lostnfound]

5. Yes, I think it is early voting & absentees..

which account for about 30% of Pinellas County votes, much more than in 2000.

Re, #2: I am going to take a special look at the extremes, and also doublecheck the precinct maps.

[goclark]

8. Great work lost and found ,send to Bev asap

Also post your info on the DU thread for Eloriel txs

[Zan_of_Texas]

9. I like the thought process.

Anecdotally, I know a lot of informed liberals who voted mail-in ballot because they thought it would be more secure than electronic.

The most liberal, informed precincts may have more of these ballots, which escaped this particular look at the votes.

Isn't Katherine Harris the US Representative of Pinellas? Never mind--she's across the bay at Sarasota.

If you look at the second chart on this:

http://ideamouth.com/floridacounties.htm

Pinellas is among those CLOSEST to the predicted vote, so least likely to be skewed. Other counties are WAY farther off, according to that analysis. I don't know if their analysis holds water -- just throwing it out there.

[lostnfound]

11. Possible skimming of Sequoia counties, via central tabulator

Skimming 5% (a swing of 10%) from 3 of Florida's largest counties could be accomplished via a simple edit to the WinEds 2.6.220 tabulator software, and would yield a 100,000 vote advantage to Shrub.

[Critical Thinker]

12. Hmmm, interesting plot

...it suggests that: when viewed against 2000 data, there was a higher growth in republican voting, despite the higher growth in democratic registrations. Kinda counterintuitive, ain't it?

lostnfound, are the VoteGrowth percentages in the plot above based on election day balloting or total votes ?

[lostnfound]

13. All votes as of 11-4-04..doesn't include major changes in Palm Beach

or minor changes in theother counties which have occurred since then.

Just checked how chart looks with latest official results -- Palm Beach really changes drastically with an additional 38,455 for * and 53,657 for Kerry in post-election day changes (absentees? provisionals?).


I think that the additional votes have no effect on the hypothesis of manipulation at the central tabulator.

[liquiduniverse]

14. This chart is based on registered Dems and Repugs

It doesn't take into account the large number of independent voters who typically sway Democratic in Pinellas. That's why Gore got more votes than Bush in 2000 despite having less Democrats than Republicans in the county.

[lostnfound]

15. True -- but it's also based on past voting behavior not just registrations

and therefore reflects the relative proportions of independents that existed in each county in the last election (by virtue of their votes cast in 2000).

[bmoney07]

10. Very interesting

Could you add Pasco on the list???? and ,aybe Sarasota should be added as well??

7: