tag:blogger.com,1999:blog-4918582430341367895.post7118832116249899262..comments2023-05-29T04:22:36.745-07:00Comments on Continuation: Statistical exploration of New Hampshire Democratic primary resultsUnknownnoreply@blogger.comBlogger10125tag:blogger.com,1999:blog-4918582430341367895.post-51600986374213274602008-01-14T11:36:00.000-08:002008-01-14T11:36:00.000-08:00Whoops, it appears those machines were ushered in ...Whoops, it appears those machines were ushered in just last year, and so there are no previous records for NH...<BR/><BR/>can anyone find the results from other states?Flipperbwhttps://www.blogger.com/profile/00563262673393490534noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-4204123799793647452008-01-13T22:15:00.000-08:002008-01-13T22:15:00.000-08:00I have an idea. I haven't seen this suggested anyw...I have an idea. I haven't seen this suggested anywhere. <BR/><BR/>What I want to do is find the results from the 2004, 2000, 1996 etc. New Hampshire primaries, and take a look at the differences between the Diebold and Hand-count votes for those elections. If we could show that it is very rare to have such a large discrepancy between machine vs hand, I think we'd have a pretty strong case.<BR/><BR/>But I can't find those documents. Anyone know where we can find them? This would be a smoking gun if we could prove it.<BR/><BR/>Anyone want to help me take this on?Flipperbwhttps://www.blogger.com/profile/00563262673393490534noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-763168817720069562008-01-13T15:54:00.000-08:002008-01-13T15:54:00.000-08:00I'm no stats expert, and this could be a spurious ...I'm no stats expert, and this could be a spurious question, but I think you have a problem in the model. In a linear multivariate regression, the explanatory variables should be uncorrelated; otherwise, some of the effect of one could be attributed to the other. It's called multicollinearity.<BR/><BR/>If size and method are so highly correlated, you have a problem: you cannot use a simple regression like this to judge the separate effect of each.<BR/><BR/>Like I said, I'm no stats expert, but if I'm just seeing things, I'd really appreciate it if you'd explain why it's not an issue.Anonymoushttps://www.blogger.com/profile/10642979708332060535noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-85537850073846014152008-01-13T13:59:00.000-08:002008-01-13T13:59:00.000-08:00Thursday 1/10: Bruce O'Dell writes:Theron Horton a...Thursday 1/10: Bruce O'Dell writes:<BR/><BR/><I>Theron Horton and I have confirmed that based on the official results on the New Hampshire Secretary of State web site, there is a remarkable relationship between Obama and Clinton votes, when you look at votes tabulated by op-scan versus votes tabulated by hand:<BR/><BR/>Clinton Optical scan 91,717 52.95%<BR/>Obama Optical scan 81,495 47.05%<BR/><BR/>Clinton Hand-counted 20,889 47.05%<BR/>Obama Hand-counted 23,509 52.95%<BR/><BR/>The percentages appear to be swapped. This seems highly unusual.<BR/><BR/>Recall that the specific model of Diebold op-scan [1.94w] and central tabulator in use in New Hampshire are proven by demonstration [Hursti Hack] to be vulnerable to insider manipulation.<BR/><BR/>Theron Horton and I are proceeding with the intra-county and demographic analysis.<BR/><BR/>More to come.<BR/><BR/>Bruce O'Dell<BR/>Co-Coordinator for Data Analysis<BR/>Election Defense Alliance<BR/>Bodell[at]ElectionDefenseAlliance[dot]org</I><BR/><BR/>What is going on here?<BR/><BR/>You already demonstrated extreme statistical anomalies. Now add this one. These numbers check out to .0001% variance. <BR/><BR/>What kind of spreadsheet code could yield such extreme events? Is it possible that this extreme statistical event is the result of a program that is calculating the computer counted votes as a function of the hand counted votes? <BR/><BR/>I'm assuming that the hand counted votes cannot be altered by the central tabulator but the computer counted votes can be altered. <BR/><BR/>Perhaps, the central tabulator is not only receiving info from subordinate terminals but it is also sending info to the subordinate terminals, adjusting the numbers according to a preset formula. <BR/><BR/>Let me know what you think. Could you write a program that would similate these events?Anonymoushttps://www.blogger.com/profile/02232277072775871371noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-55954580424187069102008-01-13T10:48:00.000-08:002008-01-13T10:48:00.000-08:00Thanks for all the comments, I'll re-run the scrip...Thanks for all the comments, I'll re-run the scripts with the new data shortly.semmelweishttps://www.blogger.com/profile/02637848845905304106noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-5922807260027290872008-01-12T16:48:00.000-08:002008-01-12T16:48:00.000-08:00Here are final results of NH primary in easy to re...Here are final results of NH primary in easy to read chart. <BR/><BR/>http://www.boston.com/news/politics/2008/nh/nh_primary_dem_results_by_town/Anonymoushttps://www.blogger.com/profile/02232277072775871371noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-75683028344634924162008-01-12T15:40:00.000-08:002008-01-12T15:40:00.000-08:00It appears that the Google spreadsheet you referen...It appears that the Google spreadsheet you reference has outdated/questionable hand vs. Diebold assignments. It doesn't match the NH government listing at <A HREF="http://www.sos.nh.gov/voting%20machines2006.htm" REL="nofollow">http://www.sos.nh.gov/voting%20machines2006.htm</A>. There are many discrepancies, (e.g. Claremont is Diebold). I wouldn't tend to trust the data source used for the spreadsheet-- Manchester being listed as "unknown" is dubious.<BR/><BR/>I'd like to see a rerun of your analysis with the corrected data.<BR/><BR/>Also, out of curiosity, I'd like to see a run with Clinton and Obama's Diebold votes swapped.Anonymoushttps://www.blogger.com/profile/03358654605558180041noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-66430439898502673732008-01-12T13:21:00.000-08:002008-01-12T13:21:00.000-08:00Look at the financing behind Diebold and ES&S. The...Look at the financing behind Diebold and ES&S. There is a heavy Jewish influence. Clinton has scored big with the Israeli lobby because of her "we will take no options off the table" stance on Iran. One of the big Isreali news organizations ranked her second behind Giuliani in terms of her willingness to serve Israeli interests. <BR/><BR/>This alone does not prove fraud.Anonymoushttps://www.blogger.com/profile/02232277072775871371noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-28474018855741227152008-01-12T11:12:00.000-08:002008-01-12T11:12:00.000-08:00This is great work, thanks!For good measure I'm go...This is great work, thanks!<BR/><BR/>For good measure I'm going to try to reproduce your calculations, but the argument is solid.<BR/><BR/>Would you like to cross-post your entry on the community blog EuroTrib.com?<BR/><BR/>It's good enough to be on the recommended diary list on Daily Kos, IMHO.Miguelhttps://www.blogger.com/profile/06568794455475900309noreply@blogger.comtag:blogger.com,1999:blog-4918582430341367895.post-34708750954634232632008-01-11T22:04:00.000-08:002008-01-11T22:04:00.000-08:00Gotta love R when it comes to statistics. It's ref...Gotta love R when it comes to statistics. It's refreshing to see you do the exact (basic) analysis I had in mind -- when blocking for district/precinct size, voting method was very significant.<BR/><BR/>To put this in perspective for non-stats people or at least assist with the understanding, the p-value is the chance (probability) that we would get the result or a more extreme result if the corresponding variable actually doesn't matter. That is, if the variable didn't actually matter, it is the chance of getting this extreme of a result (and the p-value is so extreme for the counting method that it is like flipping a coin 14 times and getting heads every single time... very unlikely).David Diezhttps://www.blogger.com/profile/03572066063477066829noreply@blogger.com