Dissonance Amongst the Delegates

CBS News and the New York Times just conducted a poll among Democratic Delegates asking the VP question. The results, Clinton received 28 (271.6 votes) percent of the vote, while Biden, the next closest responded, received 6 percent. There was however an enormous percentage of participates who were undecided, 36 percent and another 9 percent selected other. I was initially skeptical of this poll's result until I analyzed the statistics present in the sample. The sample included 970 randomly selected delegates, 23.8% of the 4081 total delegates.

Before I detail my analysis I want to first establish the basic variables used in my calculation. Lets first state that as of June 3rd (the night Obama sealed the nomination) Obama had 2172 delegates (53.2%) and Clinton had 1909 (46.8%) delegates; I draw these figures from our Democratic Delegate Estimates page that was maintained throughout the primaries. The delegate poll was conducted to include both pledged delegates and super delegates but of course CBS and the NY Times did not provide the raw numbers, so I'm left with the task of trying to figure it out. Luckily though, the poll provided additional information in the secondary question; Would Clinton as VP help Obama win the election? The matrix below uses these numbers to calculate the percentage of delegates questioned from each delegate group, Obama allocated, Clinton allocated and unpledged.

Within the text of the published poll result it states that 61% of Clinton allocated pledged (Thanks John) delegates supported her as their preferred VP choice, while just 3% of Obama's delegates supported Clinton as their first choice. If we assume that the sample is perfect, each side's percentage in the sample space is the same as their percentage of pledged delegates, then Obama would have 419.99 pledged representatives and Clinton would have 384.57. Using these figures and the percentages noted in the report an estimate to the theoretical support for Clinton as a VP can be determined; 61% * 384.57 = 234.59 for Clinton delegates, 3% * 419.99 = 12.60 for Obama delegates and 20% * 165.43 = 33.09 among the super delegates for a total of 280.278 or 28.90% if the sample were perfect. The 28.90% doesn't line up with the 28% as noted in the final poll result but its very close; Clinton and Obama delegates are over represented in the poll sample while the super delegates are under represented.

Now here comes the problem with this poll. If 410.31 Clinton delegates were interviewed and 61% of pledged delegates favored Clinton as their preferred VP choice, who did the other 39% prefer? The most likely response would be undecided, then other but this conclusion provides nothing of value. In fact its a guarantee that 14% of Clinton delegates choose undecided or other, but its also possible that all 39% voted undecided or other; absolutely inconclusive. The result of the poll essentially says this, "A majority of Clinton supporters think Obama should choose Clinton as the VP. The other half, well we have no idea what they think, it could be anybody because we didn't provide enough data." But this non-result is fascinating in itself. If 39% of Clinton delegates, who are by definition her hardcore supporters, have moved away from blindly supporting Clinton then the rest of the 18 million people who voted for her likely have as well. That's not to say that they are fully supportive, but at least they've taken off their tinfoil hats. The poll also squelches any possibility of a surprise Clinton nomination when the vote is called because much of her original support has boarded the Obama train.

Published on August 19th at 12:20 AM CT :: 1 Comment

Hillary's Popular Vote Miscalculation

With news of Hillary's upcoming concession, or something equivalent, I thought I would attempt to dispel her continuing belief that she has won the popular vote. To start off, I'll illustrate the metric she is using and then move into a more thorough calculation. To begin, the table below shows the vote totals for each primary. Each race links to a government or party affiliated election results page that depicts the numbers used in this analysis. There are four races (American Somoa, Kansas, Virgin Islands, Hawaii) in which I could not find a government or party reported page detailing vote totals. For these states I cross checked results on numerous media sites to verify accuracy.

Race Obama Votes Clinton Votes Total Votes Difference (BHO-HRC) Election Date Election Type
New Hampshire 104815 112404 285040 -7589 Jan-8 Primary
Michigan 0 328309 328309 -328309 Jan-15 Primary
South Carolina 294898 140990 532151 153908 Jan-26 Primary
Florida 576214 870986 1749920 -294772 Jan-29 Primary
Alabama 300321 223096 536635 77225 Feb-5 Primary
Alaska 6674 2194 8880 4480 Feb-5 Primary
Arizona 193126 229501 456626 -36375 Feb-5 Primary
Arkansas 82476 220136 314234 -137660 Feb-5 Primary
California 2093318 2524799 4882620 -431481 Feb-5 Primary
Connecticut 179742 165426 355561 14316 Feb-5 Primary
Delaware 50467 39984 96374 10483 Feb-5 Primary
Georgia 704247 330026 1060851 374221 Feb-5 Primary
Illinois 1318234 667930 2038614 650304 Feb-5 Primary
Massachusetts 511680 705185 1263764 -193505 Feb-5 Primary
Missouri 406917 395185 825050 11732 Feb-5 Primary
New Jersey 501372 613500 1141199 -112128 Feb-5 Primary
New Mexico 71396 73105 149779 -1709 Feb-5 Primary
New York 751019 1068496 1862445 -317477 Feb-5 Primary
Oklahoma 130130 228480 414696 -98350 Feb-5 Primary
Tennessee 252874 336245 624764 -83371 Feb-5 Primary
Utah 74538 51333 428459 23205 Feb-5 Primary
Louisiana 220632 136925 384346 83707 Feb-9 Primary
DC 93386 29470 123994 63916 Feb-12 Primary
Maryland 532665 314211 878174 218454 Feb-12 Primary
Virginia 627820 349766 986203 278054 Feb-12 Primary
Wisconsin 646851 453954 1113753 192897 Feb-19 Primary
Ohio 982489 1212362 2233156 -229873 Mar-4 Primary
Rhode Island 75316 108949 186439 -33633 Mar-4 Primary
Texas 1362476 1462734 2874986 -100258 Mar-4 Primary
Vermont 91901 59806 154960 32095 Mar-4 Primary
Mississippi 265502 159221 434110 106281 Mar-11 Primary
Pennsylvania 1046822 1260937 2307759 -214115 Apr-22 Primary
Indiana 632035 646233 1278268 -14198 May-6 Primary
North Carolina 887391 657669 1580726 229722 May-6 Primary
West Virginia 91663 239187 357031 -147524 May-13 Primary
Kentucky 209903 459210 700855 -249307 May-20 Primary
Oregon 372823 258438 636680 114385 May-20 Primary
Puerto Rico 121458 263120 387299 -141662 Jun-1 Primary
Montana 102544 75053 181906 27491 Jun-3 Primary
South Dakota 43576 54015 97591 -10439 Jun-3 Primary
Hillary's Calculation 17011711 17528570 36254207 -516859
+ Uncommited MI for Obama 17249879 17528570 36492375 -278691

Totaling these numbers we are left with what the Clinton campaign is calling a decisive victory despite excluding 16 contests. Also notice that the 'uncommitted' votes from Michigan are not included in Obama's column. The Clinton campaign does not associate the 238,168 votes as Obama's because he electively chose to have his name removed from the ballot. At this juncture Clinton has a clear advantage but if I take into consideration the decision of the Rules and Bylaws Committee (RBC) pertaining to delegate allocation and apply it to raw votes a different result arises. Following their recommendation necessitates halving the votes of Florida and Michigan requiring a recalculation of primary contests.

Race Obama Votes Clinton Votes Total Votes Difference (BHO-HRC) Election Date Election Type
1/2 Votes Michigan 0 164154.5 164154.5 -164154.5 Jan-15 Primary
1/2 Votes Florida 288107 435493 874960 -147386 Jan-29 Primary
RBC Total 16723604 16928922.5 35348137 38612.5

Simply stunning; it appears that using the determination of the RBC gives Barack Obama the popular vote lead among primaries. As a sneak peak I'll just tell you that he accrues a significant lead in the "less important caucuses" ensuring a popular vote victory. At this point I'll digress and move onto the caucus numbers to see if Obama can dent Clinton's rather impressive lead using the full votes from Michigan and Florida.

Race Obama Votes Clinton Votes Total Votes Difference (BHO-HRC) Election Date Election Type
American Somoa 121 163 285 -42 Feb-5 Caucus
Colorado 80113 38839 120971 41274 Feb-5 Caucus
Idaho 16880 3655 21224 13225 Feb-5 Caucus
Kansas 27172 9462 36723 17710 Feb-5 Caucus
Minnesota 142109 68994 214066 73115 Feb-5 Caucus
North Dakota 11625 6948 19012 4677 Feb-5 Caucus
Abroad 15214 7501 23105 7713 Feb-5 Other
Nebraska 46279 43614 93757 2665 Feb-9 Caucus
Virgin Islands 1772 149 1970 1623 Feb-9 Other
Hawaii 28347 8835 37562 19512 Feb-19 Caucus
Wyoming 5378 3311 8753 2067 Mar-8 Caucus
Guam 2264 2257 4521 7 May-3 Caucus
Caucus Total 377274 193728 581949 183546

The caucus results clearly show that Obama excelled in this style of election winning all contests but American Somoa in which only 285 people voted. If we add the result of these caucuses to the primary results we get the values shown below.

Race Obama Votes Clinton Votes Total Votes Difference (BHO-HRC) Election Date Election Type
Hillary's + Caucuses 17388985 17722298 36836156 -333313
+ Uncommited MI for Obama 17627153 17722298 37074324 -95145

Here comes the hard part. At this point Hillary has a commanding lead but the four remaining contests (Iowa, Nevada, Washington, Maine) all of which Obama won, do not report or for that matter record individual votes. All calculations done from this point are estimates, but they attempt to include "the voice of all who voted," a rallying cry for the Clinton campaign in the closing days. This is probably redundant but it needs to be said, Hillary only cares about counting the votes when the votes in question favor her. To estimate the number of votes in each of the remaining four states I will calculate the ratio of votes per delegate using the primary and caucus results in which votes are reported. However, Texas still remains an issue. For this calculation we will assume that everybody who participated in the Texas caucus also participated in the primary. Crunching the numbers shows that 3,170.5 delegates correspond to 37,074,324 reported votes. A division yields 11,693.53 votes per delegate meaning that for every 11,693.53 people that voted for a given candidate, that candidate would be allocated 1 national delegate. Multiplying 11,693.53 by the number of delegates received in each state will yield our estimate; those estimates are shown in the table below. For Iowa the initial delegate allocation will be used because some delegates originally allocated to Edwards moved to Obama after the Edwards endorsement.

Race Obama Votes Clinton Votes Total Votes Difference (BHO-HRC) Election Date Election Type
Iowa 187096 163709 526209 23387 Jan-3 Caucus
Nevada 163709 128629 292338 35081 Jan-19 Caucus
Washington 619757 292338 912095 327419 Feb-9 Caucus
Maine 175403 105242 280645 70161 Feb-10 Caucus
Estimates 1145966 689918 2011287 456048

Obama makes up 450,000 votes in these four elections. Adding these totals to the previous totals we arrive at our results.

Race Obama Votes Clinton Votes Total Votes Difference (BHO-HRC) Election Date Election Type
Total + Estimates 18534951 18412216 38847443 122735
+ Uncommited MI for Obama 18773119 18412216 39085611 360903

The final results show that no matter how Michigan and Florida are counted, just as long as all other contests are counted Obama wins the popular vote. If at this point Clinton still wants to dispute the estimates for the four remaining caucuses she only has razor thin margins to fall back on and any additional estimates will likely reveal identical conclusions; for example RealClearPolitics.com conducted a similar analysis and the final results were comparable.

Published on June 5th at 3:40 AM CT :: 0 Comments

Indiana Antics

Before the two most recent polls were released in Indiana (Downs Center [4/16], LA Times/Bloomberg [4/14]) Clinton had a comfortable lead of 9.25 percent, on average. A significant shift occurred with the release of these two new polls each showing Obama positioned with a 5 percent advantage; a swing of nearly 15 percent in the span of just a week. How did Hillary manage to squander 15 percent in this crucial state? To answer this question the actions of each candidate during the previous week must be analyzed. In the prior week Barack Obama continued his campaign in Pennsylvania and at one point referred to voters as "bitter" in the face of economic hardship. He also stated that these same "bitter" voters have given up on voting for real issues but rather focus their attention on "guns and religion." At the same time Hillary Clinton toured Indiana seeking to appeal to whiskey drinkers and firearm fanatics while touting herself as the "in touch" candidate countering what the media described as Barack's "bitter" blunder. But the poll numbers don't lie and clearly illustrate that Clinton's rhetoric completely failed.

But perhaps Obama's rise in poll numbers can be attributed to something entirely different. Perhaps people actually do feel "bitter" when they lose their job for bureaucratic reasons. Perhaps his poll numbers increased among those who are in fact "bitter". Perhaps Clinton is out of touch with the real issues. Perhaps she realizes she cannot win on the issues and is instead seeking to appeal to the same demographic for which she derided Obama for referencing; the "guns and religion" vote. Perhaps all of these ifs are true, does Obama then have the nomination wrapped up? The polls appear to be paving the way in either case.

The Indiana primary takes place on May 6th, the same day as North Carolina and fourteen days after Pennsylvania. Indiana has 72 national pledged delegates.

Published on April 18th at 7:08 PM CT :: 0 Comments

Electoral College Projection Map
Senate Projection Map