Monthly Archives: November 2003

Alabama and the 1960 Popular Vote

I have received several emails with regard to this article from the Wall Street Journal’s Opinion Journal. The article makes an argument that JFK lost the popular vote to Richard Nixon in the election of 1960. The basis of the argument is not about the alleged fraud in Illinois and Texas, but rather that it is not correct to assign all of the popular votes for Alabama’s Presidential Electors to Kennedy because six of the eleven electors were officially not pledged to Kennedy. The facts of the article are mostly correct – Alabama held a primary to determine which Democratic Presidential Electors would appear on the ballot. The result, was a ticket with split loyalties – six of the eleven Electors were “unpledged” and the remaining five were “loyalists” pledged to Kennedy (this was a different situation than in Louisiana and Mississippi – where the Unpledged Electors were on a separate ticket than the Democratic Electors).

One inaccuracy in the article is the statement that the Nixon “Slate” defeated the Democratic “Slate” 324,050 to 237,981. In fact, the citizens did not vote for the electors as a slate, but actually cast eleven ballots for not more than eleven electors (if someone wanted to vote for the six unpledged electors and five of the Republican ones, they could. If they wanted to vote for just one elector, they could). The result of this voting method produced varying totals; the difference between the highest and lowest electors is about seven thousand votes (see table below).

Alabama Election Results – 1960

Frank M. Dixon U 324,050 Cecil Durham R 237,981
Bruce Henderson U 323,018 C.H. Chapman, Jr. R 237,370
Edmund Blair U 322,593 W.H. Gillespie R 236,915
C.E. Hornsby, Jr. U 322,124 J.N. Dennis R 236,765
W.W. Malone, Jr. U 322,084 Robert S. Cartledge R 236,110
Frank Mizell U 320,957 W.J. Kennamer R 235,414
C.G. Allen D 318,303 Perry O. Hooper R 234,976
C.L. Beard D 318,266 Tom McNaron R 234,856
J.E. Brantley D 317,226 Mrs. John Simpson R 234,002
Dave Archer D 317,171 T.B. Thompson R 233,450
Karl Harrison D 316,934 George Witcher R 230,951

Unpledged: Max: 324,050; Mean: 322,471; 6/11: 175,893
Democratic: Max: 318,303; Mean: 317,580; 5/11: 144,355
Republican: Max: 237,981; Mean: 235,345; 11/11: 235,345

So, the question is, how does one fairly allocate the popular vote in Alabama in 1960? Typically, most sources that I have seen give Kennedy 324,050, the highest total of the eleven electors. However, the elector whom received that total, Frank M. Dixon, was one of those not pledged to Kennedy. The highest vote total for a Kennedy elector was 318,303 (C.G. Allen). There are several ways that the results can be allocated with math (mean, median, 5/11 for Kennedy, 6/11 for Unpledged, etc.). These would all result in different totals and, from some perspective, be correct (its the “all-of-the-above” choice). The real total would be to determine the intent of the voters. When a citizen cast his/her ballot for each elector, was that citizen choosing Kennedy or Not Kennedy (some conservative democrat not yet named). Unfortunately, we can never know.

The point I would like to make about the national popular vote, is that it does not matter. The “game” isn’t played that way. As for an analogy, I haven’t seen anybody up in arms and complaining that the Yankees rightfully won the World Series in 2003 because they scored more runs (21 – 17). The Marlins won the most games… and are therefore the winners. The US Presidential Election is about winning electoral votes (which, today, translates to winning the popular vote within each state – with exception of Maine and Nebraska where a candidate’s electors are chosen by a plurality of the popular vote within each congressional district). The players understand this and execute their game plans accordingly. If the national popular vote was the actual metric by which they would win, then the campaigns would execute their plans very differently than they do today – leading to a different outcome.

Electoral College Calculator

I’ve created a simple Electoral College Calculator for all those interested in running Election 2004 scenarios. The calculator is a radio-button form where a visitor may choose whether the Electoral Votes of each state (and D.C.) will be assigned to the Democratic candidate, to the Republican candidate, or left unallocated as a tossup. A calculate button at the bottom tallies the results of the selections. It does not presently allow for split electoral votes. Note that the javascripting and dynamic HTML do not work on older browsers. If you experience problems with the function, I recommend upgrading to Netscape version 7.* (or other Gecko-based browsers such as Safari and Camino (MacOS X), Mozilla, Firebird, etc.) or Internet Explorer version 6 or greater.

Presidential Prediction Page

With the good help of many members of the Atlas Forum, the 2004 Presidential Prediction page is up and running. The idea here is to provide an interactive set of pages for users to post their electoral predictions, including maps and descriptive comments (necessary for the maps to hold up under the scrutiny of the fellow forum-members :) . There are two map types for the 2004 election: the Prediction Map and the Confidence Map. The Prediction Map is intended for the user to call every state the way he/she sees it and tabulate the electoral votes, thereby choosing a winner (or a tie). The Confidence Map is intended to show the relative probability of the prediction on a state-by state basis (is a particular state strong-democratic? leaning democratic? or a tossup?) Here, states that are close can be gray, showing a low confidence in the assigned winner of the Prediction Map. Your prediction is intended to be dynamic (unless you are omniscient) and can be updated as more information is learned (such as the eventual Democratic nominee). I have provided template maps that may be downloaded from the page. So, please participate and enjoy. As always, please provide feedback. There are forum topics that cover both the discussion of the predictions and any technical issues with the pages.

Final Results from the California Recall

Arnold Schwarzenegger was sworn in today as the new Governor of California. The recall election results in the Gubernatorial Section of the web site have been updated per the certified results from the California Secretary of State (the official Statement of Vote 2003 Statewide Special Election October 7, 2003). The official statement has over 1.4 million more votes than were included in the preliminary results (about 16% of the overall vote total). Interestingly, the overall percentages did not change much at all. Schwarzenegger gained about 0.8 of a percentage point overall and his decade result changed in only one county, San Bernardino, from > 50% to > 60%. Bustamante dropped about one percentage point and his decade result changed in only one county, Santa Clara, falling from > 40% to > 30%. McClintock picked up about 0.3 of a percentage point. Results for other candidates are about the same.

LS Gubernatorial Results

The preliminary (100% precincts reporting) results of the Louisiana Gubernatorial run-off election have been posted in the Gubernatorial section of the website. The results differ relative to many recent polls; Blanco wins with a margin of slightly less than 4 percentage points: 52% to 48%. The results by Parish show signficant geographic support for Blanco, winning 53 Parishes to Jindal’s 12.

MS Gubernatorial Results

The preliminary (99% precincts reporting) data from the Mississippi Gubernatorial election is in. You can find the county map and data in the Gubernatorial section of the web site. The results show a modest win for Haley Barbour (R) over incumbent Ronnie Musgrove (D) by a margin of 7.3 percentage points (52.9% to 45.6%). The three minor party candidates received a combined total of 1.6%.

KY Gubernatorial Results

The preliminary (100% precincts reporting) data from the Kentucky Gubernatorial election is in. You can find the county map and data in the Gubernatorial section of the web site. It is a solid win for Fletcher (R) who leads in the unofficial tally 55% – 45% over democrat A. B. Chandler. I have not yet found preliminary county-level results for Mississippi.


I cast my ballot this evening! Positions for Alderman and School Board were all that were on the ballot here. The city bought new optical ballot-reading machines (each paper ballot was about 8 1/2 x 15!) to replace the old lever machines that were used last year. It felt a bit like I was taking a high-school test… filling out little ovals (albeit with a pen instead of a #2 pencil). Overall, I believe the paper ballot read by optical-scanning machines to be an excellent system. Very accurate (especially relative to punch-cards), much faster to count than standard paper ballots, and a (sizable) paper receipt in the event of a need to recount or examine the election.


Tomorrow is Election Day! Although most states do not have high-level offices up for election (exceptions being the Governor’s races in Kentucky and Mississippi. Note that the Louisiana Gubernatorial election runoff is on Nov 15th), many local races such as school board, aldermen, etc. are taking place tomorrow. Such off-elections traditionally have significantly lower turnout – and consequently individual votes can have a greater impact on the results. I encourage you to learn about the local offices up for election and the candidates interested in filling those positions – and then go to the polls! You can find information on voting in your particular state (offices, candidates, polling locations, sample ballots, etc.) at the Secretary of State (or equivalent agency) website. You can find a link to these websites on the Links page.

Electoral College Calculator

I was recently asked about a web-based interactive electoral college calculator. Such an application allows the visitor to dynamically color states “blue”, “red” (or “green”) to simulate various electoral college scenarios for one’s favorite candidate. I believe that this would be an excellent feature – one that could further be used in “what-if” scenarios of past elections. Unfortunately, I currently do not know how to implement such a feature. If you have knowlege in this area, please send me some tips.

In the meantime, there does exist an Electoral College Calculator at Gray Raven Ventures . This is an embedded java application. It loaded slowly on my machine, but once there was very responsive. It allows for four candidates (and they have chosen the media’s recent “standard” of blue for Democrats and red for Republicans). The map is a bit crude, but practical. It presently requires that one choose the state first, and then click on the party that you want to award its electoral votes to. This requires signficantly more clicks and mouse-movement to complete a scenario (a faster version would allow a user to choose a party and then fill all the states for that party). It also has an advanced feature to enter split electoral votes – if, for example, you would like to award one of Maine’s CDs to a different party from the statewide winner.

Perhaps the best feature is that you can manipulate this map to run “what-if” scenarios. On the left side is the ability to display the results from election (1789 through 2000). You may then change the winner of various states to see how this would have affected the outcome!

Update:Warning, This application locked my Netscape 7 browser and required me to reboot.