Electoral Projections Using LOESS

By: TJHalva | Comments [5] | Category: Methodology | 7/31/2008 2:11:34 PM CT

I rolled out a brand new algorithm that uses local regression to generate a trendline corresponding to the states poll data; I created a PDF detailing the method. The algorithm uses a degree of 3, a bandwidth of 15, and a confidence interval of 95%. If a given state has less than 5 non-partisan polls a local least squares method is used. The results of the algorithm are presented on our graphs and in our electoral college estimates. If there are two or more polls with the same end date they are averaged and treated as a single poll.

Published on July 31st at 2:11 PM CT :: 5 Comments

Sen. Stevens Faces Criminal Charges

By: TJHalva | Comments [0] | Category: Senate | 7/29/2008 12:30:26 PM CT

Senator Ted Stevens (Republican, Alaska), the highest ranking and longest serving current Republican in the US Senate has been indicted of 7 criminal counts by the US Justice Department. Stevens tried unsuccessfully in Alaska's Republican Senate Primary of 1968 before falling short to Elmer E. Rasmuson; who later lost to Mike Gravel (Democrat, Alaska). Stevens came to power in that same year after the death of Alaska's other Senator Bob Bartlett (Democrat, Alaska). In turn Stevens was then appointed by Governor Walter Hickel (Republican, Alaska) to serve out Bertlett's remaining term. Stevens was later reelected in a 1970 Special Election.

At this point it looks as though Sen. Stevens' career will end how it started, with a Senate Election lose. It is a safe bet to assume that the 7 charges will hang over his reelection bid this November. Sen Stevens' seat was already at the top of the Democrat's list for possible takeovers, and these indictments will all but seal a Democratic victory for the challenger, Anchorage Mayor Mark Begich. The 7 charges stem from a lengthy investigation involving shady buisiness dealings in Stevens' home state of Alaska. The exact details behind the charges will be disclosed in an upcoming press conference.

Update: The press conference has occurred and nothing really happened other than some procedural stuff. Stevens is charged of making false statements in relation to accepting $250,000 from VECO Corp. the largest private employer in Alaska. The person running the press conference was asked several question to which he referred the questioner to read the report in greater detail. Well guess what, the average person wasn't handed a report from the Justice Department so I tried to find my own. I went to the Justice Department website to see if I could get my hands on a PDF, but to no avail; it will be interesting to see how long it takes them to post the report as the whole incident is a giant smudge on the GOP's image. There is also a lot of talk about whether Senator Ted Stevens will resign and the ramifications this event will have on the presidential race. I'll post a direct link to the document if it ever gets published. The Department of Justice has published a press release, but not the actual indictment documents.

Elizabeth Dole is a Senate Parasite

By: TJHalva | Comments [2] | Category: Senate | 7/6/2008 9:23:51 PM CT

With the 2008 senate elections still in their dormant summer phase I started to experiment with a ratings system. I started with the greenhorns elected for the first time in 2002; there are nine in total with seven Republicans. The basic notation I landed on, which may be too idealistic, is that the performance of a senator actually has an effect on their reelection efforts. Heading over to the Library of Congress the data gathering began, by hand. I started by defining what a good, or for that matter, bad performance entails. I initially started by analyzing each senators voting record, but ultimately decided that this course would result in the introduction of partisanship.

I needed to find a way to concretely rate the senators based on their actions, but not necessarily their beliefs. I settled on this formula: The average number of cosponsors present on a bill that is inducted into law in which a given senator is also a (co)sponsor. The basic rationale behind this formula attempts to gauge the extent to which a given senator individually participates in the law making process. The general hypothesis is that a lower average will indicate a larger contribution, and a higher number will indicate less participation. I tabulated the results for all nine first time senators up for reelection in 2008 along with some other, more prominent names for use as a measuring stick. All data was gathered from the 108th, 109th, and 110th congresses.

Senator State Party Elected Avg # of Sponsors
Ted Kennedy Massachusets Democrat 1962 23.397
John Cornyn Texas Republican 2002 31.138
Chuck Hagel Neberaska Republican 1996 33.302
John E. Sununu New Hampshire Republican 2002 33.688
Larry Craig Idaho Republican 1990 34.852
Saxby Chambliss Georgia Republican 2002 35.172
Frank Lautenberg New Jersey Democrat 2002 37.194
Lamar Alexander Tennessee Republican 2002 38.727
Norm Coleman Minnesota Republican 2002 40.459
Mark Pryor Arkansas Democrat 2002 44.357
Lindsey Graham South Carolina Republican 2002 45.323
Elizabeth Dole North Carolina Republican 2002 49.750

As you can seen from the table Ted Kennedy remains a senate legend. Ted Kennedy scores very low for two reasons; first he is a seven term, second in seniority senator and he solo authors many bills that eventually become law. On the opposite end of the spectrum is Elizabeth Dole, who appears to be useless. On a further analysis of her record, the only bill to become law that she has personally written focuses on the redrawing of coral reef on maps in the Atlantic Ocean. Is that seriously the most important issue right now; the answer's no. Back on track, based on this small sample set it appears that a lower average is in fact superior, but it remains unclear to what extent this holds true and if the average has any correlation to a senator's reelection bid. I should also point out that Frank Lautenberg served as a senator in the 1990s before retiring.

My results cannot be ruled conclusive, although they are interesting. However, the aftermath leads me to believe a more thorough analysis of this application would yield a fascinating outcome, but the data is sparse. There is no way to automate the calculation process. The Library of Congress has a thorough database of legislative data, but it unfortunately is not programmatically available. I may get ambitious in the future and manually start developing an open database from the Library of Congress' site.

Published on July 6th at 9:23 PM CT :: 2 Comments

Electoral College Projection Map
Senate Projection Map