Analysis Rationale (crosspost from Yahoo-360)

Electoral and Election analysis 1904-2004 and 1948-2004

The presidential, and therefore, the electoral race 2008:

After hearing so much opinion and propaganda being thrown around, I decided to turn to statistics in order to get a more cut-and-dryed approach to the whole electoral trend phenomenon. Using Dave Leip’s election analysis website and a number of others, I have begun a large research project into electoral politics of the last 100 years. First, I made comparisons, per state over a 100 year period, and then more importantly, over a 60 year period. The numbers speak for themselves.

NOTE: For simplicity’s sake, the word “state” will be used for states and well as commonwealths and the District of Columbia.

Each state is presented in the same way and has been ranked in the listing solely based on it’s partisan performance in the presidential election of 2004. The calculation is simple: one takes the winning margin in each state from the presidential race and subtracts the national average (margin) for the respective party:

Sample:

Indiana: George W. Bush, Jr., won Indiana with a margin of 20.68%
The national average for the GOP in 2004 was +2.46%

20.68
-2.46
-------
18.22%, making Indiana the 12th most conservative state, meaning that 11 other states were won for the GOP with wider margins than Indiana.

Sample 2:

New York: John Kerry won New York with a margin of 18.29%
The national average (margin) for the DEMS is –2.46 (so a minus, minus = plus)

18.29
+2.46
----------
20.75%, making New York the 5th most liberal state, meaning that 3 other states + DC were won for the DEMS with wider margins than New York

After the partisan ranking, the number of Electors are given and a prediction for it’s number of electors as of the 2010 reapportionment (based on Census 2006 data and trends) and that state’s electoral development. According to census 2006 predictions, assuming trends continue as they have until the year 2010, the chances are good that 7 states will lose at least one elector a piece: Iowa, Louisiana, Massachusetts, Missouri, New York, Ohio and Pennsylvania, and that six states will gain at least one seat: Arizona, Florida, Georgia, Nevada, Texas and Utah. Texas is likely to gain between 2-4 electors, New York and Ohio will lose between 1-2 electors a piece, and for the first time, California may gain no electors.

Next, I simply counted the number of time the state went for each party from 1904-2004 (26 elections) and then again from 1948 (15 elections). In this way, we can see when the massive shifting of the electoral landscape started to take place in the last century.

Example, TEXAS:

“From 1904-2004, Texas went for the GOP 11 times, for the DEMS, 15 times
From 1948-2004, Texas went for the GOP 10 times, for the DEMS, 5 times”

We can easily see that the GOP won 10 of those 11 times after 1948 showing us that the ideological shift happened in the second half of the 20th century.

If there is relevant information about that particular state, I may quote out of Wikipedia or another source at this point.

Now the fun begins:

I do a calculation of the voting in that state, but only as of 1948. I do two (2) calculations:

First, I simply average the percentages for each party and present them. In the case of the independents, I average again, this time dividing the total percent by the actual number of elections in which the independent were on the ballot between 1948-2004. This average gives a feel for how this state would most likely go for independents were an established independent party on the ballot in every election.

Second, I add up all the real votes for each party, divide by total votes cast and come up with a real vote percentage. The more reliable figure, of course, is the averaging of the percentages but the real vote count gives us a feel for how many voted in this time and whether the state voted consistently or if it has varied greatly.

If a state has been very consistent in voting, then the percentage based on the real vote will come very close to the average of the winning percentages. If there is a difference of 3% or more between these two methods, then it indicates that either one party or the other won huge landslides early on, when there were far less voters than in the record breaking election of 2004, or that an independent party (for instance in MS, AL, GA, etc) actually won the state in one or two massive landslides. Such things can skew the figures. Here a humorous, fictive example:

For example the state of MARSHMALLOW

Candidate A, from the “smores” party in 1948: 700,000 out of 1,000,000 or 70%
Candidate B, from the “roasties” party in 1948: 300,000 out of 1,000,000 or 30%

A LANDSLIDE for the smores.

Now, same parties, year 2004 and a reverse landslide with a little more than 4 times as many votes:


Candidate A, from the “smores” party in 2004: 1,260,000 from 4,200,000 or 30%
Candidate B, from the “roasties” party in 2004: 2,940,000 from 4,200,000 or 70%

A LANDSLIDE for the roasties.

Now, the average of both elections:

“Smores”: 1,960,000 from 5,200,000 (37.70%)
“Roasties”: 3,240,000 from 5,200,000 (62.30%)

however the averaging of the percentages would have been a tie: 50%-50%

Now, assume 8 or 9 elections in between and you can imagine how a slight skew can happen.

Why did I pick 1948 and not 1964 or 1932? Why not the entire 20th century?

A number of reasons:

1.) 1948 was the first election post WW-II and the first election in the atomic age. The old accepted GOP and DEM alliances before the great depression were all pretty much muddied up through 4 Roosevelt landslides and quite clearly the shaking out and reshaping process happened in 1948.

2.) 1948 was the first election in the age of TV. True, most TV sets were sold after 1948 and the first television election spots first happened in 1952, but indeed television was indeed around in 1948. And the nominating conventions were televised for the first time in 1948.

3.) 1948 marked the return of the “whistle-stop” tour, an election technique that has been used ever since then.

4.) Since 1948, the shift of ideology in the NE of the USA from conservative to liberal and in the deep south from liberal to conservative has been in motion. True, the radical knee-jerk shift in the south happened in 1964-1968 as an almost violent reaction to the civil rights act, but Strom Thurmond (States Rights) already sent a huge warning shot across the DEM’s bow in 1948. And the real shift to the left in the NE began about 12-16 years after the red-shift began.

5.) Furthermore, it is so that three generations vote in an election. This means that seniors (older than 70) of today were children or teenagers in 1948 and were influenced by the events of the day, but not of events from much earlier. For this purpose, a 60 year time span is not a bad idea, which is what this time-span will be when the 2008 election happens.

6.) Lastly, this time frame allows us to see and measure the influence of three of the four major independent threats to the two-party lock on the electoral system (excluding 1912): 1948, 1968 and 1992. It is also very telling to note that in spite of the fact that Ross Perot had the highest percentage of the independent vote out of these three elections (18.91%), he won no states and therefore 0 electoral votes, whereas Strom Thurmond, with just 2.37% of the national popular vote in 1948 won 4 states and 49 electors and George Wallace in 1968, with a much higher percentage of the independent vote (13.53), won exactly those same states + one, making a disproportional low gain of 7 electors (46 total), showing that it already was difficult and has become even more difficult for an independent to win the presidency. Later, I will do an in-depth analysis of the INDEPENDENT movements, including the 1912 phenomenon.

For any republican out there who might like to criticise this time frame, he or she may want to consider that out of the 7 landslide elections in the popular vote (Eisenhower 52 and 56, Johnson 64, Nixon 72, Reagan 80 and 84, Bush 88), only one of them was for a democrat and the other six were for a republican. In the electoral college, one can calculate the number of landslides at 9, as the number of electors who cast their votes for Bill Clinton in 1996 and even more so in 1992 were disproportionately higher than the popular vote margin. In this case, it would be 3 DEM electoral landslides and 6 GOP electoral landslides. So, no matter how you look at it, starting with 1948 does not stack the deck in the DEMS favor.

After the averages, some superlatives are noted: the largest GOP and DEM victories in each state, the largest IND showing (or in some rare cases, victory), and the closest margins since 1904. Some of these fascinating tidbits of information may really surprise you.

After that, a short summary of the state’s population and population density.

Finally, a short description of the partisan trend in the governorship, the Senate, the House and also the statehouses, plus a facit conclusion.


If you read the analysis of each state carefully, you will notice that partisan trend from 2004 more or less is in agreement with the percentage analysis, perhaps off one number up or down in the ranking. And in both parties, one or more states currently deemed as solid to leaning are actually trending toward the other party and the percentage analysis correlates to this trend.