On my political I had calculated some statistics about U.S. presidential elections, specifically about how the Electoral College affected the outcome in terms of the two modern major parties.
I thought it would be interesting to do a quick calculation about time-in-office and number of terms, starting with 1860 (the first year when it was a real Republican-Democrat election).
Since then, counting the current term, 15 out of 38 terms (39%) have been Democrat administrations, and 23 out of 38 terms (61%) have been Republican. Not counting the current term, the average time for a Republican in office is 4.84 years, vs 6.98 years for Democrats. Yes, FDR throws it off; but to come down to the GOP average, you have to take out FDR, Truman, and all 3 two-termer Democrats (Clinton, Wilson and Grover "Mr. Non-Contiguous" Cleveland). To bring the GOP's average up to the Democrats, you have to throw out Garfield, Harding, Ford, Arthur, Johnson, Hayes, Harrison, Taft, Hoover, Bush I, Lincoln, and McKinley.) Looking at the list, it struck me that the iconic JFK is the shortest-term Democrat since the Civil War, and the second shortest of either party.
From the start of Lincoln's term until today, the GOP has had 1104 months as presidient, vs. 821 months for Democrats. If you look at contiguous administrations (whether or not it was the same individual running them), Republicans have an average streak of 2.875, and Democrats 2. Take away the post-Civil War era from the GOP and it falls to 2.43; take away FDR and Truman and the Democrats fall to 1.5. One interesting idea would be to look at the same data over the same period for Congressional Districts. Besides showing the trend of Democrats getting elected in areas that vote Republican for president, there's more data and therefore less noise.
The interesting thing about looking at the data over time is that it appears to converge. Streak-length appears to moderate too, at the same point, right after Truman (administration names provided below for reference because the X-axis is totally irregular with respect to time). What's interesting about this is that the "brands" the two represent, i.e. the demographics they capture based on their message and political climate of the time, have changed radically over a century and a half. Will these graphs still look this way after another century?