Median Voter Model (MVM)
The MVM has been a standard economic and political science framework to analyze elections. The basic idea is candidates move to the middle of the spectrum (to moderate positions) to receive at least 50% +1 votes. I've occasionally used it in my own research (looking at governmental accounting issues and political competition). Of course, it simplifies decisions to a one-dimensional (left-to-right) spectrum plus adds several additional assumptions. The fundamental model was developed by statistician Harold Hotelling (1929), economist Duncan Black (1947) and economist Anthony Downs (1957).
Many empirical studies suggest it works reasonably well in various voter settings, but seems to make more sense when compromise was common and political parties were "big tent," where both Republicans and Democrats could be liberal or conservative. There used to be substantial overlap by party, less so by region. Currently, at the national level, members of either party can be modestly moderate but there is virtually no crossover. The most liberal Republican is probably more conservative than the most conservative Democrat.
Gallup claims that 35%-38% of voters describe themselves as "centrist," an alternative category to far-right or far-left. People in the middle also could be called Independents. However, this just means they are not affiliated with a party. Bernie Sanders, a liberal, is an Independent. There is a moderate voter website with the motto: "Moderate is not a 4-letter word." The site lists 20 moderate Senators (10 from each party) and 30 moderate house member (15 each). There were exactly two Congressmen from my state of Texas (Henry Cuellar and Pete Gallego, both Democrats). The site also claimed that 41% of the voters in 2012 were moderate.
What happened? The voters in the middle (the ones who are supposed to decide elections) seem to have little or no influence on elections or policy. Determining why is one reason for this blog.