A typical voter in the United States has become ever more difficult to define in the 21st century. Only a split of the total number of people in a typical election will indeed vote. This group often has diverse characteristics than the whole population. The results of a typical election are unlike what they would be if every citizen actually turned out and voted, because the people who actually vote are different from the people who do not. More characteristically, yet, the group of those who vote is not representative of the total population. Records suggest that actual voters can differ in significant ways from all qualified voters. I believe these differences are the result of two main factors.
First, there are demographic characteristics that have by tradition been associated with the leaning to vote, and of a vote. Everything else being equal, certain types of people will more likely vote than others. The most important determinant of these is age and socioeconomic status. Older and better educated individuals are, universally, the more likely to vote than younger and less educated individuals.
For instance, because Republican voters tend to be higher educated than Democratic voters, the actual voter pool on an Election Day can be skewed more toward Republicans than the entire population composition would suggest. This in turn explains why, in some elections, the actual vote on Election Day can be more Republican than pre-election polls forecasted.
The second block of factors is related to voter turnout, and is distinctive, and can reflect one of a hundred characteristics that come into play in any given election. In some elections, particular factors may activate or animate certain groups of voters and cause them to vote at higher extents than their representation in the general pool of registered voters might have predicted. For example, an election that has racial overtones might activate minorities more than usual. An election that has fiercely debated labor-related issues might activate union members.
There are some strategic statistical methods that can be used to expand on our accuracy in determining typical voters, or comparably so, likely voters. Listed, we can develop case sensitive questions in our polls that reflect the most important issues in an election, we can attempt to isolate particular demographic differences that in themselves support or object to such considerations as socioeconomics or race and/or gender, age versus political interest, regionalization vs. federation, researching current and past voting behavior more accurately, etc.
Identifying a typical voter matters, even though the extent by which it matters is relative to its importance to the entity requiring its effectiveness. For political parties, yes, and for candidates, yes. But for estimating the trend of politics, maybe not, because the determinant of a likely or typical voter remains inconsistent depended on the election being sampled.