In a "
winner take all", non-parliamentary, generally two-party
republic like
the US, there are a fair amount of people who, one can be assured, will vote for one particular candidate, or one party, come what may. Estimates of the proportion of this electorate that will behave in this manner typically range from around 60 to 80 percent, generally split equally between dedicated
conservatives and dedicated
liberals. The remaining voters, who may "swing" towards either (or any) option, are known as "swing voters" - their positions on issues of the day are relatively less set, less emphatically held, or some combination of the two, they lack strong allegiance to any one party or issue-oriented interest group, they are generally uninformed about the issues and themes of the race, or they find them irrelevant or inapplicable to their lives. Swing voters generally choose between candidates on the basis of the perceived merits of that individual candidate's particular platform or agenda, typically placing more emphasis on issues that directly affect the voter - economic programs, workplace regulations, tax plans, and at the more local levels, infrastructure and planning agendas - than "abstract", ideological debates.
Of course, in a two-party system like the US, where "conservative" and "liberal" are each represented by one party, if 30% to 40% of the electorate will, without fail, throw their support to one side, then it's the remaining 20% or 40% that decides the election, and both parties dedicate most of their efforts to courting these swing voters. In a binary system where issues are typically expressed in terms of binary choices for which either party can be assumed to hold the position opposite that of the opponent party, swing voters are by definition in between, and so both parties often seek to put forth a "moderate" face and program which will alienate as few swing voters as possible. This often requires candidates, campaigning in a general election, to reverse to some extent the positions they espoused during the lead-up to the primary election. Even if only 30% of the population is overall liberal or conservative, those 30% of the general populace will represent a 60% majority of their representative parties (making for the sake of simplicity the false assumption that the entire population is split 50/50 between the two major parties), and in any case tend to represent the more politically active, and more free-spending members of the party. Thus, in order to achieve nomination, primary candidates often appeal to more "hard-line" interests within the party, later tempering these positions in the search for swing voters in the general election. Of course, it's not sufficient not to repulse swing voters, parties must seek to actively woo them to vote for their preferred candidates. In each election cycle, great effort, thought, and polling firm billable hours go into finding what issues swing voters are most likely to identify with, with the result often being a composite archetype of the swing voter populace, like 1992's "angry white male", or 1996's "soccer mom". Parties and individual candidates will then play up their respective strengths on these issues, attempt to highlight their opponents' perceived weaknesses, and implement relevant new legislative or public relations initiatives to win swing voter approval. If you play your cards right, you find yourself gaining (or retaining) a spiffy title and a staff to take your constituent correspondence at public expense. If not, well, there's always next election, right?
Of course, some say that these days, the search for "middle ground" in appealing to swing voters makes a mockery of parties' claims to stand for something, anything at all, and renders them less and less distinguishable from each other. There's more than a little "golden days" false nostalgia in remarks of this nature, and at base, this is essentially a complaint that parties must appeal to a majority of constituents in order to claim power, which is, after all, rather the point. There is some truth to them, however - while the American government or election process hasn't changed significantly in recent memory, even dispassionate analysts will acknowledge some signs of a trend away from ideological polarization and an increasing fetishization of the "swing voter" in American politics of late. On a personal level, hopeless little shits that they were, I'd love another Barry Goldwater, or even Adlai Stevenson, to root for. There are a few things that might explain this ascendency of the swing voter.
One possible argument is that we've simply exhausted all matters of importance. With issues like civil rights (for) and the dominance of the national or state governments (national) behind us, we've pretty much done everything that matters, and from here on it's all procedural and unimportant issues. More or less, this is the national analogue of Francis Fukuyama's global "end of history" argument. Of course, I think both are wrong on their faces, but it's a possibility to be considered. Next, there's the possibility that this is just a historical aberration - with the end of the cold war and the decline of the welfare state paradigm under Bush and Reagan, respectively, the parties are still in a process of reorienting themselves, and the "culture wars" of the late '80s and mid '90s left the electorate tired and disenchanted with ideological politics. There may be important things still to deal with, but there are none to deal with now, and we're back in another Gilded Age of political irrelevance (though hopefully without the rampant corruption). Another theory would be that no one actually does care anymore - with the fall of nationalism as a power to be reckoned with and the increase of world interdependence and interconnection, identification with an entire nation (remember, in the US, people once identified with state citizenship) is replaced with affinity and social affiliations - government is an increasingly irrelevant means, and never an end, and is being treated as such. On a less abstracted level, the whole thing might just as easily be attributed to new nationwide instant polling techniques and capabilities - after all, you couldn't court the swing voters before if you didn't know what they thought. Obviously, behind it all is the decline of party allegiance - straight ticket voting is down, independent registration is up, and general ambivalence about the parties is at record highs and on the rise. But that's really a matter for a node of its own.
In any case, it's a little silly to assume that the importance of the swing voter will lead to an inevitable and complete merging of American politics to a murky center. Parties who moderate their platforms too much risk losing core constituency votes to third parties - it's widely speculated that Ralph Nader cost the Democrats the 2000 presidential election under exactly these circumstances. Even parties outside of the traditional continuum can stand to benefit from disaffected swing voters - Ross Perot and his Reform Party took around 17% of the vote in the 1992 presidential race, which if a bit higher and/or differently geographically distributed might have allowed him to play a kingmaker role in the Electoral College in return for significant agenda concessions. And after all, it's possible that the importance of swing voters as a whole is being overstated - an accurate survey of all citizens would yield a "swing voter" yield far greater than the stated 20% to 40%, the important caveat being that undecided citizens are, almost by definition, least likely to vote, and tend to stay away from the polls en masse. From a game theory perspective, Timothy Fedderson noted in 1995 that the higher the stakes are, and the more chance they have of making a significant impact, the less likely the least informed and least prejudiced (in the sense that they have not made a decision beforehand) voters are to vote, and this appears to be a rational, optimal outcome.