William G. Mayer (2007) The Swing Voter in American Presidential Elections

William G. Mayer (2007) The Swing Voter in American Presidential Elections, American Politics Research, May 2007 vol. 35 no. 3 358-388

En av mycket få artiklar på swing voter-temat. Går igenom möjliga defintioner och förklarar vad som skiljer begreppet "swing voters" från t.ex. indepedents, undecideds, party switchers etc. Enkel artikel men mycket bra! Tyvärr är måttet med termometern han använder alldeles för USA-specifikt - tror inte det finns i några andra valundersökningar. Därtill förutsätter det i princip att det bara finns två kandidater.

Definition

s. 359: In simple terms, a swing voter is, as the name implies, a voter who could go either way, a voter who is not so solidly committed to one candidate or the other as to make all efforts at persuasion futile.

Med andra ord, man kan säga att swing voters i betydelsen “avgörande grupper” är de som är obestämda men som garanterat kommer att dyka upp...

s. 360: To see why this is the case, consider the situation of a voter located at –100 or –80 (i.e., at the far Democratic end of the scale). The Democrats will probably expend some effort to make sure that this voter will actually show up at the polls on election day. But as a subject for persuasive actions or communications, this voter is not a very attractive target for either party, simply because there is so little likelihood of changing his or her voting decision.

s. 361: Voters receive attention from campaigns according to the expected “pay-off” they will yield, meaning the number of votes that can be gained or at least not lost to the other side. Thus, campaigns will generally ignore or take for granted each candidate’s most committed supporters and concentrate their persuasive efforts on the undecided or weakly committed swing voters. In this respect, there is an obvious parallel between swing voters and the so-called battleground states in the electoral college.

s. 364-365: One advantage of using a scale of this sort is that it provides a nuanced, graduated measure of a voter’s convertability or “swingness.” Yet for the analysis that follows, it will be helpful to have a simple, dichotomous variable that divides voters into two categories: swing voters and non–swing voters. A close inspection of Table 1 suggests that the best way to define such a variable is to classify any voter with a score between –15 and +15 inclusive as a swing voter, with everyone else falling into the non–swing [/]voter category.6

Kan man få syn på en sådan osäkerhet i de svenska valundersökningarna?

Vi gör intervjuer innan (en del i alla fall). Men har vi frågor om hur väljarna väger mellan olika partier/block? Vi har frågor om när de bestämde sig, men har vi frågor om considerations? Mayers analyser är vettiga på många sätt, men de går inte att genomföra i SNES, för vi saknar en fråga där folk får skatta hur mycket de står och väger. Möjligen skulle man kunna använda frågorna om hur mycket folk gillar eller ogillar allianspartierna på detta sätt, och om det där sedan finns swing voters. Detta skulle kanske vara ett bättre mått än folk som bestämmer sig sent. Vi vet inte inom vilken range som dessa människor bestämmer sig.

Tre alternativ och varför de är sämre:

Political independents:

s. 366: But political independence, whatever its other uses, is not a very good measure of what it means to be a swing voter. If the point of the swing voter concept is to identify those voters who might conceivably vote for either major-party candidate, political independents fall short in several different ways.

Tre skäl: Det finns “hidden partisans”, som är svåra att rensa ut. Och alla som anger party identification röstar inte på ”sitt” parti, av specifika politiska skäl (exv. Reagan democrats).

Party switchers:

s. 368: But party switchers are simply not the same thing as swing voters. There are too many people who fit into one category and not the other, or vice versa. Most obviously, because party switchers are defined by a disjunction in voting behavior across two successive elections, using this variable as a way of identifying swing voters automatically excludes all those who did not or could not vote the last time around.

s. 368-369: If not all swing voters are party switchers, the reverse is also true: Not all party switchers are swing voters. Party switchers include all those who decided to abandon the party they voted for in the last presidential election, no matter when they reached that decision. And although some voters will not make that decision until the final days of an election campaign, many, it [/] appears, decide months or even years earlier and are thus effectively removed from the swing voter category by the time the campaign begins.

The undecided:

s. 369: The principal difference, at the theoretical level, is that the swing vote is a slightly broader concept: It includes not only those who are literally undecided but also those who have some current vote intention but are weakly committed to that choice.

s. 371: There is also some reason to think that many of those who say they are undecided may actually have a preference that they are reluctant to reveal to the interviewer.

Detta innebär ändå en del intressanta saker när det gäller att konceptualisera vad manmenar med swing voters, och vad jag menar med tesen att medelklassen har särskild vikt.

Kritik av Kelley: 

s. 374: Though at one point I considered using Kelley’s method as the basis for my own investigation, I ultimately came to believe that it had two major shortcomings. First and most important, the Kelley index, in my opinion, actually measures two things: a respondent’s comparative assessment of the major candidates and parties but also, to some extent, his or her level of political sophistication.

Kelley använder alltså ett större set av frågor i sitt mått (40 stycken, om jag inte tar fel). De kan helt enkelt vara så att det också spelar roll i vilken mån respondenten  kan/orkar besvara frågorna.

Spelar swing voters roll för vem som vinner eller inte?

s. 376: The base vote, as I am using the term here, is the opposite of the swing vote: It is those voters whose support a candidate can comfortably rely on. On average, the 18 major-party candidates shown in Table 7 held on to 96% of their base vote. The problem for most campaigns is that the base vote falls short of a majority. Hence, the principal goal of the campaign becomes to add on to the base vote enough weakly committed,  undecided, and even initially antagonistic voters to secure a majority. And that, of course, is where the swing vote becomes important.

The swing vote is most significant, then, in close elections. The basic dynamic can be seen most readily in the elections of 1976, 1980, 1992, and 2000. In each of these contests, both major-party candidates had a base vote of between 30% and 40% of the electorate. When this is the case, which candidate wins will depend on how the swing vote breaks—and in every one of these elections, the candidate who won a majority of the swing vote also won a majority of the popular vote as a whole (though in Gore’s case, this was not enough to carry him into the White House). The situation is different when the general election shapes up to be a landslide. In 1972, for example, 53% of the voters in the NES preelection survey were already part of the Republican base vote. To win the 1972 election, George McGovern had to win an overwhelming percentage of the swing voters and make some substantial inroads into the Republican base. In fact, as the figures in Table 7 indicate, Nixon held on to 94% of his base vote—and also won a narrow majority of the swing voters. Ronald Reagan in 1984 and Bill Clinton in 1996 similarly began the general election campaign with a base vote that fell just shy of a majority.

Tillhör swing-voters vissa speciella demografiska grupper? Klass är inte med!

s. 377: The swing vote, in sum, is not the be-all and end-all of American presidential elections. It is much less important in landslide elections—but, then, so are campaigns in general.

s. 382: The most important conclusion to be derived from Table 10 is that swing voters are, at least in demographic terms, a very diverse group. Of the 87 survey groups evaluated in Table 10, in only 16 cases is the group significantly over- or underrepresented among swing voters—and in only 4 cases does the difference reach 10 percentage points. To the extent that swing voters are demographically different from non–swing voters, moreover, their distinctive attributes vary from election to election. The only group that is overrepresented among swing voters in at least 8 of 9 elections is Catholics.

Är swing-voters mindre politiskt intresserade / kunniga än andra väljare?

s. 380: “Compared to voters generally,” Kelley concluded, marginal voters “were on average less well educated, less active politically, less interested in the campaign, less informed, and less attentive to politics” (p. 157). Given what has been said earlier, however, about the problematic character of Kelley’s method of identifying marginal voters—particularly the fact that it may also serve as a measure of political sophistication—the whole matter is clearly worth revisiting. Table 9 accordingly shows how swing voters and non–swing voters compare,

on average, on a variety of measures of political interest, involvement, and information. As it turns out (the data are not reported here), using the thermometer ratings rather than the likes–dislikes questions does make some difference. Swing voters, as I have defined them, are, for a substantial majority of the questions in Table 9, more involved and more knowledgeable than a comparable group based on the Kelley (1983) index (specifically, those with scores between –2 and +2, inclusive). But the differences are in most cases rather modest and not enough to undermine Kelley’s basic conclusion. Swing voters, no matter how one defines them, are consistently less involved in and  informed about politics than the rest of the electorate.