Symbolitics As Usual
A guide to non-instant election analysis.
Jan 21, 2008, Vol. 13, No. 18 • By JOHN J. DILULIO JR.
Recall the image makeover Bill Clinton's campaign undertook in 1992, as revealed in an internal document uncovered some months later by Michael Kelly and reported in the New York Times shortly after the election, on November 14. After five months of adverse publicity and sagging poll numbers, the Clintons' advisers "proposed the construction of a new image for Mr. and Mrs. Clinton: an honest, plain-folks idealist and his loving wife." As Kelly documented, this symbolitics strategy "required a campaign of behavior modification and media manipulation so elaborate that its outline ran to 14 single-spaced pages." It explicitly called for depicting Bill Clinton as an "agent of change," plus holding town-hall style forums and events "with her friends where Hillary can laugh, cry, do her mimicry." This year it took the Clinton operation only five days to react to bad news.
Whoever the candidates are this fall, the general election campaign will see unprecedented spending on television ads. Ted Brader, a young political scientist at the University of Michigan, has analyzed thousands of ads produced by presidential and congressional campaigns. In Campaigning for Hearts and Minds (2006), he details how most ads appeal either to voters' fears or to their enthusiasms. Among his many fascinating findings, it turns out that emotion-tugging ads are more, not less, effective among voters who are fairly well informed about and attentive to particular policy differences between the candidates.
In the current primary campaign, Obama is like a nonstop enthusiasm ad. Hillary hit back in New Hampshire with some teary enthusiasm of her own (sincere, I think), even as she pushed the symbolitics fear button via references that seemed calculated to get voters worrying over whether Obama will combat terrorists. Obama is gifted at speaking "poetry" on the stump, but he has probably seen the last of an uptight Hillary mumbling wonkish "prose" in reply. Indeed, he is almost certainly in for a symbolitics counteroffensive the likes of which we have never seen.
But symbolitics isn't everything. Even given all the great firsts that Obama's electoral success would represent, if he beats Hillary for the Democratic nomination, he will still need to flesh out his intelligently left-leaning (as opposed to boilerplate liberal) positions. Obama will need to rebut, say, a conservative John McCain's policies on Iraq, border security, school reform, tax cuts, abortion rights, and other matters. Hand-to-hand combat on policy details may or may not turn out to be another Obama strength.
This much, however, is certain: Spirited sermons echoing Obama's book The Audacity of Hope (2006) will soon wear thin, and above-the-fray appeals to bipartisanship will ring ever more hollow. At present, compared with every other major hopeful in both parties, Obama does seem short on policy-relevant, change-making experience. The analogy with the young John F. Kennedy will carry him only so far (see PT 109), nor can he coast indefinitely on narrative notes from the Harvard Law Review, his years as a community organizer and state legislator, his support for congressional ethics legislation, or even his opposition to going to war in Iraq.
Next, to free oneself from excessive deference to opinion surveys, it is also necessary to really and truly understand that no poll is perfect, and that averaging results from polls taken by different organizations, at different times, in different ways, with different samples and sample sizes is no sounder than relying on what the few biggest, best, and latest polls record.
Consider a poll's margin of error. Keep in mind that "plus" and "minus" count, as it were, on each side. To wit: If a legitimate poll--one with an adequate sample size, non-biased wording, and so on--finds that Obama leads Clinton by 8 points, with a margin of error of plus or minus 4 points, it has ascertained something significant but far from oracular: to wit, that on the date the poll was taken, Obama would have beaten Clinton by 16 points (plus 4 for him, minus 4 for her) or tied her (minus 4 for him, plus 4 for her) or reached a result anywhere in between.
Polling, including exit polling, has improved, but sampling is still an art, not a science. It may happen only rarely, but Election Day factors like door-to-door canvassing and other turnout tactics can beget actual returns that deviate from forecasts based on even the finest "likely voter" survey.