Primary fortunes and election tales

Start with the obvious: Hillary Clinton is very unlikely to catch up to Barack Obama in the number of pledged delegates won, and Barack Obama is very unlikely to win enough pledged delegates to capture the nomination.

As a result, the superdelegates are going to decide the Democratic presidential nomination. Whether you celebrate or bemoan this fact, it appears inescapable.

Democrats have reason to hope this reality will constrain the behavior of both candidates. The need to attack may have been both campaigns’ take-away from Ohio and Texas, but nothing is more likely to alienate superdelegates than a scorched-earth negative strategy against a fellow Democrat.

Efforts to attract superdelegate support are already in overdrive, with both camps spinning webs of affirmative arguments for their candidacies. Let’s dispense with their normative contentions for the moment — no one has ever hired me as a philosopher — and focus on the tenuous cases both are making about electability based on their performance in the primaries and caucuses.

Make no mistake, absent deep divisions rending the party I am confident either candidate can win in November. One may be more likely to win than the other, but performance in primaries gives us precious little insight into determining which one that is.

Sen. Obama argues he has won a whole series of red states, enabling him to redraw the electoral map, as well as college-educated white men, a key segment for Democrats. Sen. Clinton maintains she has won the large base and swing states Democrats need to win the Electoral College, as well as non-college-educated white men and downscale white women, two blocs Democrats also need in November.

Both seek to project their primary performance into the general election. History and logic suggest that, while tempting, that analytic strategy is fundamentally flawed.

Winning states in primaries and caucuses has little to do with winning them in general elections.

John Kerry began his march to the nomination with a come-from-behind victory in Iowa, only to lose the state to George Bush. In 2000, the loser of New Hampshire’s Republican primary (Bush) beat the winner of the Democratic primary (Gore) in that state’s general election.

Bill Clinton lost 1992 primaries in New Hampshire, Colorado and Maryland before winning those states in the general election, while primary victories in Florida, South Carolina, Mississippi and Oklahoma failed to foretell general election wins.

In 1988, Mike Dukakis won general election victories in only eight of the 25 states in which he won primary or caucus battles.

Using primary results within voter subgroups to project general election outcomes is equally problematic. The most important political fact about anyone who votes in a primary is not his or her level of education, income or even party registration. Rather, the key point of differentiation is the very fact he or she chose to vote in a partisan primary. Non-college white men who voted in the Democratic primary are importantly different from non-college white men who did not. Assuming that those who did not vote in the primary will behave like those who did ignores their most salient difference.

In 1992, Bill Clinton lost college-educated voters in New Hampshire overwhelmingly to Paul Tsongas. In the general election, Clinton won a higher percentage of the two-party vote among college-educated voters than has any Democrat before or since.

This is one of those occasions where numbers create only the illusion of precision. Because Democrats desperately want to know who the strongest candidate will be, our desire to read tea leaves is insatiable — and so we read too much into each leaf we happen upon. Sometimes, though, tea leaves are just tea leaves and nothing more.

Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982. Current clients include the majority leaders of both the House and Senate.