Saturday, December 13, 2008

Opinions with results

On August 11, 2008, writing for the ironically named conservative publication The American Thinker, Steven M. Warshawsky proclaimed, "As I wrote last December, '[t]he pundits can talk until they are blue in the face about Obama's charisma and eloquence and cross-racial appeal. The fact of the matter is that Obama has no chance of being elected president in 2008.' I am more convinced of this conclusion than ever."

On October 9, shortly after the second presidential debate, Warshawsky wrote, "I have received numerous emails from Republicans and Democrats alike, asking whether I still think Obama will lose the election. Yes, I do. But what about the polls, they ask? The polls show that Obama is winning. No, they don't, as I will explain." And he did.

Even by October 25, he did not back down: "In a few more weeks, the political environment in this country is likely to become a heckuva lot nastier. For there are real signs pointing to a McCain victory this year, whether or not the mainstream media wants to acknowledge them."

Needless to say, after Election Day he was shocked: "I cannot understand how a man like Obama became president. It contradicts everything I know, or thought I knew, about American history, culture, and politics." But he didn't conclude that his own thinking was at fault. He acted as if the country had pulled a fast one on him. He even made a whole new set of predictions about the damage that Obama and the Democratic Congress would do, and he ended by saying, "I hope that the Democrats will prove me wrong again." Judging from his track record, his hopes will likely be fulfilled!

He is hardly the only commentator to have underestimated Obama. But unlike any mainstream pundit I'm aware of, he continued to predict Obama's demise after primary season, and with firm conviction ("no chance of being elected"). To suggest as late as August that Obama might lose was reasonable, but to suggest that he definitely would lose was insane.

The sheer insularity is striking. Not surprisingly, The American Thinker has also been a repository for wild conspiracy theories about Obama, from the birth certificate business to the claim that William Ayers ghostwrote Dreams from My Father. I have no doubt that the people who believe these things will continue to believe them in the years to come. That's the beauty of having opinions: nothing can shake your belief in them, as long as you choose to consider them true. Making a specific, concrete prediction about the near future is another matter. Once you expose yourself to objective reality, you can't hide from it.

The fallacies in Warshawsky's analysis weren't hard to spot. His most telling statement was, "Why am I so confident that John McCain is going to win the election? In short, because Barack Obama is not an acceptable choice to lead the country." It didn't seem to occur to him that the American public might not share his standards of what is acceptable.

His refusal to believe the polls was also notable. The success of polls at predicting presidential winners has increased dramatically since the days of "Dewey Defeats Truman." In a very close race, there may be uncertainty. But by mid-October this year, McCain was consistently trailing Obama by at least five points, and the electoral map looked even worse for him. It was conceivable that public opinion might change before Election Day, but there was no reason to believe he was already in the lead.

Warshawsky may be an extreme case, but the punditocracy is littered with erroneous forecasts, and the pundits are rarely taken to task for them. I believe that the quality of prognostications can be a gauge of a commentator's analytical skill. We should pay more attention to them.

Let's consider some of the factors that inspire bad political predictions. One is wishful thinking. Another is projecting one's own outlook on the public. A notorious example of the latter was the title of conservative columnist Shelby Steele's early-2008 book A Bound Man: Why We are Excited about Obama and Why He Can't Win. Steele later apologized for the "stupid, silly subtitle that was slapped on to the book" and claimed it did not represent what the book was arguing. Not having read the book, I'll take his word for it.

Some people will predict a candidate's victory in the hopes of creating a self-fulfilling prophecy. That's why candidates themselves rarely admit they're losing even when it's obvious they are. The final nail in the coffin of Fred Thompson's candidacy may have been when he admitted to the press that he wasn't likely to get the nomination. For a politician in those circumstances to lie is understandable, but we expect more honesty from commentators, who should always tell at least what they believe to be the truth.

None of the aforementioned factors explain why many Democrats doubted Obama would win, even days before the election when he seemed practically unbeatable. The crucial factor here was paranoia, inspired by past defeats. They felt their own party had a knack for "snatching defeat from the jaws of victory," and they believed that Republicans might steal the election again. (To this day, there are liberals who think even the 2004 election was stolen.) They also worried about the Bradley Effect, the alleged phenomenon that public opinion polls overestimate a black candidate's support because some respondents are afraid of revealing racist motivations.

One Democrat whose predictions were spectacularly vindicated was Nate Silver of A baseball statistician by trade, Silver developed a unique method of averaging together the presidential polls to determine the winner. His final estimate had Obama beating McCain by 52.3% to 46.2%. The initial results on Election Night were 52.4% to 46.3%--within a tenth of a percent of Silver's predictions. (Subsequent counting, however, has widened Obama's lead by a whole percentage point, making Silver's estimate less accurate.) He also correctly predicted the winner in every state except Indiana, which Obama won narrowly. It's worth asking whether Silver would have been as accurate in a year that was bad for Democrats. My impression is that he doesn't let his biases interfere with his mathematical estimates.

Sometimes good predictions come from unexpected quarters. During the 1996 primary campaign, liberal comedian Al Franken correctly predicted that Dole's running mate would be Jack Kemp. He based his conclusion on a quip by Newt Gingrich that Kemp (a former NFL player) has showered with more blacks than most Republicans have shaken hands with. Franken followed this quote with a list of "Politicians Who Have Showered With Blacks," consisting of former athletes who went into politics.

In 1999, Franken released a book describing a bizarro version of the upcoming 2000 election in which Franken himself becomes the Democratic nominee, with an all-Jewish staff. And guess who his running mate is? Why, Joe Lieberman, of course. His stated reason is that he wants to balance the ticket because "I'm Reform and he's Orthodox."

After these two episodes, I began to pay more attention to Franken's musings. Maybe buried beneath the comedy was some sound insight, I figured. That's why I was puzzled when his 2005 book hinted that the next president would be Barack Obama. At the time, Obama had told the press "unequivocally" that he would not run in 2008. I thought to myself, "I guess Franken is finally wrong about something." Hmmmph. (To be fair, I should mention that the book also predicted that the Republican nominee for 2008 would be Bill Frist.)

What's Franken's secret? Good instincts, or dumb luck? You be the judge. Nobody can know with certainty what the future holds. But I value predictions, because they are opinions with results. They test a person's capacity to think objectively, without letting wishes or fears get in the way. They also tell us something about the quality of the person's reasoning. And they help us weed out the shills.

No comments: