Broadside  

Lies, Damned Lies, and Election Analysis

by R.W. Bradford

Whether it is mainline pundits or Libertarian Party leaders offering election analysis, the results are the same: idiotic claims that ignore obvious facts.

---

The morning after Kerry conceded the election, one of Liberty's editors was rather upset. "All over the media," he said, "there are reports that Bush won because ballot measures banning gay marriage brought out new voters who supported Bush." He was upset by this because he had voted for Bush and thought this was the usual left-wing attempt to smear opponents as bigots.

I told him that I doubted this theory was correct. "For one thing," I said, "so far as I can recall, the anti-gay marriage measures were mostly on the ballot in states that were strongly pro-Bush anyway. And one was on in Oregon, which is pro-Kerry anyway. The only battleground state that I can recall it was on the ballot is Ohio. But Bush won Ohio by almost a quarter million votes, and I doubt that the measure brought that many new voters to Bush there. Anyway, checking this out is a simple enough matter. All you have to do is see how Bush performed in states where the measure was on the ballot in comparison to his performance in states where it wasn't."

After a few minutes on the net, I had my answers.

There were no data that supported the notion that Bush gained votes because of the gay-marriage ballot measures, let alone that the measure had won the election for him. In fact, the data suggested that he would have done better if the gay-marriage proposals had not been on the ballot.

In states that voted on the gay-marriage ban, Bush increased his vote share from 53.33% in the 2000 election to 54.17% in the election just past. That's an increase of 0.84%. In states where gay-marriage bans were not on the ballot, Bush increased his vote share from 48.82% to 50.78%. That's an increase of 1.96%. Bush's vote share rose more than twice as much in states where voters didn't have a chance to ban gay marriages.* ¹

The evidence suggested that the gay marriage measures actually hurt Bush — and hurt him substantially. And this makes a lot of sense, if you think about it. Sure, the gay marriage measures may have brought more religious anti-gay voters to the voting booth, and these voters may have voted mostly for Bush. But it is just as likely that it brought a lot of other new voters to the ballot box: young, urban gays who were offended by the proposed ban on gay marriage. And these young voters may very well have influenced others: virtually all gays have heterosexual parents, and although these parents may not be crazy about their progeny's sexual orientation, they also may strongly prefer that their offspring get involved in more-or-less monogamous relationships, if only for health reasons. These groups and their families may very well have outnumbered anti-gay-marriage Christian voters.

This hypothesis may be wrong. But at least it is consistent with the evidence. The hypothesis that anti-gay-marriage ballot measures helped George Bush is not consistent with the evidence.

Yet on the cable news channels, in major newspapers, and on broadcast political shows the theory that the anti-gay-marriage measures had helped Bush continued to be stated, almost as if it were a fact.

In a sense, this is easy to understand. Two political groups had strong motives to propagate the groundless theory. The anybody-but-Bush crowd, most of whom believed that Bush was such an evil man that voters would surely reject him, found solace in the theory because it explained why their prediction was wrong, and it explained it in a flattering way: they had miscalculated because they had failed to appreciate how many small-town and rural religious bigots exist out there in the red states. Not only did it let them off the hook, but if they happened to earn their living from leftist politics in general and anti-Bush or anti-GOP politics in particular, it gave them an excellent way to raise funds. "We must fight against religious bigotry!" is an excellent battle cry for rallying their troops and getting donors to cough up cash.

Opponents of gay marriage, and anti-gay people in general, also had a strong self-interest in the hypothesis. It flattered them. It proclaimed them victors in the culture war. Not only were they so successful that they passed every measure banning gay marriages, but they also saved the president! It certainly wouldn't hurt their fund-raising abilities either.

But what is strange about this is that for more than a week, the theory was treated as gospel by much of the political class, and virtually no political analyst, commentator or pundit bothered to look at the actual evidence, despite the fact that the hypothesis was child's play to check out. Those who challenged the thesis were content merely to disagree; with a single exception, none bothered to look at the data.

But what is strange about this is that for more than a week, the theory was treated as gospel by much of the political class, and virtually no political analyst, commentator or pundit bothered to look at the actual evidence, despite the fact that the hypothesis was child's play to check out. Those who challenged the thesis were content merely to disagree; with a single exception, none bothered to look at the data.

That exception was a feature on Slate.com, which examined the vote totals and arrived at the same conclusion that I did. Unhappily, this analysis provided little detail and made a claim that was contradicted by the actual data, viz., that Bush's "vote share averaged 7 points higher in gay-marriage-banning states than in other states (57.9 vs. 50.9)." In actual fact, Bush's vote share in the states with ballot measures was 3.39 points higher than in other states (54.17 vs. 50.78).

It also claimed that in the 2000 election, "Bush's vote share was 7.3 points higher in these same states than in other states." This also is false: Bush's 2000 vote share in the states that had the ballot measure in 2004 was 53.33, versus 48.82 in other states, a difference of 4.51 points.

Finally, ten days after the election, syndicated columnist Charles Krauthammer examined the evidence, arrived at the obvious conclusion, and wrote a column debunking the theory.

Why the long delay?

Politics in the United States has become a sport. Like sports journalism, political journalism has remarkably little interest in looking at actual evidence. And this is fine with most political activists, as it is with most sports fans. Consumers of political punditry seem more than satisfied by the moronic, highly partisan analysis offered by Sean Hannity or Michael Moore. They actually prefer it to the more reasoned (dare I say nuanced?) analysis of Charles Krauthammer or, say, me.

For their part, political activists, reporters, and even campaign managers would rather work themselves up over theories than attempt to verify or disverify them, in exactly the same manner as sports fans, reporters, and managers.

Consider the case of the sacrifice bunt, a tactic once widely employed in baseball. The theory is this: with a runner on first with no men out, the batter should bunt, in hopes that the infielder who gets to the ball will throw to first to retire the batter, enabling the runner to get to second base, where he can be driven home with just a single. In effect, you trade an out for a one-base advance.

From the 1890s until the 1980s, this tactic was very widely employed. It never occurred to anyone who believed in it that it could easily be proved or disproved. All one had to do was examine what happens when it is tried and when it is not tried: add up how many runs are scored in innings in which a runner makes it to first safely with no one out and a sacrifice bunt is used, and how many runs are scored when a batter is allowed to try to get a hit.

Finally, sometime in the 1980s as I recall, a group of baseball scientists did just that. They quickly discovered that in innings when the bunt was tried, fewer runs were scored on average, and fewer innings failed to produce a single run. Many other old strategic controversies in baseball have been similarly resolved by the simple examination of evidence.

A lot of baseball fans, commentators, and managers oppose this sort of analysis and feel nostalgic for the good old days when they could argue endlessly about various tactics and strategies — the same way that many political fans, commentators, and managers prefer to argue endlessly about simple questions like whether gay marriage proposals helped or hurt Bush. But whether in sports or in politics, ignorance of this sort, be it willful or the product of intellectual laziness, is just plain stupid.

Political reporting and sports reporting are both inherently meretricious. Reporters in each domain depend on access to public figures — sports figures and politicians — which often leaves them reluctant to make any serious criticism at all, unless they are extremely partisan, and then their criticism is as predictable as it is tedious.

Reporters and commentators in both fields have a powerful interest in making the events they cover seem both closely competitive and crucially important. If they do not, the audience loses interest. Political journalists have even adopted the language of the sports page: closely competitive elections are "real horse races," attractive candidates are "real contenders," etc.

I just read Newsweek's special election issue, which will be expanded and published as a hardback bestseller within a few weeks. It is sports journalism, pure and simple, an account of a close race, with inside details about the managers, players, and strategies, and distaste for outsiders, like the Swift Boat veterans, who might disrupt the game.

Politics is like sports in other disquieting ways. Consider the huge importance that partisans in both fields attach to victory. Every time you hear the word "historic" in a sports story, you can be sure that the utterer actually means "trivial," as in the Boston Red Sox's "historic, come-from-behind victory" over the Yankees after falling behind three games to zero.

Whether the Red Sox beat the Yankees is of huge importance to fans, but trivial in reality. Is the slim victory of George Bush over John Kerry any more significant? It's obvious that whichever man won the presidency, the republic would survive more or less unscathed. While Kerry has a record as a left-liberal spender (a record any Massachusetts politician must have if he is going to win elections) and Bush has a record as what's now being called a "big-government conservative," as befits any successful politician from Texas, there is no reason to believe that government spending would be less under Bush. Indeed, based on past experience, the evidence is that spending will be higher under Bush — but not disastrously higher. There was little perceptible difference between the candidates even on the issue of the war in Iraq, despite the claims of idiots like Michael Moore. In fact, a case can be made that Kerry proposed a set of policies that was even more pro-Iraq War than those of Bush.

The American republic is a remarkably robust political institution: it survived twelve years of Franklin Roosevelt, a Civil War in which 5% of its entire population was killed, eight years of Bill Clinton, and more nonsense in its public discourse than an imaginative novelist could concoct. It can certainly survive another four years of George W. Bush or even eight years of John Kerry.

And political commentary and analysis remains about 97% bullshit, pure and simple. The people like it that way. Charles Krauthammer will not be honored by his fellow professionals for pointing out that they've been talking pure BS. Some will simply ignore the evidence and go on prattling about the brilliance of getting anti-gay-marriage initiatives on the ballot. Others will simply drop the issue, and move on to other spins. And I shall move on too.

Deconstructing the Libertarian Party Performance

As fairly complete election returns became available the morning after the election, libertarians like me had reason for elation. Libertarian Party candidate Michael Badnarik had apparently captured around 390,000 votes, bettering the total of 2000 candidate Harry Browne and coming surprisingly close to his much better known competitor Ralph Nader, who captured about 407,000 votes — only 4% more than Badnarik.

I was surprised by Badnarik's strength and close showing to Nader, and figured that it might indicate that the Libertarian Party was coming out of a long electoral funk. Given the tightness of the presidential race, the widely-held why-waste-your-vote sentiment among voters, and the lack of resources of the LP campaign, this seemed a very impressive showing.

But, as is my custom before I arrive at conclusions, I examined the evidence as closely as I could. I looked at state-by-state data, actual matchups between Nader and Badnarik, and kept an eye out for correlations that might help me understand what had happened, and why.

The results were not pretty. In every one of the 34 states in which both Nader and Badnarik had been on the ballot, Nader beat Badnarik by substantial margins. Badnarik came closest to Nader in South Carolina, where Nader beat him by 49%. Nader had his widest margin in New York, where he garnered more than eight times as many votes as Badnarik. In states where both were on the ballot, Nader got more than three times as many votes as Badnarik.

So how did Badnarik manage to finish a relatively close fourth to Nader in the national totals? He did so largely because his name was on the ballot in 15 states where Nader's was not. In those states, Badnarik got almost 70% of his own vote total.

This suggested to me that a good share of Badnarik's total vote may have come as a protest from individuals who disliked both major candidates and were willing to vote for any alternative. If that were the case, we'd expect Badnarik to do best in states where he had the least third-party competition. So I sorted the states by the number of presidential candidates who adorned their ballots, and totaled Badnarik's vote share in each state. The results are illustrated in the graph to the left.

Obviously, there is a very strong correlation between Badnarik's vote share and the number of other fringe parties on the ballot, suggesting that a substantial share of Badnarik's vote did indeed come from voters seeking any alternative to the major parties and not particularly enamored or even aware of the Libertarian message. ²

What about the Badnarik campaign's much touted strategy of concentrating its resources in four "battleground" states? Did the substantial expenditures for television advertising in these states pay off?

The four states were Nevada, New Mexico, Arizona, and Wisconsin. The results can be measured by comparing the LP candidate's vote share in 2004 to the LP candidate's vote share in 2000, when resources were not concentrated in those states.

States	2000 vote	2004 vote	Change
Nevada	.365%	.331%	-9.3%
New Mexico	.544%	.385%	-29.2%
Wisconsin	.256%	.216%	-15.5%
All three states	.316%	.262%	-17.0%

In Arizona, a party split meant that 2000 presidential nominee Harry Browne did not appear on the ballot. Here are the results for the other three states:

And how did Badnarik perform this year compared to the LP candidate in 2000 in the other states, where resources were not concentrated? His vote share fell from 0.364% to 0.333%. That's a drop of 8.4%.

In sum, LP vote share dropped 17% in the states where the 2004 campaign focused its money and energy, while dropping only 8.4% in other states. Think about it: the Badnarik campaign did twice as well in states that it ignored than in states where it concentrated its resources. Needless to say, this does not support the idea that heavy spending on television advertising had a positive impact. Indeed, it suggests that perhaps the more voters know about the Libertarian candidate, the less likely they are to vote for him.

There's another way to gauge how the LP fared in the election. The presidential race was widely portrayed in the media as extremely close, and the why-waste-your-vote argument was very much in evidence. Perhaps this factor drove down the LP presidential vote totals in general and in the battleground states in particular.

So let's take a look at how the LP did in its races for the House of Representatives, where such arguments were not much in evidence. There are 91 seats in Congress for which the LP fielded candidates against both major parties in 2004 as well as in 2002. The LP vote share rose in three of those races. It declined in 88. The LP vote share in all 91 races declined from 2.24% to 1.95%. That drop of 12.7% exceeds the drop at the presidential level of 7.7%.

In sum, the election results were very disappointing for proponents of the LP. After contesting nine presidential elections, running thousands and thousands of candidates for lesser offices, spending tens of millions of dollars and untold hours of volunteer time, the LP presidential vote share is down more than 68% from its 1980 showing. That share at the presidential level has declined in every election but one since 1988.

Of course, you won't get this impression from the Libertarian Party's website, where the election results were actually given as evidence of growth. Its lead feature was headlined "More than 20 Libertarians elected to office this year." A careful reading of the feature revealed that not a single candidate was elected to a partisan office.

The second lead concerned the presidential election. "Mr. Badnarik's campaign touched millions of voters and helped to increase the size and strength of the Libertarian Party," executive director Joe Seehusen told LP members. Seehusen credited Badnarik's unexpectedly close finish to Ralph Nader to the "media coverage that is showered upon celebrity candidates." He made no mention of Badnarik's huge advantage in ballot access or the huge margin by which Nader trounced Badnarik in every state in which both were on the ballot. Nor was any mention made of the 12.7% decline in vote share at the congressional level or the 7.7% decline at the presidential level.

Next up was a story about the Badnarik campaign's efforts in battleground states. No mention was made of Badnarik's dismal performance in these states.

Candidate Michael Badnarik was as optimistic as the LP executive director. "We are the new refounding fathers and mothers and we will restore liberty," he told a student newspaper twelve days after the election. "We are making incredible progress. We will use this [election] as a springboard for future success. The campaign was a huge success." He also said he planned to run for president again in 2008.

Libertarian Party professionals, of course, get their livelihood from funds raised by the party and have a powerful vested interest in convincing the faithful that their contributions were effective. Like other political professionals, they spin stories and propagate interpretations that make no sense upon examination. They are simply doing their jobs. Which is exactly what I am doing when I debunk their claptrap.

I suppose Libertarians can take solace in the fact that other third parties did even worse. Nader finished almost 40% below his 1996 vote total — and he didn't even campaign in 1996. The Greens, without Nader on their ticket, got only 112,000 votes, finishing in sixth place, though they did manage to finish ahead of the LP in six states, as did the Constitution Party. It was a bad year for third parties all around — but even within that context, it was a bad year for Libertarians.

* * *

Note on sources

Vote totals for the 2000 election are from the Federal Election Commission's website. For the 2002 congressional vote, I depended on this FEC spreadsheet. For 2004 presidential votes, I used USA Today's website. For 2004 congressional returns, I used Yahoo's election page. Raw third party presidential vote totals came from Ballot Access News' website.

I would like to thank A.J. Ferguson and Sara A. Jones for assistance with researching this article.

---

1	These numbers are slightly different than those I calculated at the time because I have used updated vote totals. In both cases, I used two-party totals because fringe party totals were not yet available and wouldn't substantially change the outcome.

BACK

---

2

There was a -.504 correlation coefficient between the percentage of votes that Badnarik received in a particular state and the number of presidential candidates on the ballot in that state. I realize this sounds like technical gobbledygook to people unfamiliar with statistical analysis. A correlation of this sort explains the dependence of one variable on another in the following way: if you square the coefficient of correlation you get the level of dependence. In this case, the square of -.504 is .254, which means that 25.4% of the variance of Badnarik's vote share in each state is accounted for by the number of fringe candidates on the state ballot.

BACK

© Copyright 2008, Liberty Foundation

Send editorial comments to letters@libertyunbound.com.
All letters to the editor are assumed to be for publication unless otherwise indicated.

Send web site comments to webmaster@libertyunbound.com.