In Saudi Arabia, the government refused to allow the question to be asked at all.

Pro-American feeling does not necessarily translate into pro-American action.

People across Western Europe mourned the 9/11 attack. But a Gallup poll conducted the week after 9/11 found that only 29% of the French, 21% of Italians, 18% of the British, 17% of Germans and 12% of Spaniards supported military action against countries that harboured terrorists. Iraq is not the reason that NATO has trouble persuading European governments to send troops to fight in Afghanistan.

Anti-American feeling is often an artefact of propaganda.

Anti-Americanism becomes stronger as the media become less free. Russian anti-Americanism began its measurable rise in the polls in the late 1990s, coinciding with the reassertion of state control over the Russian media and the imposition of a strong nationalist message.

Behind all these particular problems, there is a larger defect, maybe best explained by an analogy.

Suppose we were studying anti-black hatred. Would we begin by trying to figure out what blacks had done to justify hatred--and then offer suggestions about how they might alter their behaviour so as to give less offense? Yet that is how the conversation about anti-Americanism often proceeds.

Critics of the Bush administration cite various grievances against American policy as the cause of anti-Americanism. But this is a naive understanding of how the human mind works.

An example from U.S. politics may explain. During the prosperous 1990s, Democrats consistently gave a more favourable assessment of the U.S. economy than Republicans. Sometimes the gap spread as wide as 10 percentage points. Partisanship determined perspective, not the other way around.

In the same way, the Islamic world seems to pay attention only to those facts like the Arab-Israeli conflict that inflame their anger, while ignoring facts that might contradict anti-American prejudice, like the use of U.S. military force in the 1990s to defend the Muslims of Bosnia and Kosovo.

In the movies, trenchant remarks like these leave Congress buzzing. In real life, they make no impact at all. Lunchtime came, the chairman banged a gavel, and I stepped out of the Rayburn Building to hail a cab.

I'd seen the scene a hundred times in movies and television: the big hearing room on Capitol Hill, the chairman with his gavel, the spectators crowded at the back, the massed cameras and the lonely witness crouched into his microphone, speaking to the U.S. Congress and the world.

Real life is less dramatic. On Wednesday morning, I was the man at the microphone, speaking to a subcommittee of the House committee on foreign affairs. Instead of the thronged audience, there was a visiting group of students, sprinkled with protesters in pink shirts and underwhelmed journalists.

The topic of the hearing: a new report commissioned by House Democrats asserting that America's standing in the world had collapsed under President Bush. I was called as a Republican rebuttal witness.

The Democrats' report compiled poll data showing a spike in anti-American feeling around the world since 2002. The numbers are indeed troubling--and they surely mean something. But what exactly?

I use survey data a lot in my own writing and research. It's a great tool, when handled with care. But used improperly, polls create the appearance of fact without the reality.

Here's what's wrong with the polls' claim that America's standing has plummeted because of George W. Bush and the Iraq war:

The emotion measured by the polls is shallow and shifting.

The more intensely people care about an issue, the more difficult it is to change their minds. A shift of 5 points in public opinion on the abortion issue would be a huge change. Yet according to the annual Pew survey of international opinion, fully one-fifth of the population of Spain switched from a positive to negative view of the United States between spring, 2005, and spring, 2006.

Where anti-Americanism is strong, it predates George Bush and Iraq.

In the first Pew survey, conducted in 1999, Pakistan already ranked as the most anti-American country on Earth.

The Gallup Organization conducted a huge survey of Islamic public opinion between December, 2001, and January, 2002. It found that a majority of those surveyed regarded the United States unfavourably, with Pakistan, Saudi Arabia and Iran being the most hostile. Significant numbers in all Islamic countries regarded the 9/11 attacks as justifiable. Barely one-fifth of those surveyed accepted that the 9/11 attacks were carried out by Arab men--two-thirds denied it outright.