How to Mismanage Statistics and Hoodwink People

The latest stench wafting from the dankest, foulest, fetid, and odious corners of the blogosphere fever swamps goes something like this:

"60% of Americans don't support President George W. Bush's handling of the Iraq War."

The explanation is that Americans want the troops out. Suddenly a majority apparently believe Michael Moore's "Minutemen" (al Qaeda head-choppers) should be allowed to run roughshod over those brave Iraqis who've stood up for a chance at democracy and freedom, despite the long odds in that part of the world.

Is that the real explanation?

In a word, no. Manipulating polling data is really very simple. The first rule of manipulating polling data is to fail to distinguish types of negative responders. In all liklihood a substantial number of Americans who "don't support Bush" would favor a more robust approach to terrorism.

In recent weeks it's become clear that both Syria and Iran have funded and supported the "insurgency." That the vast majority of suicide bombers come from outside Iraq is now common knowledge to anyone who reads the news on a regular basis. To complete the third strike, troops in Afghanistan have been complaining the Pakistani border has become a safe haven for Taliban and al Qaeda terrorists.

It is not hard to imagine those on the Right losing support for Bush's handling of the "War on Terror." The only way to properly fight an enemy is to take the fight to his doorstep, his turf. There is no home-field advantage in war. These "negative responders" would not favor withdrawl and surrender to the terrorists as the manipulators of the data would have you believe.

When, or if, I do a next installment on "How to Mismanage Statistics and Hoodwink People" it will be entitled "The Form of the Question: One Word Can Make the Difference."