Of course, there are limits to how far evidence can take us. Evidence is often ambiguous, and so multiple interpretations may be consistent with what we know. Sometimes the facts are wrong. And at times we may lack the necessary evidence when decisions must be made. The Iraq war illustrates these problems particularly well.
While progressives quite effectively assessed the (stunningly weak) evidence around Saddam Hussein's efforts to obtain uranium from Niger, for instance, many prominent hawks such as Ken Pollack had important facts about Iraq's WMD capabilities wrong. The invasion of Iraq required advocating a policy without all the ideal information being available.
Information is a very valuable thing and you can buy new information --- that is why I spent $.50 today to see how the Pirates are continuing to embarrass --- and smart, evidence based liberals had the opportunity to assess the quality of information and reassess their evaluations multiple times. So now I'll start my policy-evaluation macro for Iraq:
There was significant uncertainty in the spring and summer of 2002 as to the status of the ABC programs in Iraq, and SOMETHING HAD TO BE DONE. And the AUMF in conjunction with an approach to go to the UN for coercive inspections could have been that something. If the support for the AUMF was based on a desire to gain information and reduce uncertainty as to the status of the ABC programs of Saddam Hussein, then the vote is defensible if the voter quickly changed their mind.
The inspections combined with increased US and UK surveillance of Iraqi territory should have produced significant new information. And they did. The information was null information that to the best of the US, UK and UN knowledge the biggest threat in Iraq was minor book-keeping errors and some short range ballistic missiles that either were just within the allowable range limits or under unusual circumstances 10-15% over allowable limits. By mid-January one of three conclusions should have been drawn from the new information.
1) There was nothing there
2) US intelligence sucks
3) Saddam Hussein is a competent evil genius who is able to coordinate the movement of multiple large labs, thousands of artillery shells, millions of pounds of highly fragile precise machinery, thousands of individuals without being noticed despite being under some of the most intensive electronic and visual surveillance in the history of the world.
#1 and #2 are reasonably highly correlated and the available supporting evidence strongly supported some combination of the two. The uncertainty costs had been dramatically reduced so opposition to the war would be a coherent position for an AUMF yes-voter to have if this is the justification.[/end macro]
I can understand people being wrong when there is bad information and insufficient information. I can understand people wanting better information before they lend their support to an extremely expensive proposition. I can not understand anyone who says with a straight face that the information environment by January 2003 was the same information environment of August 2002. We knew that either Saddam Hussein was an evil genius with magical Voldemortian powers to apparate away visible signs of massively capital intensive programs or he had next to jack shit.
Anyone who relies on informational uncertainty to argue that their support to invade Iraq in order to (potentially) disarm it was still retrospectively a good decision is either full of shit or delusional.