Exit Polls "got it wrong" (Again!)

WSJ.com - The Numbers Guy

I commend this article to you, and actually, the series that this author is doing. He lays out just how wrong the exit polls were, and possible explanations as to why.

Perhaps, for me, the most interesting reason for the innaccuracy in the polling is that Hamas voters intentionally refused to answer polling questions, to protect the sanctity of the elections. Not just worried that Fatah would get concerned, and come out to vote in greater numbers--worried that Fatah would burn ballot boxes. That intrigues me, since so many are convinced that exit polls are right, and actually, apparently believe "more right" than actual counts of votes (remember 2000, and 2004?) I am still amazed at that. Imagine, if you have coins in a jar, and I guess how many (even if I apply some methodology for estimating--counting the number visible... times how many one could fit in the diameter, or some such) and you count out, one by one, how many--which count would you hold as more accurate?

Now, the academic in me finds something else interesting in this article. Dr Said, one of the pollsters, critiques his decision to change his methodology, when it appeared that he was getting the "wrong answer." Alas, researchers often fall prey to expecting an answer, and questioning their results when they don't get that answer. In this case, Dr Said even questions whether his subconscious got in the way.

Imagine if US pollsters actually admitted that their results could, perhaps, have been biased by their own subconscious desires to see on candidate elected over another. Hmmmm....