Observers, myself included, underestimated both the Muslim Brotherhood and many peoples' desires to return to something approximating the old order.
I blew it. There is no other way of putting it. The following two sentences from a CFR "Expert Brief" that I posted on May 21st are, without a doubt, my Scott Norwood moment:
The declining fortunes of the Brothers' presidential candidate, Mohamed Morsi, who is trailing badly in the polls, signals the group is paying the price for the decision to run a candidate despite earlier commitments not to do so. Although Egyptians supported the Brotherhood in parliamentary elections, the Brothers' about face on the presidential elections clearly evokes the hypocrisy of the Mubarak era.
My only consolation is that I am in pretty good company. A variety of polls (which can be read here and here), news reports ( such as these from the Washington Post and The Wall Street Journal ) and expert commentary all came to the conclusion that Morsi was going to lose badly. To be sure, a number of keen observers--all friends--got it right ( such as the Sandmonkey and Elijah Zarwan). Still, the fact that so many well-informed observers were "wide-right" on the elections speaks to the perils of prediction.
So what went wrong? Israel's President Shimon Peres allegedly once said, "Polls are like perfume, beautiful to smell, but poison to drink." Peres never won an election in his own right and may have ignored too many polls, but his underlying point remains valid. Take the results of polling, especially if you don't have access to the methodology and a full picture of the questions, with a healthy dose of caution. Indeed, when it came to Egypt's first round, many observers gave the pre-election polling a bit too much credibility. Clearly, we should have known better. Some of my colleagues have commented on the small sample size in some of the pre-election polling, but I am not sure that is the problem. In general, the sample sizes of the Egyptian polls I looked at were less than a 1,000 people, but that is not much smaller than public opinion polls conducted in the United States.