The expectations game, Keystone edition

From the Economist's Election '08 blog:

The Clinton campaign would have us believe that after all the resources Barack Obama has poured into Pennsylvania, anything but a huge victory for him in that state's primary should raise grave doubts about his electability. But of course, Mr Obama has always trailed Hillary Clinton by huge margins in the Keystone State, which is precisely why he's had to spend so much cash, in hopes of avoiding an embarassing rout.

Nevertheless, what is the appropriate "expectations" benchmark with polls bouncing erratically week to week? Does Mr Obama "beat expectations" if, as per the most recent polls, he holds Mrs Clinton below the double-digit leads she was showing in surveys a month ago? Or are new expectations established by the polls conducted this month? Or, rather, by whatever results are released in the next few days? How long before we decide that a "benchmark" as fluid as "expectations" is essentially meaningless, allowing both campaigns to spin in their preferred direction by selecting their preferred baseline?

Amen to that.

Alas, I'm afraid the media will probably fulfill the prediction of commenter "yea" last week:

who could have predicted this would happen? obama is trailing by a huge amount in a state he's never been to. he shows up in the state and the margin starts to decrease slowly. all of a sudden he surges and even takes the lead in a few polls. things stabilize as the election gets closer, and then there is a natural drift back to hillary that allows her to win the state. she wins the state be a smaller margin than anyone thought possible 4-5 weeks ago, yet by a bigger margin than most of the late polls indicated. hillary then claims the momentum.

*sigh*