Last week, Twitter entered the discussion when a professor claimed that an analysis of tweeting did an unusually good job of predicting the results of U.S. House elections in 2012. A formula based largely on the number of “tweets” for candidates correctly predicted 92.8 percent of House races. The implication was that this model might soon replace traditional polling.

There were many problems with the claim (chronicled by Mark Blumenthal, senior polling editor of The Huffington Post). At a very basic level, though, when more than 90 percent of House incumbents routinely win re-election, the 92.8 percent figure isn’t so impressive.

Still, while the professor claimed too much too soon for the new techniques, the polling industry faces the same challenge as the wine stores dealing with new apps. New technology will fundamentally alter the ways that polls are conducted. Other online techniques will replace polling entirely in some situations. These shifts will be good for everyone except those who defend the status quo.