Polls, in themselves, are a useful tool for gauging public opinion on a given issue. The math behind them is amazing, if you think about it: By surveying a subset of the population at random, polls can model with pretty good accuracy what people think on just about any question.
In the hands of journalists, however, polls can be a dangerous thing. A case in point was the polling leading up to the recent Greek referendum on austerity, which predicted a closer vote than the landslide "no" vote that resulted. In the video above, put together by The Huffington Post's Ben Craw, one pundit after another stresses how close the election would be. But when the results came in, an astonishing 61 percent of Greeks had voted against austerity.
HuffPost's Mark Blumenthal has a good rundown of possible reasons the polls were so off this time -- among other things, the vote was called quickly while public opinion was volatile, and some very confusing ballot language made it hard to know what a "yes" or "no" vote meant.
But here's what journalists so often miss when it comes to polling: Polls aren't infallible. A poll is only as good as the underlying methodology, which is why journalists should pay attention to the details rather than only reporting the top-line numbers.