Real Clear Politics has a great examination of recent Iowa polling, and averages the latest information to get these numbers:
Averaging has some challenges but probably paints a better picture than any individual polling effort. By this measure the race is pretty much a dead heat (though I HATE ther terms "statistical dead heat" because it misleads people into thinking there is no difference between different numbers. There is a difference. When two numbers are within the polls margin of error it still means the top person is "more likely" to win, but suggests far more uncertainty than if the numbers are outside of the margin of error.
For example Richardson with under 10% has virtually no chance of an Iowa top 3 placement, but any of the top 3 could win this race.
However, I continue to believe Clinton will win due to superior organization and strategy and the football game which will affect men more than women and thus Clinton's popularity among Iowa women will be enhanced, but the game is not likely to be fully reflected in polls.
Prediction: Clinton Wins Iowa and also New Hampshire.