The other day, when I wrote that Donald Trump was in big trouble in Iowa (because he is), many commenters said that it doesn't matter, because Iowa does not reliably predict the winner.  So I decided to look into this.  I looked at the history of contested Republican contests in 2012, 2008, 2000, and 1996, looking at the early states of Iowa (Feb. 1), New Hampshire (Feb. 9), South Carolina (Feb. 20), and Nevada (Feb. 23).

In 1996 the results were as follows:

Iowa: Bob Dole 26%, Pat Buchanan 23%

New Hampshire: Pat Buchanan 27%, Bob Dole 26%

Nevada: Bob Dole 51%, Steve Forbes 19%

South Carolina: Bob Dole 45%, Pat Buchanan 29%

Nominee: Bob Dole.

From this dataset, it would seem that virtually tying for first place in Iowa and winning Nevada and South Carolina decisively determined the nominee.

In 2000:

Iowa: George W. Bush 41%, Steve Forbes 30%

New Hampshire: John McCain 49%, George W. Bush 30%

Nevada: No data available.

South Carolina: George W. Bush 53%, John McCain 41%

Nominee: George W. Bush

From this year, it seems that winning Iowa and South Carolina decisively determined the nominee.

In 2008:

Iowa: Mike Huckabee 34%, Mitt Romney 25%

New Hampshire: John McCain 38%, Mitt Romney 32%

Nevada: Mitt Romney 51%, John McCain 13%

South Carolina: John McCain 33%, Mike Huckabee 30%

Nominee: John McCain

From this year, it seemed that winning New Hampshire decisively and virtually tying in South Carolina determined the nominee.

2012:

Iowa: Rick Santorum essentially tied with Mitt Romney with 25% each.

New Hampshire: Mitt Romney, 39%, Ron Paul 23%

Nevada: Romney 50%, Newt Gingrich 21%

South Carolina: Newt Gingrich 40%, Romney 28%

Nominee: MItt Romney

From this year, it seems that tying Iowa and winning New Hampshire and Nevada were most important.  Please don't write me in the comments section how Santorum really won Iowa by a handful of votes a week later.  That's not important.

So where does this leave us?

In three of the four contests, winning or coming close to winning Iowa seems to be a prerequisite to winning.  So if Donald Trump loses to Ted Cruz by a large margin, he's in big trouble.  But if he loses by a small margin, it may not mean much.

Winning Nevada seemed important in only two out of four contests, so I am not sure much can be concluded from that.

Winning in New Hampshire was important in only two of the past four contests, and it is important to note that in one of the two in which it is important, Iowa was basically a tie.  So if Trump wins New Hampshire but loses Iowa by a big margin, that may be a bad sign.  But if he wins New Hampshire and loses Iowa by a small margin, that may be a good sign.

Winning or at least virtually tying in South Carolina was important in three out of four contests.  That's pretty important.  And the one time when South Carolina didn't matter, the winner was the winner of New Hampshire primary.  Since Ted Cruz is unlikely to win New Hampshire due to the mushy moderate makeup of the population, a win in Iowa is not sufficient.  He must also win South Carolina to win the election.

In conclusion, Iowa is pretty important, more important even than New Hampshire.  Winning or getting a close second there is a prerequisite to being the nominee but isn't sufficient in and of itself.  I suspect that the nominee will be the person who wins or comes a close second in Iowa and who wins South Carolina.

At the moment, Trump seems likely to be a close second in Iowa and to win South Carolina, based on latest polls.  But if Ted Cruz continues to surge (he just went from the teens to the 20s) in S.C. in two weeks or an Iowa victory gives him a big boost, that can change things.  Either way, these are the two states to keep your closest eyes on.  You can see my detailed map of primaries and caucus dates and rules here.

This article was written by Ed Straker, senior writer of NewsMachete.com, the conservative news site.