The New York Times published the results of a study of 11,000 interviews with voters who call themselves Republican-leaning (but may not all be Republicans – more on that below).  They broke it down by congressional district to gauge the level of support for Donald Trump.  The purpose of their study seems to be to show that places where people do racist web searches correlated nearly perfectly with places where Trump has his strongest support (really!).  But let's ignore this wacko angle (unless we also want to see if Obama is most popular in places where blacks kill blacks) and see if we can get something useful from the data:

Apparently, Trump has a very strong amount of support in the Northeastern states, the South, and parts of the Midwest.

In the Northeast, his support is strongest in New York State, Pennsylvania, New Jersey, and New Hampshire.  It is very weak in Vermont, Delaware, and Maryland, which are not as important, since they have fewer delegates.

In the Midwest, Trump is extremely strong in Michigan and very strong in Illinois, as well as in Missouri.  His support in Indiana and Wisconsin and Iowa is much weaker.  I get the impression that the more urbanized Midwestern states have stronger support for Trump and the more rural areas are less supportive, which would explain why he's struggling in Iowa.

In the South, Trump is doing extremely well, although he is not as popular in the crucial state of South Carolina as he is in the surrounding states.  If you're curious about his weak support in Arkansas, that's because all of Mike Huckabee's supporters live there.

Texas is looking more and more like Ted Cruz country, and the farm and mountain states look to be Trump's weakest region, with the notable exception of Nevada, for reasons that elude me.

The west coast – California, Oregon, Washington – is a mix, with Trump stronger in urban areas and weaker in rural ones.

You could conclude from this map that Trump is going to "Cruz" to the nomination.  After all, in the most populous states, he seems to be at or ahead of the magic 32% he needs to get enough delegates to be nominated.

But a few caveats are in order.  First, the margin of error of the data is 8.7% for each congressional district – they did over 11,0000 interviews spread out over 435 or so congressional districts.  Trump does better with Republican-leaning voters who are actually registered Democrats or independents than he does with registered Republicans.  That means that in states where only registered Republicans can vote in primaries, these numbers probably overstate Trump's support.  He has the support of 40% of unregistered voters, 36% of independent voters, and 43% of Democratic voters who call themselves Republican-leaning voters, but only 29% of registered Republican voters.

The survey also found that Trump has only 29% support among voters who say they are 80% likely to vote, and 40% support among those who are least likely to vote.  That's not a good sign for Trump voter turnout.

In other words, Trump's support is being inflated by people less likely to vote.  This has been supported by anecdotal accounts of people who come to Trump rallies all excited but less sure they will actually vote.

On the other hand, if you're trying to measure enthusiasm, it is hard to deny that Trump has consistently had the biggest, most enthusiastic crowds at his rallies, so who is to say what his turnout will be?

So I think what we take away from this is not that Trump is at 43% in New York or 24% in Vermont; with the margin of error and doubts about whether the sample selected are likely voters, I don't think we can reliably point to specific numbers.  But I think the data is still useful to determine Trump's regional strengths geographically.  All other things being equal, he is strongest in the Northeast, parts of the Midwest, and the South.

If he can win South Carolina, one of his weakest Southern states, he will probably win nearly all the Southern states, and, combined with his Northern, Midwestern, and other strengths, probably will have enough to win him the nomination.  That's why I think South Carolina is probably more important than Iowa (where he is expected to lose) or New Hampshire (where he is expected to win).

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