Yes, he does say that it looks like it was crucial in Clinton's win. He take the MSNBC exit polls and looks at the question "In deciding your vote for president today, was the race of the candidate". This is what he finds :

I'm not even going to go there describing the intentions behind the votes, but Matt should have included the data right above that question because it was asked twice but with 2 difference sets of answer choices :

Yet there are more exit poll numbers that just popped right out at me from the web page. For one, the numbers of minorities is minuscule.

There's also the percentage of people who decided at the last minute who they were going to vote for. From the 21% of undecided, a staggering 55% went to Clinton, confirming suspicions that the negative attacks and red phone ad did work.

And more interesting is the data that may prove that gender doesn't really matter as long as there's a decision by race or ethnicity to be made :

What do I think about these numbers? Lord. Where to begin.

Let me just stress one thing : When it comes to whether gender issues trump race, let's just never forget that it is white women who have most benefitted from the civil rights movement and affirmative action. Those numbers seem to prove the point.

And it brings even more context to Roger Simon's comment about Clinton's victim and victimizer game.

UPDATE:
Ian Welsh calls the sexist vote for Obama.
That may be true but it wasn't enough to give him a win in the state. So I still believe it's accurate to say that those numbers around race gave Clinton the advantage.