I'd have a problem with such disparity, but reading back on the December NPD thread, this all makes sense. I don't think vgchartz is wrong for the month, I think NPD was wrong for December, and is now making up for it by adding and subtracting some of it's December errors to January and February sales.

I just hope they're finished now and don't end up chopping sales off of March too.

Conspiracy!
MS paid NDP off to make their numbers very slightly over Wii (and PS3) for Feb so they can brag about it at GDC (in March).

One thing that makes the NPD numbers seem a bit strange, is that NPD never adjusts their previous data. Now, say they're 8% off during the month of December, and believe that the Wii sells 8.1 million (and the accurate being 7.5 million)

Later, they get data in that it was in fact 7.5 million. Now, what do they do? They never actually edit their previous numbers, so they're left with two options:

A) Have their totals be incorrect. If they do, this could cause them to lose a lot of credibility, what if Nintendo tells that they've shipped 25 million in NA, and NPD has the Wii at 25 million in the US alone?

This makes the total incorrect, but their monthly sales correct.

B) "Fix" the problem by editing the future numbers. They now believe that the Wii sold 1 million in January, but but it at 900,000 to start "fixing" their 600k overtrack.

This makes their total correct, but the monthly sales incorrect.

As we don't know which they're doing, we sort of have to suspect that either the totals or the monthly sales are incorrect. Therefore, you cannot put NPD as being 100% correct.

As a result of this, VGChartz is bound to have a difference from NPD. If they didn't, that would be a far better reason for VGChartz to lose credibility.