I've encountered this problem several times.
You have to pay specific attention to your data source. I've used CSV files generated through Access without date formatting being applied and ended up with a blend of US and UK dates.
So you'll get 12/25/2010 and 25/10/2010 which is easy, however, 03/03/2010 is not quite as simple to determine if you have a mix.
Check the data before hacking it apart, you could be making more of an issue for yourself by trying to correct it than not. If you are getting this data from another source, the issue most likely lies there. More often than not I find this to be the case.
However, if you've got consistent US dates then this works well for me, as Jeff has already suggested.
=DATEVALUE(MID(A1,4,2)&"/"&LEFT(A1,2)&"/"&RIGHT(A1,2))