If you take the entirety of Google’s index to be a representative sample of English, sure. But the problem is that it’s not, and when you’re dealing with differences in frequencies as low as 1 or 2 hits among several billion, comparing them isn’t going to yield statistically meaningful results anyway.