[Will88]

Are there any numbers you like?

It's not a matter of what anyone likes, but whether or not it's valid to generate stats from denominators whose makeup has changed significantly over the time period used. Ask a statistician or mathematician about stats from inconsistent denominators.

[expat_panama]

OK

Are there an applicable numbers that you can provide as being "valid to generate stats from denominators whose makeup has changed significantly over the time period used" or have you rejected all statistician and mathematician stats?

[1rudeboy]

It's completely valid, as long as the change in denominators (if the change exists) is announced and open. That way, statisticians or mathematicians can challenge the new calculation by comparing it to what the calculation would have been under the old method.

The fact that people rarely take the last step is a strong indication that they can't challenge the new number on the basis of statistics or mathematics. Thus the simple, and transparent, efforts to create uncertainty without doing any work.