You should get credit for having the balls to use accurate and statistics in the same sentence when discussing a highly politicized topic. 
The marriage/divorce rate is least susceptible to political manipulation. I never argued it was perfect. The point (which I mistakingly thought painfully simple) is you can't follow 10, 100 or even 1000 marriages X amount of time or come up with a real rate. Also, a ton of variables exist that we cannot technically determine without interviewing and following every married individual. Also, there are still many people separated who are technically married yet not living as man and wife.
We all know those stats can say anything someone doing the math want. Polls are pretty much the same thing. I know a pollster and even he'll admit it comes down to how a question is asked when pushed hard enough.
There are just too many variables to ague Tony's method measures what it's supposed to measure.
There are not countless variables which is why you can't name them and you keep referring to shit like "seperated" which is neither here nor there. If you're not divorced, you're not divorced. Period.
Of course you'll never get the "real" rate. But it's more than feasible to take a representative, stratified national sample of say 1000 marriages (assuming we want to be about 95% or so confident), check up on them each year to determine if they are divorced (not seperated, not arguing, not thinking about, not "in the process", but DIVORCED), for a period of say 5 years, and then infer that the 5 year divorce rate = X.
If you were evaluating casuality, then you would have a point about numerous variables. As it stands, we're just talking about divorce.
Now, whether or not anybody has given funding or really even wants to give funding...well, that's another story.
