Quote Originally Posted by Nessus View Post
I'm missing something; what is the purpose of any model but to describe the world within some accuracy parameters?
To abstract away the main/general components of reality from the noisy/idiosyncratic ones.

Quote Originally Posted by wiggin View Post
AFAIK there is no set of historical data that follows their model to steady-state conditions. Thus, making claims that the model does more that describe a potential reason for the observed dynamics in the unsteady case is unscientific. I'm not sure the paper's authors made that claim (though their analysis certainly is suggestive of it), but the article covering it most definitely did.

I love modeling; I think it's a very useful tool. Yet a too-simple model can have just as many flaws as an overly complex one. I feel that this model was far too simple to really understand what's happening with religious affiliation in parts of the world, and any extrapolations from there are extremely suspect.
I think those are two separate issues. Even a good model doesn't really allow you to extrapolate into the future. At the very least, you'd need the predicted events to have been within the realm of the sample (i.e. religion would have to have been extinct before).

I just read the paper. The assumptions it makes are pretty problematic, and little attempt is made to justify them. One can obtain any conclusion one wants as long as one is willing to make certain assumptions.