Not formal research, but an interesting read on how specious a lot of the "math in nature" observations can be:

http://www.lhup.edu/~dsimanek/pseudo/fibonacc.htm

I personally suspect that confirmation bias plays a huge rôle in this business.