AIC scores for a two (or three) part model

#1
Data is # of individuals through time. I fit logistic and exponential growth curves to the data. An event (population bottleneck) takes # to nearly zero at one point. The growth curves are fit in parts: one on either side of the event.

One thing I need to assess, though, is whether splitting the fit into two parts creates a measurably, meaningfully better fit of the data than just modeling it all as one times series.

How do I calculate the "total" AIC for the two curves which describe the entire population history? I assume it is not simply adding the AIC scores you can get from each piece separately.

Any help is greatly appreciated. Thanks!
 

hlsmith

Omega Contributor
#2
Can't you format data to have it all in one model by using series of dummy codes, etc?

So pretty pics of the data would be nice!
 
#3
Can't you format data to have it all in one model by using series of dummy codes, etc?

So pretty pics of the data would be nice!
Here's some example data. The black line is the curve in question, and we're fitting different growth curves to it, broken into pieces.