Compare between multiple nonlinear regressions

[butlerga]

As a bit of a statistical amateur, just wondering if anyone has a suggestion regarding testing for differences across multiple nonlinear regressions in MLwiN? The data I am trying to analyse/compare between is six age-and-growth data sets for a species of freshwater fish from Australia. It is not a repeated measures exercise but rather six spatially and/or temporally different data sets. Basically for each sample/fish we have a length (mm) and an age (yrs and portion of yr). Any suggestions (or how to do it would be good ) or ideas would be great.

[joneskel]

One way would be to pool all the six different studies into one data set and create and indicator dummy for each of the six different studies. Then in addition to your overall model add in the set of dummies to see if this makes a difference. You can also interact the dummies with the predictor variables to see if the relationship is different for each study. The interval and tests window will allow you to undertake the testing of the differences. If non are significant you can pool the results over the 6 studies and drop all dummies and interactions.