eronje wrote:Thanks for catching my negative SE blunder, I have negatively biased estimates on the brain...
I still don't understand why the SE is zero for that particular dataset. I don't get an SE of zero for similar datasets of low sample size using the Mt model.
Depends on the data, but unbless you set p=c in M(t), estimates are trash. With only 3 occasions, there are precious few models you can fit anyway: M0, Mt (with appropriate constraints), Mb. And thats it. Save yourself the hassle, run those 3 as pre-defined models, model average, and call it a day.