I think Jack Gallant and colleagues get it right by basically applying the method that works well in machine learning. Train your model on one dataset, and then

I think Jack Gallant and colleagues get it right by basically applying the method that works well in machine learning. Train your model on one dataset, and then validate on a held out test set. That should get rid of the majority of all statistical artifacts (as long as you do the split right of course).