Nice article. Much of psychology suffers from the failure to replicate experiments, for various reasons like funding, time pressure, and difficulties in obtaining the population required. I've worked in sensorimotor control for several years and recently some researchers have come up with the idea of putting together a database of studies on perturbations during reaching (which is a very widely used paradigm) because they can so often be divergent due to tiny changes in the experiments.

I'd love to see more of this kind of thing in psychology in general, just as I'd like to see registration of medical trials from pharmaceutical companies (with both negative and positive results published) to avoid the all-too pervasive publication bias.