Yes and no. In many ways the probelems we are addressing are much more complicated than were being addressed 100 years ago. Often advanced statistical methods n

Yes and no. In many ways the probelems we are addressing are much more complicated than were being addressed 100 years ago. Often advanced statistical methods need to be used to make inferences. We are often looking for 2% signal over noise, trying to understand complex dynamical systems. Even without anything nefarious, and careful execution of rigorous methods, unmeasured 3rd variables can be different across study samples, and you can't measure everything. Even if you could, you won't have the statistical power. Random assignment is nice, but not every field can do it ethically. Even with random assignment, there is no guarantee that the effect of unmeasured 3rd variables will average to zero. That is. Science is harder than many give credit for. That said, the competition and incentives in science are perverse. Peer review quality has declined, if only because everyone is in a rat race to survive, and being a peer reviewer does not do much for getting you a job or tenure.