One thing that allows for this is the oft neglected "method of discovery". The usual "scientific process" concerns itself with how to keep an experiment correc

One thing that allows for this is the oft neglected "method of discovery". The usual "scientific process" concerns itself with how to keep an experiment correct/truthful - but how do we decide what experiments to conduct in the first place? Scientific funding can be biased towards the successes of the past, and so future funding can be guided by political agendas.