Randomization is a generally accepted principle of sound experimental design and common practice among working scientists. But Bayesian statisticians reject it, most often because of decision theoretic argument against randomization. I trace it back to Abraham Wald's Theory of Inductive Behavior and argue that Bayesians should concur with Ronald Fisher 's criticism of Wald's analysis of randomization. The paper ends with a Bayesian argument in favor of randomization: randomization can lead to an increase in expected utility.

Even in a theory corroboration context, attention to effect size is called for if significance testing is to be of any value. I sketch a Popperian construal of significance tests that better fits into scientific inference as a whole. Because of its many errors Chow's book cannot be recommended to the novice.

In this paper I give a Bayesian criterion for when an experiment is a test of the theory of the apparatus, rather than a test of the theory of the phenomena, and describe strategies used to ensure that tests of the theory of the phenomena are possible. I extend this framework to low dose electron microscopy which has a stochastic instrument theory and which provides an exception to a thesis by Robert Ackermann on the independence between theory and instrumentation.