An alternative to the frequentists and Bayesian principled arguments
The Fiducial principeled argument is an alternative to the frequentists and Bayesian principeled arguments. It was suggested and promoted by R.A. Fisher although never widely accepted. In it's simplest form, it is used to created estimation intervals using a pivotal quantity. This quantity is derived using the sampling distribution alone, and yet it is attributed to the distribution of the parameter. It seems the origin of the controversy around this argument, is the justification and sufficient conditions for the sampling distribution to coincide with the parameter distribution. For a discussion, see:
On the Difficulties Inherent in Fisher's Fiducial Argument A. P. Dempster Journal of the American Statistical Association , Vol. 59, No. 305 (Mar., 1964), pp. 56-66
