Maybe you mean that a frequentist method assumes a certain model, for example normal. A frequentist method will sometimes give similar results to a Bayesian analysis that uses a non-informative prior. For example, if we want to estimate a value from repeated measurements with Gaussian noise the frequentist result is equivalent to the Bayesian result if a "flat" (improper) prior is used. [http://en.wikipedia.org/wiki/Jeffreys_prior#Gaussian_distrib...]