Yes - Laplace originally proposed this, it's a good approach (and incidentally the basis for Bayesian meta-analysis). Google "skeptical prior" for more.