Approximate methods

Of extreme practical importance are the approximate methods, which enable us not only to avoid having to use Bayes' theorem explicitly, but also to avoid working with probability distributions. In particular, propagation of uncertainty, including due to statistical effects of unknown size, is done in this way in all routine applications, as has been remarked in the previous section. These methods are discussed in Chapter , together with some words of caution about their uncritical use (see Sections , and ).