In fields that are mainly nonexperimental, such as economics and finance, it is unescapable to compute test statistics and confidence regions  that are motivated by previous examinations of the same realized data.  We remind and formalize the inadequacy of the Bayesian and Neyman-Pearson inference theories for such a practice. We present elements of a general econometric theory, called the neoclassical inference theory, that is immune to multiple use of the same data, modulo approximation error. In line with this econometric theory, an empirical saddlepoint approach is proposed.