This paper introduces a systematic way to analyze fuzzy data in both engineering fields and economics, with emphasis on fuzzy engineering economics. The approach is statistical in nature, in which fuzzy information and data are treated as bona fide random elements within probability theory. This provides not only a coexistence for randomness and fuzziness in the complex task of handling all kinds of uncertainty in real-world problems, but also a statistical theory supporting empirical analyses in applications. This can also viewed as a complement to two usual approaches in the literature, namely, either using only fuzzy methods, or using some forms of fuzzifying statistics. We will give illustrating and motivating important examples, in the area of regression (for prediction purposes) with seemingly unobservable variables, in which, fuzzy rule-based technology provides nonlinear models for estimating unobservables (from determinants/causal variables), followed by statistics with fuzzy data in linear regression models. The main contribution of this paper is the rigorous formulation of statistics with fuzzy data using continuous lattice structure of upper semicontinuous membership functions (random fuzzy closed sets) which can be used in a variety of useful applied situations where fuzziness and randomness coexist.