The theory of ex -ante welfare measures is well establlished and has been addressed extensively in papers relating to the valuation of environmental resources when environmental variables have a random component. However, there have been many new developments in incorporating risks and uncertainty into economic models, and perhaps more importantly, there seems to be remaining confusion as to how to empirically implement such models. To date, a variety of estimation techniques have been utilized, with varying degrees of success in deriving an ex-ante welfare measure under risk. This manuscript assesses the state of the art by discussing the sources of risk, uncertainty, and error in utility models that incorporate risk. Weare most interested in how to incorporate these ideas into empirical models and we examine how econometric estimation methods can best be used to obtain ex-ante welfare measures. We also present the current thinking on endogenous versus exogenous risks as well as subjective versus "expert" risk measures, and discuss some of the advantages and disadvantages likely to be encountered when using subjective-based risk estimates in empirical applications based on alternatives to the expected utility models.