A recent survey indicated that many procedures view risk in a safety-first context. Traditional methods used to impose safety-first constraints in optimization models have often been difficult to implement. This is particularly true when endogenous decisions affect the distribution of the chance-constrained random variable. This paper presents a method whereby probabilistic constraints can be easily imposed upon finitely discrete random variables. The procedure uses a linear version of the lower partial moment stochastic inequality. The resulting solutions are somewhat conservative but are less so than the results using the previously published mean income-absolute deviation stochastic inequality.