Burn-in and stress screening are becoming increasingly popular in the commercial electronics industry as customers become increasingly sensitive to failures occurring in the useful life of a product or system. For example, thermal stress screening (TSS) is an assembly-level electronics manufacturing process that evolved from the burn-in processes used in NASA and DoD programs. While burn-in subjects the product to expected field extremes to expose infant mortalities (patent failures), TSS briefy exposes a product to fast temperature rate-of-changes and out-of-specifications temperatures to expose failures that would occur during the useful life of the product. As a result, TSS ages the product in hopes of bringing out defects that would otherwise occur in the field (called latent defects). In support of this known failure behavior, the classical bathtub curve should be modified to aid in the economic modeling of various screen types. We have conducted extensive modeling efforts that have resulted in a systematic approach to explicitly model the latent failures in the bathtub curve. In this paper, we describe the efforts that have been dedicated to modeling latent failures known to exist in many products and systems. The resulting failure distribution is a truncated, mixed Weibull distribution. This model is proving to be an effective and relatively simple means to model the complex nature of failures of a system. With this increased flexibility, we can measure the impact of stress screens in varying conditions and ultimately design optimal screens.