Vibration has long been a part of laboratory tests for packaged-products. Starting with circular-synchronous motion, then progressing to sinusoidal resonance-search-and-dwell tests, and now with today’s PSD-based random vibration, tremendous progress has been made in the simulation of actual distribution vibrations and the quality of test results. In recent years, however, it has become recognized that the PSD approach can omit, or “average out”, infrequently-occurring but nonetheless significant and potentially damaging events. A number of proposals have been made and test approaches undertaken to address this issue, including high/low spectra, control of kurtosis, synthesis of non-stationary random, use of non-Gaussian vibrations, and shock-on-random. This paper will describe the issue and summarize the various suggested approaches in simple terms. Recommendations regarding further research and possible industry consensus will be made.