The stress level is the amount of stress that you put on a unit during a pass/fail test (probit analysis) or an accelerated life test. For example, you might use a number of stress values to test under common stress loads. To test a drill press that runs at 4 different speeds under typical use conditions, include the 4 different stress levels in your test.
When the expected life is too long to test practically, you can use stress levels higher than typical use conditions to cause early failures. For example, an engine has an average life of thousands of hours at a standard speed. To cause early failures, engineers test the engine at twice the standard speed. The engineers finish the test in a fraction of the time. Then, the engineers extrapolate time to failure at the normal use conditions from the elevated speed in the test.
Use your knowledge of the system to choose elevated stress levels. Stress levels that are too high may cause the system to suddenly fail instead of accelerating wear. For example, a stress test of a paper egg carton's compression strength at several thousand pounds of pressure would cause an immediate breakdown. This test would not offer much insight into the carton's strength under normal conditions.
You must enter the design and test stress levels. By default, Minitab determines an allocation of units across stress levels that minimizes the asymptotic variance of the parameter you choose to estimate. Alternatively, you can provide the allocation.
The efficiency of the test plan is measured in terms of the variance of the parameter that you want to estimate. However, a highly efficient test plan (one with small variance) can produce parameter estimates that are not accurate. In particular, accurate results depend on having enough failures at each stress level. To get accurate results, a common guideline is that the expected number of failures at each of the test stresses should be 4 or more.