As noted, the purpose of the demonstration test is to provide assurance of the ability to meet a specified maintainability requirement. The demonstration test is defined as a set of numerical requirements and associated risk levels that will govern the design and decision criteria of the demonstration test. For the test plans to be described, this specification involves decisions regarding the following:

- Type of maintainability index to be specified
- Acceptable and unacceptable values of the index
- Associated risk levels

For example, the test specification might be as follows:

H_{0}: Mean corrective-maintenance time = 40 minutes

H_{1}: Mean corrective-maintenance time = 80 minutes

=
0.20, _{} = 0.10

H_{0} and H_{1} are the null and alternative hypothesis, respectively. a and b are the producer and consumer risks, respectively. For the defined specifications above, the demonstration test must be designed such that the probability of rejecting a system whose mean corrective maintenance time is 40 minutes is 0.20, while the probability of accepting a system whose mean corrective maintenance time is 80 minutes is 0.10. This is presented below in
the form of probability equations.

P (reject | M_{ct} = 40 min.) = 0.20

P (accept | M_{ct} = 80 min.) = 0.10

Typically the maintainability index is that specified in the procurement specification. The index should represent a measure that is directly influenced by the equipment design so that the producer can plan for high assurance of a pass decision, but bears the responsibility for a reject decision. The index should also be appropriate for, and measurable in, the demonstration-test environment. If a demonstration of the chosen maintainability index is to be required, then adequate sampling and statistical-evaluation procedures, such as those described in this appendix, should be available for demonstrating conformance to the requirement. Finally, the specified index and risk values should not lead to sample sizes (i.e., the number of maintenance tasks required to assess compliance) that would exceed available test resources.