Downloaded from https://www.everyspec.com 'MACP 7W".198 14,000 m -1.000 BE0Y-YE REGRESSION ANE FEUNY 10.000 9W1.000 1,100 do- T 1so 0 The flast test can sometimes be used to are considered. Sample sizes can be oati- determine If a product meets Its contractual mated from Table 2-25 and 2-26. reliability requiroments. The second can be used to compare one design with another (Ref. 18). Example No. I Illustrates the procedure. The relationships to be used in performing 2.10 ACCEPT/REJECT TES1TS-SINOMIAL these tests are summarized in Tables 2-23 PARAMETER and 2-24.
Table 2-23 presents the tech- niques for comparing the mean life of a Tests of hypothesis and a-signIficance can product with a previously defined standard. be performed on the binornial parameter, The variance may either be known in ad- These techniques can be used to compare vance or estimated from the data, Table s-reiabUhties of various designs and for com- 2-24 presents the techidques for testing one paring achieved s-reliablity with a require- design against another. Again, the condi- mont, These procedures are described In ' tions of both known and unknown variance Ref. 18. 1 Text continues on page 2-109.1 ,,"