Downloaded from https://www.everyspec.com AMCP M-7t- it may be deleted by a contract modification cpiles of experimental design have been de- when calendar time and funds blgin to run velo.ed so that valid inferences can be drawn. low. These principles are briefly discussed in this paragraph (Refs. 1-4). 34.3 TEST PROt•hDURES A detailed oft oC tewt procedurs must m 34.1 THE POPULATION developed for qualty assurance tests and rell. In experimental design, the population is .ability demonstration tests (Refs. I aid 2). th .a of objects about which inferences ar The procedures am a crucial part of the test to be drawn (Ref. 0. Usually the popula- and must be carefully designed.
The test tion is on actual rather than an abstract procedures ought to Include the followinglpu.I;qipetrlaiit:etnhw , ip~roup, In equipment reliability testtnih, how- , I, PurpEw of test over, Inferences must be made about Items not actually in existence at the time or the 2. Test itsm%-desaiption and sample test, but which will be produced at some selection time in the future. 'For example, when a demonstration test Is to be performed on a 3, Teat monitoring and review procedures complex equipment, only three or four of' the equipments may actually exist and only one or two may he used ror the test. It is 4.
Test equipment required assumed that if the equipments are manufaco- tured under the same production processes STest equipment calibration procedures at a future time, then rellability Inferences 6. Test equipment prooflng based on tests of small numbeov of equip- wents are valid. If a major redesign or retrofit takes place as a rmsult of field ex- 7. Environmental conditions to be parience, the results of previous tests may applied not he valid and additional testing may be 8. Operating conditions tuqulmred. There are no formal rules for dealing with 9. Test-point identification the problem of population deflnitlan, except 10. Deflnition of limits of satisfactory that it I.
neesaary to define as completely performanceas po ble the equipment test procedures perfomanceand test conditions from which inferences arm II. Procedures for conducting t to be made. 12, Test report procedures and docu- The method of drawing tiu random owrple ments. 0.~eteitmines the population about widch in- e ,trences are to be made. Ask the question: "From what population is my sample a truly 34 EXPERIMENTAL DESIGN randosit sample?" Generally, the population about which lnforr.nces legitimately can be Reliability testing usually is performed on made is much more restricted than the a small sample of items, Based on the test e01jtIneer would likv.
ThIls consideration Is results, inferences are made about the popula- extremely important 'n evaluating much of tion from which the somple avia drawn. 16in- the theoretical/empiricml research in reliability. 3-8 k M " I I -