Most ISTA® Procedures and Projects require a minimum of one packaged-product to be tested. A single "pass", however, does not provile high confidence that other seemingly identicial packaged-products will also pass the same test. This is due to inherent variation in packaging materials, package components, and the package contents as well as other statistical considerations. ISTA generally recommends replicate testing, using new samples each time. Having three successful tests of identicial packaged-products helps impove the assurance; five or more are recommended when possible. Even ten successful replicate tests, however, do not guarantee that all future tests will also be successful.
There is no definite rule about appropriate sample size; it may depend on the purposes of testing, the desired confidence level, and the availability of samples. ISTA Procedures and Projects specify a minimum number of samples required to run the test and achieve packaged-product Certification through ISTA. In addition, commendation for replicate testing is generally made. ISTA's policy is that if any sample fails any of the tests, then the entire test is considered failed.
The aditional testing time for larger sample sizes need not be a barrier to better test technique. For example, mos vibration test systems will allow the user to test many packages simultaneously, this saving consderable time. In this way, a sample size of five would have essentially the same elapsed time for vibration testing as a sample size of one.
Occasionally, proper samples are not available to meet minimum requirements for a est protocol. The use of non-functional "dummy" products is not allowed in most cases, but samples with minor, identifiable damage, such as minor surface scratches, may be acceptable. The key is: when the test is complete, can we determine if the product was damaged according to the Product Damage Allowance statement developed before testing began?
Another technique is to re-use a product for several test sequences with appropriate inspection to insure that the product has not been damaged. The user must be cautious to o re-use a product that has become more susceptible to damage due to prior testing. In this way, one product and three packages could be used to achieve a sample size of three. The test would be run three times, re-packaging the single available product each time.
Source: ISTA Staff