Sampling by Variables
Webinar: Sampling by Variables When: Wednesday May 26, 2021 – 03:00 PM Duration: 60 Minutes.
ANSI/ASQ Z1.9 (formerly MIL-STD 414) is an acceptance sampling standard for product characteristics that are measured in terms of real numbers. While the textbook approach is very straightforward, users need to be aware of considerations such as (1) the assumption that the data follow the normal or bell curve distribution and (2) switching rules for normal, tightened, and reduced inspection. This presentation will cover the basics of ANSI/ASQ Z1.9 along with these other considerations. The handout for the presentation will meanwhile include a technical appendix that shows where the operating characteristic (OC) curves in the standard comes from.
Areas Covered in the Session:
1. Introduction to ANSI/ASQ Z1.9, formerly MIL-STD 414. (The examples in the webinar will use the public domain tables from MIL-STD 414 as opposed to the copyrighted ones in ANSI/ASQ Z1.9; the procedures are however identical.) It is vital to realize that the entire standard relies on the assumption that the quality characteristic follows the normal or bell curve distribution.
· Other considerations include the need for random sampling, i.e. each item has an equal chance of being measured, and the need for the items to come from (as an example) a lot that has been subjected to homogenous processing conditions.
2. The sampling plan depends on (1) the lot size, (2) the level of inspection, and (3) the acceptable quality level (AQL) at which it is desirable to accept most (roughly 95%) of the lots.
· Switching rules meanwhile dictate normal, tightened, or reduced inspection.
· Plans depend on whether the standard deviation is known or must be inferred from the sample statistics.
· Plans also depend on whether we are dealing with a 2-sided or 1-sided specification, and whether the (1) known standard deviation, (2) sample standard deviation, or (3) sample range is used to infer the variation. In addition, when there are two specification limits with separate AQLs, it is possible to use ANSI/ASQ Z1.9 regardless.
· When the sample range is used, it is vital to note that average ranges (e.g. the average of two ranges for groups of 5 from a specified sample size of 10) are used as opposed to the range of the entire specified sample (e.g. the range of the 10 measurements). If the sample size is 15, then the average of 3 ranges for groups of 5 would be used similarly.