iSixSigma

Calculations behind attribute sampling tables (SAE ARP9013/1)?

Six Sigma – iSixSigma Forums General Forums General Calculations behind attribute sampling tables (SAE ARP9013/1)?

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • #53499

    Lucas
    Participant

    Can anyone point me to an explanation for the binomial, poisson and hypergeometric probability equations in relation to how they generate sampling tables (SAE ARP/9013/1, ANSI/ASQC Z1.4, and old MIL-105, etc.)?

    I want to know how to define inputs and generate the outputs rather than just refer to the tables so I truely undertand the methods.

    Inputs seem to be lot size, acceptance defect rate threshold (average, max, or ?) of accepted lots, risk of accepting a lot when in fact it is beyond the designed defect rate limit (or risk of rejecting a lot within the limit), assumed actual defect rate of the lot.

    Outputs seem to be sample size (and in this case a lot rejection if 1 defect is found).

    I desire to fully understand and compare this system to others in terms of knowing, controlling, and improving quality to drive performance.

    0
    #190403

    Mikel
    Member

    Pick up an SQC book and read it. Grant and Leavenworth is my favorite, but there are many good ones.

    0
    #190408

    Lucas
    Participant

    Thanks for the reading reference Stan; maybe my question was “too large.”

    My initial research has so far clarified operating characteristic curve characteristics; false and true, positive and negative hypotheses quantification. Practical selection of cutoff point is my next goal of understanding. Sensitivity and Specificity intuitively seem like measures that would not be readily available, so my guess is that I will learn they are modeled using common distributions (as an assumption).

    Maybe I can post my own answer as I learn; and hopefully learn together with others who have similar interest.

    Best of all good things,
    Will

    0
Viewing 3 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic.