iSixSigma

CpK For Attributes Data

Six Sigma – iSixSigma Forums Old Forums General CpK For Attributes Data

Viewing 9 posts - 1 through 9 (of 9 total)
  • Author
    Posts
  • #27393

    Paul E
    Participant

    My perception is that, there is not way to calculate capability for attributes data, a process under good capability will show no defects in a time period, or zero defect found in the samples reviewed by our inspector.

    Please give your comments.

    Bye

    0
    #66984

    Praneet
    Participant

    Hi Paul
    There are two parts to your question
    (a) There is way to calculate Process Capability for attributes data. U can use either the Binomial Distribution or Poissons distribution for the same.

    Binomial data
    Binomial data is usually associated with recording the number of defective items out of the total number of items sampled. For example, if you are a manufacturer, you might have a go/no-go gauge that determines whether an item is defective or not. You could then record the number of items that were failed by the gauge and the total number of items inspected. If you are an assembler, you could record the number of parts sent back due to poor fit in the assembly process and the total number of parts purchased. Or, you could record the number of people who call in sick on a particular day, and the number of people scheduled to work that day. These examples could be modeled by a binomial distribution if the following conditions are met:

    ·Each item is the result of identical conditions.
    ·Each item can result in one of two possible outcomes (“success/failure”, “go/no-go,
    ” etc.).
    ·The probability of a success (or failure) is constant for each item.
    ·The outcomes of the items are independent of each other.

    In so far as the second part of ur question is concerned then as long as you have a stable process then you will be having rejections as per your established process capability limits and you can accordingly design your sampling plans to inspect the components produced by you.

    Regards and have a great day

    0
    #67003

    “Ken”
    Participant

    Paul,

    Your thoughts are correct. A capability study using pass/fail (attribute) data consists of estimating the fraction defective observed in the process over a reasonable period of actual production. Capability indices are typically not calculated with attribute data. However, large samples having np>5 could conceivably produce resonable estimates of Cp and Cpk if you set specifications for minimum and maximum defective rates. These specifications are rarely, if ever, set.

    The typical way of conducting an attribute capability study uses the fraction defective estimate with the computed upper confidence bound based on the binomial distribution. Making this estimate allows you to estimate the worst case defective rate expected from the process. A very conservative estimate indeed! You would report this number to management in support of a capability investigation. If you’ve ever done a confidence interval calculation, then you know the sample size is important in estimating the upper confidence bound. In previous, discussions I provide approximate sample sizes required to confidence claims supporting zero observations of defects. Give them a look for additional detail.

    Good luck,

    Ken

    0
    #67088

    RR
    Member

    Do we have to go through all that complexity? Doesn’t the centerline represent your capability?

    0
    #67094

    “Ken”
    Participant

    RR,

    The centerline provides you with only 50% confidence in the estimated quantity. If 50% is good for you, then the centerline is enough!

    Ken

    0
    #67277

    Paul E.
    Participant

    Ken, thanks for yor inputs.

    Paul.

    0
    #67279

    Paul E.
    Participant

    Praneet, thank you so much for your comments

    Paul.

    0
    #89467

    Elizabeth
    Participant

    When studying capability of a process, is it the same to measure 50 units and calculate Cp and Cpk values as it is to inspect 462 (pass/fail) to guarantee a 95% confidence level that the defect rate is no greater than 1?

    0
    #155666

    Pramod Chand
    Participant

    My view is that CpK for attributes Data can be measured. It is simple that first you convert this data in systematic defined way and now assign numerals as per their weightage, now this weighted numbers can be used for calculating CpK of  attribute data.

    0
Viewing 9 posts - 1 through 9 (of 9 total)

The forum ‘General’ is closed to new topics and replies.