iSixSigma

Continuous Data with Attribute Characteristics

Six Sigma – iSixSigma Forums General Forums Tools & Templates Continuous Data with Attribute Characteristics

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #54486

    Tejas Bangre
    Participant

    I have a dataset of 30 data points and is supposed to be continuous data but in the dataset there are only 3-4 values that get repeated so the dataset follows no distribution and cannot even be transformed. How do I calculate capability for such a data?

    0
    #195582

    Doug Von Feldt
    Participant

    This could depend on what you are trying to accomplish. Are you trying to use a control chart or trying to determine cpk/ppk. There may not be enough data to calculate capability if it is continuous and certainly not if the data is attribute data. Do you have specifications for the data? You could just plot the data in a I-MR control chart just to show the data.

    0
    #195583

    Tejas Bangre
    Participant

    Doug,

    I do have specifications for the data and usually I run the capability sixpack which also gives out the control charts but I am more interested in calculating cpk/ppk. I encounter such datasets many times and usually get stranded. For reference the data set is as follows (LSL=0.821, USL=0.827)

    0.826
    0.827
    0.823
    0.824
    0.824
    0.823
    0.823
    0.823
    0.824
    0.822
    0.822
    0.823
    0.823
    0.823
    0.822
    0.822
    0.824
    0.823
    0.824
    0.823
    0.824
    0.824
    0.824
    0.823
    0.824
    0.824
    0.823
    0.823
    0.823
    0.824

    0
    #195585

    MBBinWI
    Participant

    @tejasbangre -if your spec is to 3 decimal places, then your measurement tool must be at least to 4 decimals. Make sure to do a good gage study to ensure adequate measurement system.

    0
    #195595

    Tejas Bangre
    Participant

    MBBinWI: Thanks for your suggestion, I think that’ll resolve issues with some of my datasets. But, even when I go to 4 decimal point for tolerances upto 3 decimals I am still facing this same issue as I am having repeating values. Is there any way to find capability for such a dataset?

    0
    #195612

    Bill McNeese
    Member

    MBBinWI is correct. You need to look at your measurement system. If you use your data to create an individuals control chart, you will see that the moving range chart only has 3 possible values. This is what Dr. Wheeler calls “chunky data” – it is excessive round off error in the measurement. This leads to out of control points on the X chart, which may not be valid. I have attached a workbook with the control chart in it.

    You can still calculate the capability using the standard formulas – it just doesn’t mean much because of the measurement system. Always look at the control chart before doing a process capability analysis. The control chart will tell you what is happening in the process.

    0
Viewing 6 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic.