iSixSigma

Attribute Data: definition

Six Sigma – iSixSigma › Forums › Old Forums › General › Attribute Data: definition

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • March 5, 2007 at 2:56 pm #46297

    Schuette
    ★ 10 Years ★
    Participant

    The definition of Attribute Data on this site’s dictionary (link at bottom) contains one highly disputable statement, “No analysis can be performed on attribute data.”
    While it’s true that the many variable/continuous data analysis tools (histograms, boxplots, etc) cannot be used, there are ways to analyze Attribute data.  Pareto analysis is probably the most common technique.  There are other less technical ways to analyze Attribute data, e.g. bar charts, etc. 
    If this is agreed, can the definition be updated?
    https://www.isixsigma.com/dictionary/Attribute_Data-95.htm

    0
    March 5, 2007 at 4:22 pm #152754

    The Force
    ★ 10 Years ★
    Member

    Attribute data was defined in the link as binary — I agree that no pure analysis can be done here that’s why it was also indicated that it needs to be converted to discrete (either proportions, ratio, or count) to be able to interpret and analyze these.

    0
    March 5, 2007 at 5:04 pm #152755

    Mikel
    ★ 10 Years ★
    Member

    The definition is wrong – the part about analysis should be removed.
    Analysis techniques such as Contingency Tables are commonly used
    for Attribute data.

    0
    March 5, 2007 at 8:52 pm #152764

    Measurement Theory
    ★ 10 Years ★
    Participant

    The definition is based on the Euclidian rather than Steven’s theory of measurement. Within this theory it is correct. While the two theories are incompatible all textbooks nowadays subsume “attribute” data under Steven’s theory and leave out the whole discussion that we are really dealing with apples and organges. Good catch, however, it will take another hundred years or so to correct the “little” changes that were introduced in the 1930s and 1940s.

    0
    March 6, 2007 at 10:09 pm #152818

    PV
    ★ 20 Years ★
    Participant

    As the name suggests characteristics of “attribute data”  can be explained only by certain attributes they have i e good/bad, pass/fail, red/blue/green, accept/reject, defect/non defect etc. This kind of data cannot be explained in terms of a mean or std deviation or other metrics.
    As said earlier analysis can be performed on this data using contingency tables etc

    0
  • Author
    Posts
Viewing 5 posts - 1 through 5 (of 5 total)

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