iSixSigma

Bias and Reference Value

Six Sigma – iSixSigma › Forums › Old Forums › General › Bias and Reference Value

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • October 26, 2007 at 3:37 pm #48522

    mcintosh
    ★ 10 Years ★
    Participant

    I want to check the bias of lab equipment used in testing of a powder.  The problem I have is finding a reference value.  There are no true standards for this product so I can’t relate anything back to a standard.  One option is to set aside a batch of the powder and call it the master sample.  But how do I determine a reference value?  I could have the best lab person run the sample 10 times and call that the reference value?  But then, if someone else runs it 10 times to see if there is bias, it could just the difference between the operators.  Any advice?

    0
    October 27, 2007 at 2:08 am #164023

    Dr. Scott
    ★ 10 Years ★
    Participant

    Tom,
    Run the Gage R&R the way in which you suggest in your inquiry. As far as having an accurate measure, you will have to rely on the either the best lab you can find or:
    Test the 10 samples, multiple times each, the use the average of the multiple tests of each 10 as the “standard”. Or:
    Ask the powder God what the true value is.
    I would be more concerned with the Gage R&R at this point unless you know the “true” value of each.
    Regards,
    Dr. Scott 

    0
    November 5, 2007 at 12:20 pm #164392

    Rhineg
    ★ 10 Years ★
    Member

    Go to http://www.nist.gov and look for Standard Reference Materials.  You can choose Physical Properties and then select Metrology. There are further refinements in selection, which will depend on what you are trying to determine with your assay (optical, thermodynamic, microscale dimensional, …)

    0
    November 5, 2007 at 3:19 pm #164404

    Pablo Zaltz
    ★ 10 Years ★
    Participant

    Dear Tom,Is not a problem not having a standard of your product because what you want to check is your lab equipment. The best reference material is the one that never changes with time. So…Research if there is a material known to be stable and appropriate for testing the bias of your measurement equipment ***in the same measuring range you are using***. If this material exists, get a sample and follow these steps: – perform repeated measurement of 10 sets of 5 test portions each,
    – calculate -for each set- the mean value,
    – setup a control chart having your mean as the center and ± 2SD as warning limits and ± 3SD as your action limit,
    – establish a procedure for periodic testing that reference material,
    – chart results and evaluate if it’s inside action limits; if so, you equipment is working OK.Good luck,
    Pablo Zaltz

    0
  • Author
    Posts
Viewing 4 posts - 1 through 4 (of 4 total)

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