iSixSigma

Gage RR ANOVA vs XBar and R Method

Six Sigma – iSixSigma › Forums › Old Forums › General › Gage RR ANOVA vs XBar and R Method

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • February 18, 2010 at 5:54 pm #53298

    Ashman
    ★★★★★★★★★★
    Member

    I am confused by some results I’m getting from Minitab’s Gage R&R analysis. I thought the ANOVA and XBar and R methods would have very similar results. However, I’m getting some big differences.
    ANOVA:
    Total Gage R&R %Contribution = 18.89 (Repeat 15.84, Reprod 3.04)
    Total Gage R&R %Study Var = 43.46 (Repeat 39.8, Reprod 17.45)XBar and R:
    Total Gage R&R %Contribution = 3.88 (Repeat 2.7, Reprod 1.18)
    Total Gage R&R %Study Var = 19.71 (Repeat 16.44, Reprod 10.86)
    Why would there be such a big difference in each result? I know the ANOVA takes into account the operator by part variation and that is why it is considered more accurate.

    0
    February 18, 2010 at 6:15 pm #189514

    Mikel
    ★★★★★★★★★★
    Member

    Better give us your numbers. I don’t believe your results.

    0
    February 18, 2010 at 9:27 pm #189527

    Ashman
    ★★★★★★★★★★
    Member

    8 parts, 2 operators, 3 trials.  This is leak test data.  Its either a non-leaker (0), fine leak (E-08/9), or gross leak (E-4). 
    I believe the ANOVA % is high due to the large variance.  My understanding is that Xbar/R just looks at the range. 

    Part
    Operator
    Measurement

    1
    1
    1.40E-08

    2
    1
    7.70E-09

    3
    1
    3.00E-08

    4
    1
    1.00E-04

    5
    1
    0.00E+00

    6
    1
    0.00E+00

    7
    1
    0.00E+00

    8
    1
    0.00E+00

    1
    1
    1.80E-08

    2
    1
    5.50E-09

    3
    1
    2.90E-08

    4
    1
    5.50E-04

    5
    1
    0.00E+00

    6
    1
    0.00E+00

    7
    1
    0.00E+00

    8
    1
    0.00E+00

    1
    1
    1.50E-08

    2
    1
    6.60E-09

    3
    1
    3.50E-08

    4
    1
    6.90E-04

    5
    1
    0.00E+00

    6
    1
    0.00E+00

    7
    1
    0.00E+00

    8
    1
    0.00E+00

    1
    2
    4.60E-09

    2
    2
    4.50E-09

    3
    2
    4.50E-08

    4
    2
    8.00E-04

    5
    2
    0.00E+00

    6
    2
    0.00E+00

    7
    2
    0.00E+00

    8
    2
    0.00E+00

    1
    2
    8.70E-09

    2
    2
    3.20E-08

    3
    2
    4.40E-08

    4
    2
    7.90E-04

    5
    2
    0.00E+00

    6
    2
    0.00E+00

    7
    2
    0.00E+00

    8
    2
    0.00E+00

    1
    2
    8.10E-09

    2
    2
    5.40E-09

    3
    2
    3.80E-08

    4
    2
    5.10E-04

    5
    2
    0.00E+00

    6
    2
    0.00E+00

    7
    2
    0.00E+00

    8
    2
    0.00E+00

    0
    February 18, 2010 at 9:59 pm #189528

    Mikel
    ★★★★★★★★★★
    Member

    My take is you have an attribute study that you have assigned
    numerical values to. All of your variation (99%+) is in part 4.It’s not an issue with xbar vs ANOVA. It’s an issue with a misuse of the
    tool.This is also quite possibly a destructive test (leak rate will change over
    time).

    0
    February 18, 2010 at 11:27 pm #189530

    Mike Carnell
    ★★★★★★★★★★
    Participant

    Just a thought but maybe plot the readings by part in the order they were actually run. At least do #4.
    Just my opinion.

    0
    February 19, 2010 at 10:34 am #189535

    Craig
    ★★★★★★★★★★
    Participant

    A high leaker will show more variability over time…correct? This is why  the measurements are less repeatable for the high leaker. It does appear to be variables data, not attributes data and It doesn’t appear to be a destructive test either. I think your comment about leak rate changing over time hits the nail on the head. (the part in the study is changing throughout the study). I’d like to see Mikes suggestion (how do the leak rates plot out by run for number 4?)
    Also, if I were the original poster I would be studying the formulas in the XBAR/R versus ANOVA methods to understand where the estimates come from. It really does appear that the method for estimating variation between the two approaches is a factor in this case.

    0
  • Author
    Posts
Viewing 6 posts - 1 through 6 (of 6 total)

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