# ANOM control limits HELP!

Six Sigma – iSixSigma Forums Old Forums General ANOM control limits HELP!

Viewing 6 posts - 1 through 6 (of 6 total)
• Author
Posts
• #32571

Carl H
Participant

I am interested in using an ANOM chart, but want to know how to manually derive the control limits.  I know MINITAB and other packages auto calc this for me, but I want to compare the ANOM limit calcs to the  standard XbarR or S limit calcs.
Anyone have the limit calculations?
Thanks
Carl

0
#87127

faceman888
Participant

It just does a hypothesis test with the H0 being that each mean is = to the grand mean.  I think Ha is not equal (2 sided test) So I guess you would just calculate the decision limits for a t test for the appropriate number of df’s and your selected significance.

0
#87128

Erik L
Participant

Carl,
The control limits for an ANOM analysis are set at:
Grand Avg +/- (H)*(est. sd(xbar))
H is specified by three parameters:

degrees of freedom for estimate of sd(xbar)
alpha-level
# of subgroup averages to be compared
There are tables for the H statistic in Understanding Industrial Experimentation, Wheeler.
Regards,
Erik

0
#87171

Loehr
Member

Hi Carl,
Erik has the right formula, I only want to mention that a ANOM graph has decision limits, not control limts.  It’s a small point, but it’s important to understand the difference between ANOM and control charts.  Wheeler does a great job of describing this difference.
I have used ANOM with good results and wish you the best.

0
#87175

Carl H
Participant

I did open the macro for MINITAB Anom and quickly found that the H statistic calculation looks pretty cumbersome…
I will try to get to the H table values you described.
One thing I found interesting is that for random generated normal data with 10 pt subgroups, Anom seemed to have slightly tighter decision limits versus XbarR or S  control limits.
I used ANOM on residuals from a time series regression I did with MONTH as the factor.  It did show significance for one month using ANOM where it did not using control charts of resid. subgrouped by month.
Thanks
Carl

0
#87178

faceman888
Participant

Carl,
One contributor might be the significance level of the decision limits.  The default in ANOM in mintiab is 95%.  The default for control chart limits is +/- 3 std dev which I think gives a significance around 99.74%.  That coupled witht  the different statistics — H vs some estimate of std dev should account for the difference between the ANOM and control chart limits.

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

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