Comments for PQ Systems, Inc. | Quality Blog

Comment on Stats tip: Within and between subgroup variation clarified by matthewsavage

matthewsavage — Fri, 03 Feb 2012 13:25:30 +0000

Subgroup standard deviation is typically the variability within the subgroup of X measurement values. X might be 5 if your sample size is 5. i.e. How much variation do you have among the 5 values.

Total standard deviation is the variability with your subgroup AND between your subgroups. This is often referred to the standard deviation of the individual values.

When are they the same? When the variability within your subgroup is the same as the variability as the within and between your subgroups. In short, if your process does not have big swings from start to finish, your total variability will likely be similar to the variability you see within a sample.

Comment on Stats tip: Within and between subgroup variation clarified by maria

maria — Thu, 02 Feb 2012 19:39:44 +0000

when are total standard deviation and subgroup standard deviation the same?

Comment on Gage R&R study questions answered by pqsystems

pqsystems — Mon, 19 Sep 2011 12:21:54 +0000

Although we no longer support GAGEpack 4.0 (this version is about a decade old now) GAGEpack 4.0 and GAGEpack 5.0 up to release 5.0.34 perform the calculations for the R&R percentages consistently with the 2nd edition of the MSA manual.

What the MSA 3rd edition recommends is that the specification range be divided by 6 and this figure used to determine the percentages. The new standard appeared in release 5.0.40 and above.

GAGEpack 9.0, the current release, is compliant with the methods shown in the MSA Fourth edition.

Matt Savage
PQ Systems

Comment on Gage R&R study questions answered by Daniel River

Daniel River — Thu, 15 Sep 2011 20:47:10 +0000

For the older software Gage R&R 4.0.106, what was the K-sigma used by the software? Was it 5.15 sigma? What does the newer GagePack use for its calculations? Thanks.

Comment on Visit us in Sacramento at the National Association for Healthcare Quality’s 36th Annual Educational Conference by Matt Savage

Matt Savage — Thu, 15 Sep 2011 13:28:17 +0000

If you have 9 cases that took ~ 1 minute you would say the average is 1 minute. Now let’s say your 10th case took 91 minutes. The average is now 10 minutes. An average of 10 minutes is misleading, but I could say 90% of our cases took ~ 1 minute.

Comment on Visit us in Sacramento at the National Association for Healthcare Quality’s 36th Annual Educational Conference by Jan fuchs

Jan fuchs — Tue, 13 Sep 2011 17:46:33 +0000

I have run across twice in the last two days a discussion on fractile data which I am not familiar with.They give a turnaround time for 90% of the cases. What is the calculation for this? Do you do a cumulative number up to 90% of the cases or do you take the top 90% and “throw out” the 10% outliers. Example of reporting, EMS responded in 3.4 minutes 90% of the time while the over all average was 6.8 minutes. Thanks

Comment on IHI founder, Don Berwick, picked to head Medicare and Medicaid by Part D Cost

Part D Cost — Sat, 10 Sep 2011 11:35:44 +0000

Don Berwick is a great choice for the position. His ideals would hopefully bring the system to a better light.

Comment on Gage R&R study questions answered by pqsystems

pqsystems — Tue, 30 Aug 2011 16:29:53 +0000

Hi Ernesto,

In respect to the good and bad parts, there needs to be bad parts or how would you know if you can find them? This says in essence that you should not be checking (doing R&Rs) processes that do not make bad parts. In the MSA manual they indicate in their example that 25% of the parts selected for the study were in the gray area close to the lower specification and 25% were in the gray area close to the upper specification. If you have only a one sided specification, I would recommend at least 25% and probably closer to 50% of the items to be in the gray area close to the specification.

Gordon K. Constable, PhD
PQ Systems

Comment on Gage R&R study questions answered by Ernesto Villanueva

Ernesto Villanueva — Tue, 30 Aug 2011 15:06:06 +0000

I need to validate an equipment,Good and Bad parts are going to be tested in order to know if this machine is able to reject damaged parts.

-The gage R&R should include only the good parts?
- I only have an acceptance criteria, lower limit = 80 “x”, good parts usually 120 – 180 “x”, bad parts usually 0 “x”.

Comment on PQ in the news by Mary Alice Nicholson

Mary Alice Nicholson — Thu, 18 Aug 2011 18:59:54 +0000

I am so excited to have just found Dr. Cleary’s web site re: her children’s book, Claire and the Talking Hat. I know my daughter Clare, a reading specialist, will be so excited to investigate your work. She constantly looks for remedial reading resources for her library.
Knowing your creativity and talent, I’m sure it will be a very attractive addition for her young students. They always learn much easier when the books they read are equally fun and imaginative.
Mary Alice Nicholson, Ph.D.
ret. Marriage & Family Therapist