## Forum Replies Created

Viewing 22 posts - 1 through 22 (of 22 total)
• Author
Posts
• #172871

Member

GE is not scraping Six Sigma.  They are adding Lean as part of the tool kit. They have abandonned their in-house Six Sigma Training.  Six Sigma is alive and well in at least one GE partnership, Penske Truck Leasing.

0
#156122

Member

We must work for the same folks.  Sometimes they actually say “Thank You” too.

0
#156117

Member

Look in upper right hand area under search.

0
#155151

Member

Pareto was my first thought, but no statisticak reference from that.  So Chi Square it is.  To set this up you take the total opportunities for each type denial to occur (I extracted this from you data) 70, then list be each denial reason the number of times for Yes denied for this reason and No, not denied for this reason.
Chi-Square Test
Expected counts are printed below observed counts
Reason   Yes     No         Total
1         37     33          70 (Actual Values)
23.33   46.67  (Expected values)
2         25     45          70
23.33   46.67
3          8     62          70
23.33   46.67
Total     70     140         210
Chi-Sq = 8.005+4.002+0.119+0.060+10.076+5.038 = 27.30
DF = 2, P-Value = 0.000
Yes there is a statistical difference in the occurences of each type denial.  From the formula you can determine that Reason 1 is used more frequently than the model expected and Reason 3 is used less frequently more than expected.  The expecteed values are derived from the numbers, there were a total of  70 denials of which each of the 3 reasons could have been used, for a total of 210 (70*3=210)opportunities. There were 70 actual denials of the 210 opportunities and since each of the three reasons had an equal opportunity to occur, the models expects to see each reason having a 33.3% chance of occuring (70/210=33.3%).  33.3% of the 70 Yes =23.33 (the expected value for Y on each reason. For the No (140/210= 66.67%).  66.67% of the 140 No =46.67

0
#154988

Member

Anon, another example of degrees of freedom.
Shirts come in following sizes: XS, S, M, L, XL (3 degrees of freedom.
Or shirts sizes 4,6,8,10,12,14,16,18,20,22 (9 degrees of freedom) The increase in distinct catagories reduces the noise by more strictly defining the district catagories.

0
#154838

Member

Yes, it would be distinct catagories.

0
#154783

Member

I would recommend the Moods Median, robust to non-normality.  And if the data is normal, the mean and median will be close to equal anyway.

0
#154782

Member

F test likes normality.. alot.  Or better said, F test is not robust to non-normality in the least.
Degrees of freedom – How many different types does the test see.  We teach that a minimum of 4 is required in order to use the test.  Higher is better.  It comes down to the noise in the process.  With fewer degrees of freedom, there is more noise that is hidden among the “types” that the test sees.

Hope it helps.

0
#154740

Member

Yes the 99.73 does equate to 3.0 as your confidence level.  Make sure you square, not square root.

0
#154674

Member

Chris, Excellent use of lean concepts.  We will have to bring that into our training.

0
#154672

Member

We currently use both the catapult and the helicopter for training Black Belts.  The helicopter is easier, but more prone to “noise” caused by design defects from the individuals constructing the helicopters.  The catapult isn’t cheap, see attached web site.
http://www.4ulr.com/products/statisticalanalysis/trainingaids_22a.html

0
#154670

Member

jnu, Mike is correct in the explanation of the 1.96.  Give me an email address and I’ll send you a sample size calculator in excell.

0
#154606

Member

Confidence levels – 95% confidence level is again the generally accepted level.  Mitigating factors would be risk and costs.  Higher risk and cost would perhaps lead to a higher confidence level required.  As far as degree of precision you would look at +/- pounds or inches.  Formula is n (sample size) = (1.96*s)squared
2
s is standard deviation and the 2 is either 2 pounds or 2 inches. In the above example if the standard deviation is 3 inches, then you would collect 9 samples. (1.96*3)/2 squared.  or sample size = (1.96*3)/2=2.94.   2.94*2.94=8.64).
Your degree of precision (in example = 2) means that you would be able to measure accurately to plus or minus 2 pounds or inches.  The lower the number or the tighter degree of precision requires higher sample sizes.

Let me know if this helps.

0
#154574

Member

30 is a generally accpted sample number when dealing with continouos data such as weight and height.  If you could get this sample of 30 and determine the standard deviation for the height and weight, you could then determine how many samples would be needed depending upon the degree of precision you are willing to accept.  Send me an email for more

0
#152788

Member

I recommend thaat you go over the operational definitions again with the operators.  The one operator is (or is not) seeing something different from the other two.  You mention repeatability, did you also test for reproducibility?

0
#151885

Member

If the data is “normal”, then yes you can use ANOVA for analysis of the central tendency (mean).  If the data is non-normal then you would need to use a nonparametrics to analyze the central tendency (median), such as Moods median.  Don’t forget to analyze the variation as well, using HOV ( Homogeneity of Variance).

0
#151664

Member

In Minitab unless you have the GE add on “Six Sigma” I do not believe you can do a process report.  However, there are tools available that you can determine capability without the USL.  Look at the menu tabs here on Isixsigma to the right for the Sigma Calculator.

0
#151210

Member

As I side issue I recommend that you look at Net Promoter Score (NPS).  A derivitive from the VOC, it boils down to one question (How likely are you to recommend our company).  This has been receiving a lot of attention as late. http://www.netpromoter.com

0
#151127

Member

The Leadership Contract and Systems and Structures Analysis are initially developed up front in Define or Measure for DMAIC or DMADV.  For Lean activities we use the Leadership Contract at the onset so that we (the Quality folks) and they (the operators) reach an agreement as to what is expected.  Not limited to the leaders either, can go all the way to the machine operator.  Even counter-signing between management and operators has been used with positive effects.  The Communication plan is initiated early as well especially with Lean activities.  We start communicating prior to any activity simply to let people know that there is going to be an event.  For Lean activities we also develop the Behavior Monitoring Plan and the Influence Strategy prior to activity so that we can readily capture and define the behaviors we see  (or don’t see) and have actionable strategies inplace to help align the behaviors activity intent .  For DMAIC and DMADV we show the deliverables in Improve and Verify/Validate but the development should be an ongoing activity.

0
#151116

Member

Steven, As a Change Acceleration Proces (CAP) Facilitator/Instructor and a DMAIC/DMADV/Lean MBB I have learned the hard way that the soft skills are imperative to successful solution development and deployment.  There are 5 CAP deliverables required for my organizations Improvement Projects; Leadership Contract, Influence Stategy, Communications Plan, Attitude/Behavior Monitoring Plan, and Systems and Structures Analysis.  Since instilling these requirements projects have developed faster and been deployed more successfully than before.  Buy in from the Champion and Process Owners is essential from the beginning, if they don’t see a need to change you are handicapped from the start.
“It’s not necessary to change. Survival is not mandatory.”–W. Edwards Deming

0
#90215

Member

What industry? What kind of measurements ( attribute or variable)? Do you want speed or accuracy improvement?

0
#90214