iSixSigma

P-chart

Viewing 34 posts - 1 through 34 (of 34 total)
  • Author
    Posts
  • #45854

    Talaid
    Participant

    I am trying to do a P-chart for Defective Data. I do have my subgroups = No Constant Sample
    these subgroups data is not large enough, some subgroups has less than 10 data, and the largest subgroup is 20 data, I think this is too small to Constrcut a P-Chart, what shall I do, is there something else I could do to have a SPC chart, where I want to show Stages before and after improvement..  any Suggestions Please..!!

    0
    #150712

    Rohan Kadam
    Member

    Andy,I do have a similar problem. I do collect data from each operator on daily basis & note down the defective piece. I do 100% inspection of items produced by the Operator. While Drawing the P-Chart, we plot Date on X Axis & Rejection% on Y axis.Is that technically correct??How do we interpret P Chart

    0
    #150731

    Ken Feldman
    Participant

    Wheeler suggest using an I/MR chart instead of the p chart with you plotting the proportion.

    0
    #150783

    Talaid
    Participant

    Hi Darth
    I will try the I/MR chart by calculating the proprtion, this is no problem.. but How come No one mention this way in the Six Sigma text..  is this a solid and used method..??!!

    0
    #150786

    Keith
    Participant

    What is your process ?  Charting is only meaningful in the context of a process.
    What are you measuring to determine “defective” ?
    Six sigma is very misleading with its stupid emphasis on defectives rather than measurements.

    0
    #150788

    Rohan Kadam
    Member

    Hi keith,I have explained my process in previous Reply, could you plz advice me, whether my thinking is appropriate or not

    0
    #150794

    Ken Feldman
    Participant

    Yes, it is a common practice.  If you have not learned about it in your SS training, I would question the competence of your instruction.

    0
    #150796

    Anonymous
    Participant

    You can tell that it is a Six Sigma issue since things like p, np, c and u charts were definately introduced by Six Sigma. Perhaps you need to spend more time understanding the quality profession than displaying your ignorance in a public forum. There charts were around decades before Six Sigma existed. They address defects in one form or another and they were some of the first Quality tools available.

    0
    #150815

    Shiran Ranasinghe IBM
    Member

    The general approach to on-line quality control is straightforward: We simply extract samples of a certain size from the ongoing production process. We then produce line charts of the variability in those samples, and consider their closeness to target specifications. If a trend emerges in those lines, or if samples fall outside pre-specified limits, then we declare the process to be out of control and take action to find the cause of the problem…
    P chart. In this chart, we plot the percent of defectives, the control limits in this chart are not based on the distribution of rare events but rather on the binomial distribution. Therefore, this chart is most applicable to situations where the occurrence of defectives is not rare (e.g., we expect the percent of defectives to be more than 5% of the total number of units produced).

    0
    #150816

    Mikel
    Member

    Dear Mr. IBM,Did you know there is a famous company with same last name as you? They are famous for their great successes and equally great failures and today they make loads of money on overpriced and over staffed consulting gigs. You know – for big companies that do what is safe instead of what makes sense. I believe you will find that p charts are based on a normal approximation of the binomial which is an approximation of poisson. I believe you will find that np and n(1-p) need to be greater than 5 or you get wierd limits of either less than 0% or greater than 100%.

    0
    #150817

    mcintosh
    Participant

    As my three year old grandson would say, “You sound like a grumpy old boiler Anonymous! “
    I’ve been passively reading posts on here for several years.  Usually just for humor and very rarely respond, so you should be proud of your self to be such a stand out to prompt participation.
    Things will get better.  As you mature things like, people not understanding SPC will not cause you to be such a grumpy old boiler.
    In the mean time, I think your being anonymous is understandable.

     

    0
    #150818

    Shiran
    Member

    Dear Mr Stan,
    Your comments are simply remarkable… Keep up the good work… !!! And keep doing what you do best …
    S

    0
    #150824

    Keith
    Participant

    You are quite correct.  Six sigma has introduced much mis-information, bad statistics and poor training.
    However lets help this poor fellow who is obviously the victim of another 6s consultant who has no idea what he is talking about.
    We first need to understand what the process is.  This fellow has told us nothing in this thread about what he is doing or trying to achieve. 
    Are you manufacturing, a service industry or what ? 

    0
    #150828

    Mikel
    Member

    The arrogance of trying to impress someone by attaching IBM to your name and then giving bad information is simply remarkable. I hope you find what you do well so you can stop trying to impress us with who you happen to work for.

    0
    #150832

    Anonymous
    Participant

    What does Grumpy have to do with it?
    There are tools that have been around long before Six Sigma that address defects. Defects have been a focus of Quality for years. All of a sudden it is Six Sigma that focuses on defects? So we are indicting Shewhart and Donald Wheeler for the same issue that you have with Six Sigma? 

    0
    #150833

    Anonymous
    Participant

    Mis-information, bad statistics and poor training.
    Interesting. Since you are blaming Six Sigma this must mean you have industry wide information? What do you have behind those accusations or is this just a personal issue?

    0
    #150845

    Alvarez
    Participant

    Grumpy, you are wrong.  Six sigma introduced the focus on defects … and because defects relate to the specification, this is six sigma’s most fundamental flaw.  Even as Bill Smith so stupidly said, you can improve quality simply by changing the specs levels !
    Shewhart and Wheeler focus on reducing variation, not defects.

    0
    #150851

    Charles H.
    Participant

    Reducing variatoin by itself is not good enough.  It has to be relative to some target value.  If I reduce variation on a process with a negative Cpk, I will do a really good job of producing nothing but garbage.  I will also need to shift the mean value of my data to be on target, which is often nominal spec.  You have to do both.  We cannot look at process improvment concepts in isolation.  Continual improvement is the sum of the parts and knowing what to do when.

    0
    #150856

    Deep thoughts
    Participant

    Andrews,
    You should market you insights into a booklet called “Deep thoughts for the professional who has never read a book about quality or six sigma”. … P.S. this is not grumpy under a different name. Just someone who gets a kick out of the strength of the conviction with which utter nonsense is being published on this site.

    0
    #150868

    Deeper thoughts
    Participant

    Grumpy,
    You should market your insights into a booklet called “Deep thoughts for the unprofessional who is too stupid to understand a word about quality or six sigma”.  P.S. This is not the tooth fairy under a different name.  Just someone who gets a kick out of the strength of the conviction with which utter nonsense is being published on this site.

    0
    #150869

    Talaid
    Participant

    good discussion guys..
    all I have is a straight forward question.. if I can not use a p-chart, can I use a I/MR chart..  Darth said yes as the book says.. well I know some is asking what is process.. the problem.. we are trying to control the process for solving technical issue for web and IT apps. for internal and external users/customers.. these cases/requests comes by email from the tech. support by email to the developrs/ sys. analyst to work and solve these cases. the problem is: many of these new unaasigned cases sets in the group list sometimes many day before any one assign it to her/him self and start working on it.. so we have assigned senior level system analysts to take turns every quarter to be in charge of assigning these cases to other individuals to work on them.. now we have to control this process.. we only have 1 biz. day maximum for any case arrive to be assigned.. so the SPC must control this.. each day we extract a report from the data base counting how many cases came in and how many cases assigned out to the tek. guys to work on.. the condition any case has has been setting Unassigned for over 1 Biz. day is considered Defective Case.. so I want to use a p-chart but the number of unassigned cases daily or even weekly does not exceed 20 defectives.. even I have at least 5% defectives most of the times not always (if I take over a week period then it will be for sure at least %5 defectives) so I asked what SPC chart shall I use in such case.. I got the suggestion to use I/MR chart and I calculte the defectives in percentage or proprtions.. I think I will do it weekly… ANY MORE SUGGESTION.. OR MORE INSIGHT TO THIS..

    0
    #150882

    mand
    Member

    The ideal approach is to chart the buz hours taken to service each request. Use a histogram (without the stupid commonly seen, superimposed normal distributions) and XmR chart.

    0
    #150885

    Anonymous
    Participant

    Andy,The defects that are charted in the p, np, c and u charts are defects that are not related to specs?I see now since you said “Six Sigma introduced the focus on defects…” it must be true. Better check when these charts that track defects – not variation – came into use. It was way before Six Sigma. Shewhart, Wheeler, Deming, and Juran all advocated te use of control charts. Those charts include the attribute charts that track defects and those defects are related to specifications – that is how they became defects. You can throw your nonsense about specifications, defects, etc. out there but you know this simply reducing variation doesn’t get it done unless of course you have never actually done any of this and had to build good product for a customer. CharlesH explained simply enough. Give it a try.

    0
    #150886

    Charles H
    Participant

    My dear friends,
    I believe that a Six Sigma professional requires many diverse skill sets.  Obviously, knowledge and ability to apply different tools and methods are a basic requirement.  Also required, in my humble opinion, is the ability to present one’s thoughts and knowledge in a coherent manner, with decorum and respecting the dignity of those we interact with, even those that we may have issues or disagreements with.  The level of rhetoric displayed here is beneath a Six Sigma professional. in my humble opinion.  Bringing personal insults against one another and their families is not a part of the mix.  Let’s try to keep it on a higher plain, shall we?

    0
    #150888

    torres
    Participant

    Can anyone tell, what interpretation you can draw from P chart.What is the significance of Control Limits, when 100% quality is required by the end customer.Can we really differentiate between Special & Common cause in P Chart.e.g:- out of 1800 items supplied, one item is defective. Even if the deviation is into Upper control limit, can we say firmly is it because of special Cause or Common Cause.I draw P Chart on the basis of Daily Production, i.e, Percent Defective. I am hardly able to see any value addition.Can you expert, put some light on this topic.

    0
    #150890

    Sloppy Joe
    Member

    Quite a few assumptions there ole buddy!
    I like Anonymous – at least she has a sense of humour – unlike the guy with the poka yoke, handle bar moustache.

    0
    #150895

    CI Man
    Participant

    we got a similar type of situation, only that our cases come in bulk every day and have to be assigned within the day as well, unassigned cases are considered as defects. at the end of the day the supervisors log the prop and simply use XmR chart for monitoring and control.  hope this helps.

    0
    #150901

    Ken Feldman
    Participant

    Apparently there are a lot of hacks doing control charts without the slightest understanding as to what they are doing!!!!!  Again, do we blame SS or the execution and deployment of it’s methodology and tools?
    Can anyone tell, what interpretation you can draw from P chart. The same thing you can interpret from any other control chart.  That is, whether the process behavior is stable or whether there are signals of common cause or special cause variation.  In this case it is the proportion defective.
    What is the significance of Control Limits, when 100% quality is required by the end customer. For the millionith time, control limits and specs are not the same.  The control limits tell you what the process is doing and your spec of 100% tells you what the customer wants.  If they don’t match then you have a problem.
    Can we really differentiate between Special & Common cause in P Chart. Yes, within certain constraints.  If your proportion defective is very large or very small the control limits may send artificial signals.  The P chart is based on a binomial assumption but the control limits are constructed as a symmetrical boundary.  The binomial is reasonably symmetrical until your p falls in a range beyond about .9 or .1.  That is why some recommend using an I/MR chart instead of the P chart.
    e.g:- out of 1800 items supplied, one item is defective. Even if the deviation is into Upper control limit, can we say firmly is it because of special Cause or Common Cause.  You can’t firmly say anything since the control limits are not intended to encompass 100% of the possible common cause just a whole bunch of it.  You can have a point outside the upper limit and it still be common cause.  But that is rare.  Obviously you can’t have more than 100% nor less than 0%. 
    I draw P Chart on the basis of Daily Production, i.e, Percent Defective. I am hardly able to see any value addition. Well, if you don’t see any value, big DUH, stop doing it.  If you aren’t meeting your customer specs, then stop worrying about control charts and figure out how to improve your process.
    Can you expert, put some light on this topic. The expert has shed light.

    0
    #150910

    p chart
    Participant

    As always, there is little to argue with the dark side’s black and white summary of basic facts. There are two things that I think should be added to the list of the basics of control chart.
    1. The control chart was developed under the assumption of sampling. Thus the limits are both a function of the proportion defefective and the total sample size. Chart the same proportion defective based on a base of 100, 1000, 10,000 or 100,000 and the picture changes quite dramatically.
    2. In this same vein, if you take all units produced (1,800) and determine the proportion defective (1), your total proportion defective is 1/1800 = 0.000556. Because you are not sampling there is also no room for random error (bias etc. is a different story). So, yes the p-chart control limits have no logical or substantive meaning.
    Control charts are conceptualized for the purpose of the economic control of ongoing processes. They are not meant to be easy to digest graphical cartoons so that managers get colorful, scientifically looking displays of yesterday’s performance. Maybe Deming should have added the one critical step: THINK to his Plan, Do, Check, Act.
    Six Sigma is not responsible for foggy thinking and the nonsensical use of the tools kit of quality. To this day, there is still a world of a difference between a hammer and a skrewdriver … .  

    0
    #150911

    Charles H
    Participant

    Good post.  Just one thing – Deming never used plan, do, check, act – he used plan, do, study, act – and he was very emphatic about it, even saying “It was always study – I don’t know who changed it, but it is study, not check!”  To Deming, words had meaning and he was the consumate thinker, always developing theories, studying them and changing them when disproven.  Study has the element of analysis and thinking contained within it.  Check does not.  Let’s be true to Dr. Deming and use the PDSA, not PDCA cycle.

    0
    #150912

    Mikel
    Member

    Amen. Well said.
    There were loads of hacks doing quality before Six Sigma and there still are.
    There were loads of young punks who thought they knew everything when I was a pup and there still are.
    Some things never change.
    Deming knew what he was talking about with the “hope for instant pudding”. Unfortunately Six Sigma is  currently the most visible face of that today; Lean is the second most visible face. Most who have been trained in either do not have the knowledge or experience to do anything more than to go learn more under the mentorship of someone who knows enough to guide them, if they are really lucky.

    0
    #150914

    p chart
    Participant

    Charles,
    You are absolutely correct … I guess, eventually, we all get sucked into the CNN sound bits of the “general business literature” and recite the slogans that we have heard over and over and over again … until they finally become the eternal and unquestionable truth. Deming would not have been amused :-). Thanks for the reminder.

    0
    #150917

    Charles H
    Participant

    No worries, I know what you mean. 
    Speaking of Deming being “amused,”  I wonder, can anyone find a picture of him with a smile on his face?  I do think he had a sense of humor, but he kept it to himself.  ;-)  He was a great man and it is sad how he has been forgotten.  Yeah, we all refer to him and such, but no one seems interested in studying the man and his theories, anymore.  The WEDI seminars aren’t being well attended these days.  I do hear that the Deming Scholars MBA Program at Fordham is still going, though.  Maybe, someday, we will get back to his message and theories and give them their due.  That is my eternal hope.

    0
    #150918

    p chart
    Participant

    Charles,
    What I find interesting is the fact that in most MBA and PhD programs Six Sigma plays a very minor role. By contrast quality and systems theory are very well represented and alive in operations management classes. In general business administration education has moved away from integrating fields such as operations management, marketing, finance etc. towards specialization simply due to the fact that PhDs are trained in distinct areas of business administration. As a result, the kind of integrative view of Deming (a statistician by training who turned into a business strategist) has fallen to the sideways. Instead, bits and pieces of his tremendous intellectual achievements are now being packaged into quality, statistics, management, strategy etc.

    0
Viewing 34 posts - 1 through 34 (of 34 total)

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