Read the pages and make a very brief summary of them with your own words, please. It is what you understand. Don't make copy-paste, please. Mention important parts only. Also, you will put your comments and ideas about the topic. Please don't write item by item. Write the summary in paragraph form. UCL or LCL, or there will be a "run" of several points in arow above or below the average line. When a penetration ora lengthy run appears, this is the control chart's signal thatother tools is that it does not help us understand whether something is wrong that requires immediate attention.As long as the plots stay between the limits and don'tcongregate on one side or the other of the process averageline, the process is in statistical control. If either of these con-ditions is not met, then we can say that the process is notin statistical control or simply is "out of control"-hence theControl Charts4The problem with the run chart and, in fact, many of thethe variation is the result of special causes-things such aschanges in the materials used, machine problems, lack ofemployee training-or common causes that are purely ran-dom. Not until Dr. Walter Shewhart made that distinction1.79.1113 1517 19 21Workdayin the 1920s was there a real chance of improving processesthrough the use of statistical techniques. Shewhart, then anemployee of Bell Laboratories, developed the control chartto separate the special causes from the common causes.In evaluating problems and finding solutions for them, itis important to distinguish between special causes and com-mon causes. Figure 15.27 shows a typical control chart. Dataare plotted over time, just as with a run chart; the differenceis that the data stay between the upper control limit (UCL)and the lower control limit (LCL) while varying about thecenterline or average only so long as the variation is the re-sult of common causes (i.e., statistical variation). Whenever aspecial cause (nonstatistical cause) impacts the process, oneof two things will happen: Either a plot point will penetratename of the chart.If you understand that it is the UCL, LCL, and processaverage lines added to the run chart that make the differ-ence, you may wonder how those lines are set. The position-ing of the lines cannot be arbitrary. Nor can they merelyreflect what you want out of the process, for example, basedon a specification. Such an approach won't help separatecommon causes from special causes, and it will only compli-cate attempts at process improvement. UCL, LCL, and pro-cess average must be determined by valid statistical means.Determination of UCL, LCL, and process average is fullycovered in Chapter 18, which is dedicated to the use of con-trol charts in statistical process control.All processes have built-in variability. A process thatis in statistical control will still be affected by its naturalrandom variability. Such a process will exhibit the normaldistribution of the bell curve. The more finely tuned theprocess, the less deviation from the process average and thenarrower the bell curve. (Refer to Figure 15.19, HistogramA and Histogram B.) This is at the heart of the control chartand is what makes it possible to define the limits and processFIGURE 15.28 Chart of Operator Defects for November.might include data relative to the environment, the peopleinvolved, the machine(s) used in the process, materials, andso on. Grouping data by common element or characteristicmakes it easier to understand the data and to pull insightsfrom them.Consider an example from a factory floor. One of thefactory's products requires five assemblers, all doing thesame thing at the same rate. Their output flows togetherfor inspection. Inspection has found an unacceptably highrate of defects in the products. Management forms a teamto investigate the problem with the objective of finding thecause and correcting it. They plot the data taken over the lastmonth (see Figure 15.28).The chart in Figure 15.28 plots all operator-induced de-fects for the month. The team believes that for this product,zero defects can be approached. If you were going to react tothis chart alone, how would you deal with the problem? Youhave five assemblers. Do they all contribute defects equally?This is hardly ever the case. The data can be stratified by theoperator to determine each individual's defect performance.The charts in Figure 15.29 do this.The five stratified charts in Figure 15.29 indicate thatone operator, Assembler B, is responsible for more defectsthan the other four combined. Assembler A also makes moreUCLaverage.Control charts are the appropriate tool to monitor pro-cesses. The properly used control chart will immediatelyalert the operator to any change in the process. The appro-priate response to that alert is to stop the process at once,preventing the production of defective product. Only afterthe special cause of the problem has been identified and cor-rected should the process be restarted. Having eliminateda problem's root cause, that problem should never recur.(Anything less, however, and it is sure to return eventu-ally.) Control charts also enable continual improvement ofprocesses. When a change is introduced to a process that isoperated under statistical process control charts, the effectof the change will be immediately seen. You know when youhave made an improvement. You also know when the changeis ineffective or even detrimental. This validates effectiveimprovements, which you will retain. This is enormouslydifficult when the process is not in statistical control becausethe process instability maskschanges deliberately made.To learn more about statistical process control and con-trol charts, study Chapter 18.Process AverageLCLSamplesthan twice as many errors as Assembler C or Assembler Dand eight times as many as Assembler E, the best performerof the group.FIGURE 15.27 Basic Control Chart.The performance of Assembler A and Assembler B mustbe brought up to the level of the others. Possible causes of theoperator-induced defects could be inherent skill, training,vision, attitude, attentiveness, and environmental factors,such as noise, lighting, and temperature in the operator'sworkstation area. The charts provide an indication of theplace to start making changes.The Pareto charts of Figure 15.6 also represent stratifi-cation. Figure 15.6 started with a series of defect types thatresults, good or bad, of anyweremost costly (the first chart). Then it took the worstcase, Miswires, and divided it into the kinds of miswires(the second chart). Then the worst kind, Hand Wrap, wassplit into several categories (the third chart). The dominantHand-Wrap defect category was operator induced. Finally,the Operator category was stratified by individual operator(the fourth chart).The power of stratification lies in the fact that if youstratify far enough, you will arrive at a root cause of theSTRATIFICATIONStratification is a simple tool in spite of its name. It involvesinvestigating the cause of a problem by grouping data intocategories. This grouping is called stratification. The groupsDefects

Dr. Walter Shewhart' s control chart reflects the Difference between Special causes and common…

Answered: Read the pages and make a very brief…

10th Edition

ISBN:9780134641287

Author:Elliot Aronson, Timothy D. Wilson, Robin M. Akert, Samuel R. Sommers

Publisher:Elliot Aronson, Timothy D. Wilson, Robin M. Akert, Samuel R. Sommers

Chapter1: Introducing Social Psychology

Section: Chapter Questions

Problem 1RQ1